{ "article": "REMAG: GPU-first SIESTA for magnetic materials", "as_of_local_date": "2026-10-10", "release_commit": "848091532f8f2b7268cd83e165458106d95c407d", "current_suite": { "run": "easy-magnetic-campaign-v181-20261010T060613Z", "summary": { "scientific_qualified": false, "performance_qualified": false, "local_cpu_verification_pending": false, "remote_complete": true, "budget_exhausted": true, "declared_workflows": 7, "observed_workflows": 4, "unstarted_workflows": [ "FePt16 full-SOC x/z", "Fe16 noncollinear", "Fe5PB2 full-SOC x/z" ], "remote_all_workflows_passed": false, "allocation_elapsed_seconds": 6457.169878971, "geometric_mean_speedup": null, "total_suite_speedup": null, "scope": "One A100-80GB and 16 reserved CPU cores. Current v178 binary, default tile modes, fresh processes and unchanged saved inputs. More upstream parallel layouts; no claim of exhaustive tuning, production convergence or general SIESTA average.", "qualified_workflows": 2, "allocation_released": true, "qualified_at_utc": "2026-10-10T08:07:25.065995+00:00", "aggregation_scope": "Equal-workflow geometric mean only after all seven declared workflows and all their required upstream solver families pass direct pristine-CPU science/export/convergence checks. Directional x/z pairs count once. Partial results retain explicit coverage, without a full-suite average." }, "workflows": [ { "name": "Fe3C collinear", "cases": [ "fe3c16_collinear" ], "required_families": [ "native", "elsi" ], "screens": [ { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 144.476083417 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 294.10390538800004 }, { "config_key": "elsi2-r16-b8", "family": "elsi", "passed": false, "wall_seconds": 38.36134438299996 }, { "config_key": "elsi1-r16-b8", "family": "elsi", "passed": false, "wall_seconds": 127.34655580399999 } ], "finalists": [ { "family": "native", "config_key": "native2-t4", "config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 152.31288163600004, "candidate": 2.834449581999934 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 135.462752691, "candidate": 2.839490099000045 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 144.06467930099996, "candidate": 2.9413141740001265 } } ], "case_timings": [ { "case": "fe3c16_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 152.31288163600004, 135.462752691, 144.06467930099996 ], "median_seconds": 144.06467930099996, "iterations": [ 73, 73, 73 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 685.0, 685.0, 685.0 ] }, "remag": { "seconds": [ 2.834449581999934, 2.839490099000045, 2.9413141740001265 ], "median_seconds": 2.839490099000045, "iterations": [ 65, 65, 72 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2049.0, 2049.0, 2049.0 ] }, "speedup_over_upstream_gpu": 50.73610904709045, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 144.06467930099996, "candidate": 2.839490099000045 }, "remote_provisional_speedup": 50.73610904709045, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 50.73610904709045 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 2.834449581999934 }, { "repeat": 2, "passed": true, "seconds": 2.839490099000045 }, { "repeat": 3, "passed": true, "seconds": 2.9413141740001265 } ], "remote_qualified": false, "performance_qualified": false, "speedup_over_upstream_gpu": null, "missing_families": [ "elsi" ], "selected_config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native2-t4", "selected_family": "native", "medians": { "upstream": 144.06467930099996, "candidate": 2.839490099000045 }, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 152.31288163600004, "candidate": 2.834449581999934 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 135.462752691, "candidate": 2.839490099000045 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 144.06467930099996, "candidate": 2.9413141740001265 } } ], "remote_provisional_speedup": 50.73610904709045, "scientific_qualified": false, "best_verified_finalist_speedup": 50.73610904709045 }, { "name": "Fe54 collinear", "cases": [ "fe54_collinear" ], "required_families": [ "native", "elsi" ], "screens": [ { "config_key": "elsi2-r8", "family": "elsi", "passed": false, "wall_seconds": 202.782551561 }, { "config_key": "elsi2-r4-b32", "family": "elsi", "passed": true, "wall_seconds": 102.93532000699997 }, { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 124.33507310599998 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 270.22637843300004 } ], "finalists": [ { "family": "native", "config_key": "native2-t4", "config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 131.79213680500015, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 129.12238525899966, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 125.92732839500013, "candidate": 7.414528536000034 } } ], "case_timings": [ { "case": "fe54_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 131.79213680500015, 129.12238525899966, 125.92732839500013 ], "median_seconds": 129.12238525899966, "iterations": [ 74, 73, 71 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 695.0, 695.0, 695.0 ] }, "remag": { "seconds": [ 7.425738153000111, 7.843675305000033, 7.414528536000034 ], "median_seconds": 7.425738153000111, "iterations": [ 107, 127, 114 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3229.0, 3285.0 ] }, "speedup_over_upstream_gpu": 17.388491567916684, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 129.12238525899966, "candidate": 7.425738153000111 }, "remote_provisional_speedup": 17.388491567916684, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 17.388491567916684 }, { "family": "elsi", "config_key": "elsi2-r4-b32", "config": { "route": "elsi-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 104.26391460900004, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 103.34408651400008, "candidate": 7.414528536000034 } } ], "case_timings": [ { "case": "fe54_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 104.23722245900012, 104.26391460900004, 103.34408651400008 ], "median_seconds": 104.23722245900012, "iterations": [ 79, 79, 79 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2378.0, 2370.0, 2378.0 ] }, "remag": { "seconds": [ 7.425738153000111, 7.843675305000033, 7.414528536000034 ], "median_seconds": 7.425738153000111, "iterations": [ 107, 127, 114 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3229.0, 3285.0 ] }, "speedup_over_upstream_gpu": 14.037287648890057, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 }, "remote_provisional_speedup": 14.037287648890057, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 14.037287648890057 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 7.425738153000111 }, { "repeat": 2, "passed": true, "seconds": 7.843675305000033 }, { "repeat": 3, "passed": true, "seconds": 7.414528536000034 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 14.037287648890057, "selected_config": { "route": "elsi-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "selected_config_key": "elsi2-r4-b32", "selected_family": "elsi", "medians": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 }, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 104.26391460900004, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 103.34408651400008, "candidate": 7.414528536000034 } } ], "remote_provisional_speedup": 14.037287648890057, "scientific_qualified": true, "best_verified_finalist_speedup": 14.037287648890057 }, { "name": "Fe3C noncollinear", "cases": [ "fe3c16_noncollinear" ], "required_families": [ "native" ], "screens": [ { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 429.20419752099997 }, { "config_key": "native1-t4", "family": "native", "passed": true, "wall_seconds": 433.5560178679998 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 370.9022678799997 } ], "finalists": [ { "family": "native", "config_key": "native2-r4-b32", "config": { "route": "native-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 360.35077029400054, "candidate": 3.3475332680000065 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 360.2013342050004, "candidate": 3.2520779189999303 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 366.6469657509997, "candidate": 3.2553646079995815 } } ], "case_timings": [ { "case": "fe3c16_noncollinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 360.35077029400054, 360.2013342050004, 366.6469657509997 ], "median_seconds": 360.35077029400054, "iterations": [ 100, 100, 100 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2182.0, 2182.0, 2180.0 ] }, "remag": { "seconds": [ 3.3475332680000065, 3.2520779189999303, 3.2553646079995815 ], "median_seconds": 3.2553646079995815, "iterations": [ 71, 65, 69 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2199.0, 2199.0, 2199.0 ] }, "speedup_over_upstream_gpu": 110.6944424622948, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 360.35077029400054, "candidate": 3.2553646079995815 }, "remote_provisional_speedup": 110.6944424622948, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 110.6944424622948 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 3.3475332680000065 }, { "repeat": 2, "passed": true, "seconds": 3.2520779189999303 }, { "repeat": 3, "passed": true, "seconds": 3.2553646079995815 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 110.6944424622948, "selected_config": { "route": "native-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "selected_config_key": "native2-r4-b32", "selected_family": "native", "medians": { "upstream": 360.35077029400054, "candidate": 3.2553646079995815 }, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 360.35077029400054, "candidate": 3.3475332680000065 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 360.2013342050004, "candidate": 3.2520779189999303 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 366.6469657509997, "candidate": 3.2553646079995815 } } ], "remote_provisional_speedup": 110.6944424622948, "scientific_qualified": true, "best_verified_finalist_speedup": 110.6944424622948 }, { "name": "Fe2B full-SOC x/z", "cases": [ "fe2b12_soc_z", "fe2b12_soc_x" ], "required_families": [ "native" ], "screens": [ { "config_key": "native1-t4", "family": "native", "passed": true, "wall_seconds": 371.74052989200027 }, { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 403.99989258799997 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 553.965018502 } ], "finalists": [ { "family": "native", "config_key": "native1-t4", "config": { "route": "native-1stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "gpu_execution_verified": true, "remote_qualified": false, "pairs": [], "scientific_qualified": false, "performance_qualified": false, "speedup_over_upstream_gpu": null } ], "candidate_repeats": [], "remote_qualified": false, "performance_qualified": false, "speedup_over_upstream_gpu": null, "failure": "Insufficient reserved time for the next complete trial and checkpoint", "budget_exhausted": true, "scientific_qualified": false } ], "hardware": [ "NVIDIA A100-SXM4-80GB, GPU-aa998a4d-5016-320b-ebe3-7abc148490f5, 580.95.05" ], "cpu_cores_requested": 16, "candidate_settings": { "REMAG_NONLOCAL_H0_POISON_HOST": "0", "REMAG_NONLOCAL_H0": "1", "REMAG_NONLOCAL_H0_SHADOW": "0", "OMP_NUM_THREADS": "1", "OPENBLAS_NUM_THREADS": "1", "MKL_NUM_THREADS": "1", "REMAG_GRID": "1", "REMAG_DM": "1", "REMAG_GRID_EPOCH": "1", "REMAG_GRID_VERIFY_EPOCH": "0", "REMAG_NORMALIZE": "1", "REMAG_OCCUPATIONS": "1", "REMAG_RESIDENT_SPECTRA": "1", "REMAG_HS": "1", "REMAG_HARTREE": "1", "REMAG_RESIDENT_GRID": "1", "REMAG_RESIDENT_SCF": "1", "REMAG_DEVICE_SCF": "1", "REMAG_DEVICE_VECTOR_CACHE": "1", "REMAG_BLOCK_SOLVER": "0", "REMAG_BLOCK_WIDTH": "32", "REMAG_BLOCK_LOCAL_SOLVER": "jacobi", "REMAG_FORCE_SHADOW": "0", "REMAG_FORCE_SHADOW_RESIDENT": "0", "REMAG_GRID_FORCE": "1", "REMAG_ATOMIC_GRID_FORCE": "1", "REMAG_ATOMIC_FORCE_SHADOW": "0", "REMAG_FINAL_GRID": "1", "REMAG_FINAL_GRID_SHADOW": "0", "REMAG_FINAL_DENSITY": "1", "REMAG_FINAL_DENSITY_SHADOW": "0", "REMAG_ORBITAL_BASIS_SHADOW": "0", "REMAG_ORBITAL_BASIS": "1", "REMAG_ATOMIC_BASIS_SHADOW": "0", "REMAG_ATOMIC_BASIS": "1", "REMAG_TWO_CENTER_SHADOW": "0", "REMAG_PAIR_FORCE_SHADOW": "0", "REMAG_NONLOCAL_FORCE_SHADOW": "0", "REMAG_INTEGRAL_FORCE": "1", "REMAG_POPULATION": "1", "REMAG_TRACE_IO": "0", "REMAG_TRACE_CUDA": "0", "REMAG_GPU_INITIALIZATION": "1", "REMAG_GPU_HISTORY": "1", "REMAG_FINAL_SPECTRA": "1", "REMAG_GATE_H_EXPORT": "1", "REMAG_DEFER_SCF_LOG": "1", "GFORTRAN_FORMATTED_BUFFER_SIZE": "1048576", "GFORTRAN_UNBUFFERED_ALL": "0", "GFORTRAN_UNBUFFERED_PRECONNECTED": "0", "REMAG_BLOCK_SHARED_JACOBI": "1", "REMAG_INITIAL_MESH": "1", "REMAG_DENSITY_UNIQUE_PAIRS": "1", "REMAG_SCF_SOLVER_LANES": "8", "REMAG_SCF_WARM_START": "1", "REMAG_RADIAL_BATCH": "1", "REMAG_GRID_POTENTIAL_SPINS": "1", "REMAG_SCF_STAGE_PROFILE": "0", "REMAG_HOST_ID": "03524688dfa14fe899c126d2d5962295", "LD_LIBRARY_PATH": "/opt/remag-candidate/remag-native:/opt/elpa-gpu/lib:/usr/local/cuda/lib64", "REMAG_BLOCK_FUSED_MATRIX": "0", "REMAG_BLOCK_FP64_VECTORS": "0", "REMAG_BLOCK_FP64_MATRIX": "0", "OMPI_ALLOW_RUN_AS_ROOT": "1", "OMPI_ALLOW_RUN_AS_ROOT_CONFIRM": "1" }, "upstream_settings": { "LD_LIBRARY_PATH": "/opt/elpa-gpu/lib:/usr/local/cuda/lib64", "REMAG_TRACE_CUDA": "0", "REMAG_TRACE_IO": "0", "REMAG_SCF_STAGE_PROFILE": "0", "REMAG_HOST_ID": "03524688dfa14fe899c126d2d5962295", "OMPI_ALLOW_RUN_AS_ROOT": "1", "OMPI_ALLOW_RUN_AS_ROOT_CONFIRM": "1" }, "binaries": { "/opt/remag-candidate/Src/siesta": "e9c444827228c6b680dc5f9a62bf54f422a0220e5a6add27766ecbc0a2da40d1", "/opt/remag-candidate/remag-native/libremag_gpu.so": "479381a3d20051a2540c12c823d2143bee091cd53c51d8ee4bfb2068dcaa50b6", "/opt/upstream-gpu/Src/siesta": "59781a7c103181e30ca77cced4c5c1256c101c5fbbb88033263e8e5a83d44c66" }, "plan": { "created_at_utc": "2026-10-10T06:01:18.054405+00:00", "suite": "Expanded known-converging magnetic workflows with native and applicable ELSI finalists", "selection": "Ten existing physical inputs across seven workflows, selected for demonstrated CPU/candidate convergence before new timing. Four compounds (Fe3C, Fe2B, FePt, Fe5PB2) plus two elemental-Fe controls. Fe4N excluded for known convergence difficulty; all future failures retained.", "candidate_run": "fp64-matrix-qualification-v178-20261009T215709Z", "candidate_image": "im-1U2ZnxmQQlDqw1VEGqVh5x", "workflows": [ { "name": "Fe3C collinear", "cases": [ "fe3c16_collinear" ], "configs": [ "native2-t4", "native2-r4-b32", "elsi2-r16-b8", "elsi1-r16-b8" ], "required_families": [ "native", "elsi" ] }, { "name": "Fe54 collinear", "cases": [ "fe54_collinear" ], "configs": [ "elsi2-r8", "elsi2-r4-b32", "native2-t4", "native2-r4-b32" ], "required_families": [ "native", "elsi" ] }, { "name": "Fe3C noncollinear", "cases": [ "fe3c16_noncollinear" ], "configs": [ "native2-t4", "native1-t4", "native2-r4-b32" ], "required_families": [ "native" ] }, { "name": "Fe2B full-SOC x/z", "cases": [ "fe2b12_soc_z", "fe2b12_soc_x" ], "configs": [ "native1-t4", "native2-t4", "native2-r4-b32" ], "required_families": [ "native" ] }, { "name": "FePt16 full-SOC x/z", "cases": [ "fept16_soc_z", "fept16_soc_x" ], "configs": [ "native2-t4", "native2-t1", "native2-r4-b32" ], "required_families": [ "native" ] }, { "name": "Fe16 noncollinear", "cases": [ "fe16_noncollinear" ], "configs": [ "native1-t4", "native2-t4", "native2-r4-b32" ], "required_families": [ "native" ] }, { "name": "Fe5PB2 full-SOC x/z", "cases": [ "fe5pb2_32_soc_z", "fe5pb2_32_soc_x" ], "configs": [ "native2-t4", "native2-r4-b32" ], "required_families": [ "native" ], "trial_timeout_seconds": 1800 } ], "cases": { "fe54_collinear": { "remote_reference_run": "scf-solver-lanes-regression-v116b-20261008T174638Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/scf-solver-lanes-regression-v116b-20261008T174638Z/fe54_collinear/cpu", "input_sha256": "705a76444e50a1096898f71b64f769c5575128c4ab93acce486f2149784e1005", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" }, "prior_cpu_iterations": 89, "atoms": 54, "spin": "polarized" }, "fe16_noncollinear": { "remote_reference_run": "scf-seeded-regression-v119b-20261008T191317Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/scf-seeded-regression-v119b-20261008T191317Z/fe16_noncollinear/cpu", "input_sha256": "3602afc93642d3aaf393530d0c440b3dc65b57af72360e4a390fdcfe28f732ec", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" }, "prior_cpu_iterations": 77, "atoms": 16, "spin": "non-collinear" }, "fe3c16_collinear": { "remote_reference_run": "carbide-gpu-v169-20261009T170851Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/carbide-local-reference-v168-20261009T164908Z/fe3c16_collinear/cpu", "input_sha256": "833c95fbd300f7ee745f87027f613379f47fbcff7d5c12f8abd930b0c1e42cbf", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" }, "prior_cpu_iterations": 68, "atoms": 16, "spin": "polarized" }, "fe2b12_soc_z": { "remote_reference_run": "boride-gpu-v173-20261009T202117Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/boride-local-reference-v171-20261009T172616Z/fe2b12_soc_z/cpu", "input_sha256": "4ed2ed5b40f6cd2fc689061fcefbfa31fd6ed935e3ff8544ffb870a307935ddf", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 47, "atoms": 12, "spin": "spin-orbit" }, "fe2b12_soc_x": { "remote_reference_run": "boride-gpu-v173-20261009T202117Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/boride-local-reference-v171-20261009T172616Z/fe2b12_soc_x/cpu", "input_sha256": "1987a287dd32acdf7fe7e190e9363673b31b257ba2d3722a62674d69f3e031ac", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 75, "atoms": 12, "spin": "spin-orbit" }, "fe3c16_noncollinear": { "remote_reference_run": "carbide-gpu-v169-20261009T170851Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/carbide-local-reference-v168-20261009T164908Z/fe3c16_noncollinear/cpu", "input_sha256": "d4561abad7720d8b554784fa3a44c5f92e46164651e5422beee89182333bb05e", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" }, "prior_cpu_iterations": 72, "atoms": 16, "spin": "non-collinear" }, "fept16_soc_z": { "remote_reference_run": "upstream-gpu-refresh-v143-20261009T101214Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/upstream-gpu-refresh-v143-20261009T101214Z/fept16_soc_z/cpu", "input_sha256": "ea290e77b26b088836049ef4ef3376ef8067509fa1cd2a7dc1910b601c28e46b", "pseudopotentials": { "Fe_fept_SOC.psf": "47a51f8ef02f4f64146a55f5445d5c7142489b1dafdeb300d25adc1ccc9f1978", "Pt_fept_SOC.psf": "2be713b0971d72ab0b47d1965b0539e494a20a42a2002a09cd9d5f3eec0d6c47" }, "prior_cpu_iterations": 87, "atoms": 16, "spin": "spin-orbit" }, "fept16_soc_x": { "remote_reference_run": "upstream-gpu-refresh-v143-20261009T101214Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/upstream-gpu-refresh-v143-20261009T101214Z/fept16_soc_x/cpu", "input_sha256": "2636385ff78d16342c8d79e6ed814a52b8009ae73b594122d7fbfb620f2fb5a5", "pseudopotentials": { "Fe_fept_SOC.psf": "47a51f8ef02f4f64146a55f5445d5c7142489b1dafdeb300d25adc1ccc9f1978", "Pt_fept_SOC.psf": "2be713b0971d72ab0b47d1965b0539e494a20a42a2002a09cd9d5f3eec0d6c47" }, "prior_cpu_iterations": 81, "atoms": 16, "spin": "spin-orbit" }, "fe5pb2_32_soc_z": { "remote_reference_run": "candidate-magnet-gpu-v179-20261009T222951Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/candidate-magnet-local-reference-v177-20261009T214755Z/fe5pb2_32_soc_z/cpu", "input_sha256": "fa042284e6a283e00cf013ef12060a1fb760e7354ce7c9189574636561b99123", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "P.psml": "56914dbba0f5c0a625efb4c0f2a1a34c50dff4100ef3dbd0a9454229ebd1c207", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 88, "atoms": 32, "spin": "spin-orbit" }, "fe5pb2_32_soc_x": { "remote_reference_run": "candidate-magnet-gpu-v179-20261009T222951Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/candidate-magnet-local-reference-v177-20261009T214755Z/fe5pb2_32_soc_x/cpu", "input_sha256": "76d3a02f41466bcc148615565c74f1ab175171006fb90263dae2e81e2a08dbca", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "P.psml": "56914dbba0f5c0a625efb4c0f2a1a34c50dff4100ef3dbd0a9454229ebd1c207", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 128, "atoms": 32, "spin": "spin-orbit" } }, "configs": { "native2-t4": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "native1-t4": { "route": "native-1stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "native2-t1": { "route": "native-2stage", "block_size": 64, "threads": 1, "mpi_ranks": 1 }, "elsi2-r8": { "route": "elsi-2stage", "block_size": 64, "threads": 1, "mpi_ranks": 8 }, "native2-r4-b32": { "route": "native-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "elsi2-r4-b32": { "route": "elsi-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "elsi2-r16-b8": { "route": "elsi-2stage", "block_size": 8, "threads": 1, "mpi_ranks": 16 }, "elsi1-r16-b8": { "route": "elsi-1stage", "block_size": 8, "threads": 1, "mpi_ranks": 16 } }, "repetitions": 3, "candidate_tile_flags": { "REMAG_BLOCK_FUSED_MATRIX": "0", "REMAG_BLOCK_FP64_VECTORS": "0", "REMAG_BLOCK_FP64_MATRIX": "0" }, "trial_timeout_seconds": 600, "allocation_timeout_seconds": 7200, "gpu_limit": 1, "convergence": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "baseline_selection": "Screen allowed solver-only settings. Repeat the fastest passing native finalist and, for each collinear workflow, the fastest passing ELSI finalist three times. Use the lowest median among fully passing finalists. Do not claim a full-suite aggregate if a required family lacks three passing repetitions.", "aggregate": "Equal-workflow geometric mean only after all seven declared workflows and all their required upstream solver families pass direct pristine-CPU science/export/convergence checks. Directional x/z pairs count once. Partial results retain explicit coverage, without a full-suite average.", "scope": "One A100-80GB and 16 reserved CPU cores. Current v178 binary, default tile modes, fresh processes and unchanged saved inputs. More upstream parallel layouts; no claim of exhaustive tuning, production convergence or general SIESTA average.", "cpu_cores_requested": 16, "budget_status": "configured", "budget_policy": "Hard allocation ceiling plus an earlier software deadline; no process begins without enough remaining time for its per-case timeout and checkpoint reserve. No automatic rerun after interruption, cap extension or follow-on GPU launch.", "budget_hours": 2, "budget_basis": "Conservative first-allocation assumption while the optional user budget preference is unanswered. No expansion or follow-on GPU allocation without resolving the remaining campaign budget." }, "inputs": { "fe54_collinear": { "input_fdf": "SystemName REMAG regression fe54_collinear\nSystemLabel remag\nNumberOfAtoms 54\nNumberOfSpecies 1\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n8.61 0.0 0.0\n0.0 8.61 0.0\n0.0 0.0 8.61\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.435000000000 1.435000000000 1.435000000000 1\n0.000000000000 0.000000000000 2.870000000000 1\n1.435000000000 1.435000000000 4.305000000000 1\n0.000000000000 0.000000000000 5.740000000000 1\n1.435000000000 1.435000000000 7.175000000000 1\n0.000000000000 2.870000000000 0.000000000000 1\n1.435000000000 4.305000000000 1.435000000000 1\n0.000000000000 2.870000000000 2.870000000000 1\n1.435000000000 4.305000000000 4.305000000000 1\n0.000000000000 2.870000000000 5.740000000000 1\n1.435000000000 4.305000000000 7.175000000000 1\n0.000000000000 5.740000000000 0.000000000000 1\n1.435000000000 7.175000000000 1.435000000000 1\n0.000000000000 5.740000000000 2.870000000000 1\n1.435000000000 7.175000000000 4.305000000000 1\n0.000000000000 5.740000000000 5.740000000000 1\n1.435000000000 7.175000000000 7.175000000000 1\n2.870000000000 0.000000000000 0.000000000000 1\n4.305000000000 1.435000000000 1.435000000000 1\n2.870000000000 0.000000000000 2.870000000000 1\n4.305000000000 1.435000000000 4.305000000000 1\n2.870000000000 0.000000000000 5.740000000000 1\n4.305000000000 1.435000000000 7.175000000000 1\n2.870000000000 2.870000000000 0.000000000000 1\n4.305000000000 4.305000000000 1.435000000000 1\n2.870000000000 2.870000000000 2.870000000000 1\n4.305000000000 4.305000000000 4.305000000000 1\n2.870000000000 2.870000000000 5.740000000000 1\n4.305000000000 4.305000000000 7.175000000000 1\n2.870000000000 5.740000000000 0.000000000000 1\n4.305000000000 7.175000000000 1.435000000000 1\n2.870000000000 5.740000000000 2.870000000000 1\n4.305000000000 7.175000000000 4.305000000000 1\n2.870000000000 5.740000000000 5.740000000000 1\n4.305000000000 7.175000000000 7.175000000000 1\n5.740000000000 0.000000000000 0.000000000000 1\n7.175000000000 1.435000000000 1.435000000000 1\n5.740000000000 0.000000000000 2.870000000000 1\n7.175000000000 1.435000000000 4.305000000000 1\n5.740000000000 0.000000000000 5.740000000000 1\n7.175000000000 1.435000000000 7.175000000000 1\n5.740000000000 2.870000000000 0.000000000000 1\n7.175000000000 4.305000000000 1.435000000000 1\n5.740000000000 2.870000000000 2.870000000000 1\n7.175000000000 4.305000000000 4.305000000000 1\n5.740000000000 2.870000000000 5.740000000000 1\n7.175000000000 4.305000000000 7.175000000000 1\n5.740000000000 5.740000000000 0.000000000000 1\n7.175000000000 7.175000000000 1.435000000000 1\n5.740000000000 5.740000000000 2.870000000000 1\n7.175000000000 7.175000000000 4.305000000000 1\n5.740000000000 5.740000000000 5.740000000000 1\n7.175000000000 7.175000000000 7.175000000000 1\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n1 0 0 0.0\n0 1 0 0.0\n0 0 1 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin polarized\n%block DM.InitSpin\n1 2.5 0.0 0.\n2 2.5 0.0 0.\n3 2.5 0.0 0.\n4 2.5 0.0 0.\n5 2.5 0.0 0.\n6 2.5 0.0 0.\n7 2.5 0.0 0.\n8 2.5 0.0 0.\n9 2.5 0.0 0.\n10 2.5 0.0 0.\n11 2.5 0.0 0.\n12 2.5 0.0 0.\n13 2.5 0.0 0.\n14 2.5 0.0 0.\n15 2.5 0.0 0.\n16 2.5 0.0 0.\n17 2.5 0.0 0.\n18 2.5 0.0 0.\n19 2.5 0.0 0.\n20 2.5 0.0 0.\n21 2.5 0.0 0.\n22 2.5 0.0 0.\n23 2.5 0.0 0.\n24 2.5 0.0 0.\n25 2.5 0.0 0.\n26 2.5 0.0 0.\n27 2.5 0.0 0.\n28 2.5 0.0 0.\n29 2.5 0.0 0.\n30 2.5 0.0 0.\n31 2.5 0.0 0.\n32 2.5 0.0 0.\n33 2.5 0.0 0.\n34 2.5 0.0 0.\n35 2.5 0.0 0.\n36 2.5 0.0 0.\n37 2.5 0.0 0.\n38 2.5 0.0 0.\n39 2.5 0.0 0.\n40 2.5 0.0 0.\n41 2.5 0.0 0.\n42 2.5 0.0 0.\n43 2.5 0.0 0.\n44 2.5 0.0 0.\n45 2.5 0.0 0.\n46 2.5 0.0 0.\n47 2.5 0.0 0.\n48 2.5 0.0 0.\n49 2.5 0.0 0.\n50 2.5 0.0 0.\n51 2.5 0.0 0.\n52 2.5 0.0 0.\n53 2.5 0.0 0.\n54 2.5 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "705a76444e50a1096898f71b64f769c5575128c4ab93acce486f2149784e1005", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" } }, "fe16_noncollinear": { "input_fdf": "SystemName REMAG regression fe16_noncollinear\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 1\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.74 0.0 0.0\n0.0 5.74 0.0\n0.0 0.0 5.74\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.435000000000 1.435000000000 1.435000000000 1\n0.000000000000 0.000000000000 2.870000000000 1\n1.435000000000 1.435000000000 4.305000000000 1\n0.000000000000 2.870000000000 0.000000000000 1\n1.435000000000 4.305000000000 1.435000000000 1\n0.000000000000 2.870000000000 2.870000000000 1\n1.435000000000 4.305000000000 4.305000000000 1\n2.870000000000 0.000000000000 0.000000000000 1\n4.305000000000 1.435000000000 1.435000000000 1\n2.870000000000 0.000000000000 2.870000000000 1\n4.305000000000 1.435000000000 4.305000000000 1\n2.870000000000 2.870000000000 0.000000000000 1\n4.305000000000 4.305000000000 1.435000000000 1\n2.870000000000 2.870000000000 2.870000000000 1\n4.305000000000 4.305000000000 4.305000000000 1\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin non-collinear\n%block DM.InitSpin\n1 2.5 30.0 0.\n2 2.5 30.0 0.\n3 2.5 30.0 0.\n4 2.5 30.0 0.\n5 2.5 30.0 0.\n6 2.5 30.0 0.\n7 2.5 30.0 0.\n8 2.5 30.0 0.\n9 2.5 30.0 0.\n10 2.5 30.0 0.\n11 2.5 30.0 0.\n12 2.5 30.0 0.\n13 2.5 30.0 0.\n14 2.5 30.0 0.\n15 2.5 30.0 0.\n16 2.5 30.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "3602afc93642d3aaf393530d0c440b3dc65b57af72360e4a390fdcfe28f732ec", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" } }, "fe3c16_collinear": { "input_fdf": "SystemName REMAG regression fe3c16_collinear\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n2 6 C\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.081 0.0 0.0\n0.0 6.754 0.0\n0.0 0.0 4.515\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.184600000000 0.059400000000 0.334000000000 1\n0.184600000000 0.440600000000 0.334000000000 1\n0.815400000000 0.559400000000 0.666000000000 1\n0.315400000000 0.559400000000 0.834000000000 1\n0.684600000000 0.440600000000 0.166000000000 1\n0.684600000000 0.059400000000 0.166000000000 1\n0.315400000000 0.940600000000 0.834000000000 1\n0.815400000000 0.940600000000 0.666000000000 1\n0.034600000000 0.250000000000 0.837700000000 1\n0.965400000000 0.750000000000 0.162300000000 1\n0.465400000000 0.750000000000 0.337700000000 1\n0.534600000000 0.250000000000 0.662300000000 1\n0.898400000000 0.250000000000 0.446700000000 2\n0.101600000000 0.750000000000 0.553300000000 2\n0.601600000000 0.750000000000 0.946700000000 2\n0.398400000000 0.250000000000 0.053300000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin polarized\n%block DM.InitSpin\n1 2.5 0.0 0.0\n2 2.5 0.0 0.0\n3 2.5 0.0 0.0\n4 2.5 0.0 0.0\n5 2.5 0.0 0.0\n6 2.5 0.0 0.0\n7 2.5 0.0 0.0\n8 2.5 0.0 0.0\n9 2.5 0.0 0.0\n10 2.5 0.0 0.0\n11 2.5 0.0 0.0\n12 2.5 0.0 0.0\n13 0.0 0.0 0.0\n14 0.0 0.0 0.0\n15 0.0 0.0 0.0\n16 0.0 0.0 0.0\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nPAO.EnergyShift 300 meV\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "833c95fbd300f7ee745f87027f613379f47fbcff7d5c12f8abd930b0c1e42cbf", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" } }, "fe2b12_soc_z": { "input_fdf": "SystemName REMAG regression fe2b12_soc_z\nSystemLabel remag\nNumberOfAtoms 12\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.077 0 0\n0 5.077 0\n0 0 4.254\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.167300000000 0.667300000000 0.000000000000 1\n0.332700000000 0.167300000000 0.000000000000 1\n0.832700000000 0.332700000000 0.000000000000 1\n0.667300000000 0.832700000000 0.000000000000 1\n0.667300000000 0.167300000000 0.500000000000 1\n0.832700000000 0.667300000000 0.500000000000 1\n0.332700000000 0.832700000000 0.500000000000 1\n0.167300000000 0.332700000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 0.0 0.\n2 2.0 0.0 0.\n3 2.0 0.0 0.\n4 2.0 0.0 0.\n5 2.0 0.0 0.\n6 2.0 0.0 0.\n7 2.0 0.0 0.\n8 2.0 0.0 0.\n9 0.0 0.0 0.\n10 0.0 0.0 0.\n11 0.0 0.0 0.\n12 0.0 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "4ed2ed5b40f6cd2fc689061fcefbfa31fd6ed935e3ff8544ffb870a307935ddf", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } }, "fe2b12_soc_x": { "input_fdf": "SystemName REMAG regression fe2b12_soc_x\nSystemLabel remag\nNumberOfAtoms 12\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.077 0 0\n0 5.077 0\n0 0 4.254\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.167300000000 0.667300000000 0.000000000000 1\n0.332700000000 0.167300000000 0.000000000000 1\n0.832700000000 0.332700000000 0.000000000000 1\n0.667300000000 0.832700000000 0.000000000000 1\n0.667300000000 0.167300000000 0.500000000000 1\n0.832700000000 0.667300000000 0.500000000000 1\n0.332700000000 0.832700000000 0.500000000000 1\n0.167300000000 0.332700000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 90.0 0.\n2 2.0 90.0 0.\n3 2.0 90.0 0.\n4 2.0 90.0 0.\n5 2.0 90.0 0.\n6 2.0 90.0 0.\n7 2.0 90.0 0.\n8 2.0 90.0 0.\n9 0.0 90.0 0.\n10 0.0 90.0 0.\n11 0.0 90.0 0.\n12 0.0 90.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "1987a287dd32acdf7fe7e190e9363673b31b257ba2d3722a62674d69f3e031ac", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } }, "fe3c16_noncollinear": { "input_fdf": "SystemName REMAG regression fe3c16_noncollinear\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n2 6 C\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.081 0.0 0.0\n0.0 6.754 0.0\n0.0 0.0 4.515\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.184600000000 0.059400000000 0.334000000000 1\n0.184600000000 0.440600000000 0.334000000000 1\n0.815400000000 0.559400000000 0.666000000000 1\n0.315400000000 0.559400000000 0.834000000000 1\n0.684600000000 0.440600000000 0.166000000000 1\n0.684600000000 0.059400000000 0.166000000000 1\n0.315400000000 0.940600000000 0.834000000000 1\n0.815400000000 0.940600000000 0.666000000000 1\n0.034600000000 0.250000000000 0.837700000000 1\n0.965400000000 0.750000000000 0.162300000000 1\n0.465400000000 0.750000000000 0.337700000000 1\n0.534600000000 0.250000000000 0.662300000000 1\n0.898400000000 0.250000000000 0.446700000000 2\n0.101600000000 0.750000000000 0.553300000000 2\n0.601600000000 0.750000000000 0.946700000000 2\n0.398400000000 0.250000000000 0.053300000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin non-collinear\n%block DM.InitSpin\n1 2.5 30.0 45.0\n2 2.5 30.0 45.0\n3 2.5 30.0 45.0\n4 2.5 30.0 45.0\n5 2.5 30.0 45.0\n6 2.5 30.0 45.0\n7 2.5 30.0 45.0\n8 2.5 30.0 45.0\n9 2.5 30.0 45.0\n10 2.5 30.0 45.0\n11 2.5 30.0 45.0\n12 2.5 30.0 45.0\n13 0.0 30.0 45.0\n14 0.0 30.0 45.0\n15 0.0 30.0 45.0\n16 0.0 30.0 45.0\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nPAO.EnergyShift 300 meV\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "d4561abad7720d8b554784fa3a44c5f92e46164651e5422beee89182333bb05e", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" } }, "fept16_soc_z": { "input_fdf": "SystemName REMAG regression fept16_soc_z\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe_fept_SOC\n2 78 Pt_fept_SOC\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.46 0.0 0.0\n0.0 5.46 0.0\n0.0 0.0 7.58\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.365000000000 1.365000000000 1.895000000000 2\n0.000000000000 0.000000000000 3.790000000000 1\n1.365000000000 1.365000000000 5.685000000000 2\n0.000000000000 2.730000000000 0.000000000000 1\n1.365000000000 4.095000000000 1.895000000000 2\n0.000000000000 2.730000000000 3.790000000000 1\n1.365000000000 4.095000000000 5.685000000000 2\n2.730000000000 0.000000000000 0.000000000000 1\n4.095000000000 1.365000000000 1.895000000000 2\n2.730000000000 0.000000000000 3.790000000000 1\n4.095000000000 1.365000000000 5.685000000000 2\n2.730000000000 2.730000000000 0.000000000000 1\n4.095000000000 4.095000000000 1.895000000000 2\n2.730000000000 2.730000000000 3.790000000000 1\n4.095000000000 4.095000000000 5.685000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.5 0.0 0.\n2 0.3 0.0 0.\n3 2.5 0.0 0.\n4 0.3 0.0 0.\n5 2.5 0.0 0.\n6 0.3 0.0 0.\n7 2.5 0.0 0.\n8 0.3 0.0 0.\n9 2.5 0.0 0.\n10 0.3 0.0 0.\n11 2.5 0.0 0.\n12 0.3 0.0 0.\n13 2.5 0.0 0.\n14 0.3 0.0 0.\n15 2.5 0.0 0.\n16 0.3 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "ea290e77b26b088836049ef4ef3376ef8067509fa1cd2a7dc1910b601c28e46b", "pseudopotentials": { "Fe_fept_SOC.psf": "47a51f8ef02f4f64146a55f5445d5c7142489b1dafdeb300d25adc1ccc9f1978", "Pt_fept_SOC.psf": "2be713b0971d72ab0b47d1965b0539e494a20a42a2002a09cd9d5f3eec0d6c47" } }, "fept16_soc_x": { "input_fdf": "SystemName REMAG regression fept16_soc_x\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe_fept_SOC\n2 78 Pt_fept_SOC\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.46 0.0 0.0\n0.0 5.46 0.0\n0.0 0.0 7.58\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.365000000000 1.365000000000 1.895000000000 2\n0.000000000000 0.000000000000 3.790000000000 1\n1.365000000000 1.365000000000 5.685000000000 2\n0.000000000000 2.730000000000 0.000000000000 1\n1.365000000000 4.095000000000 1.895000000000 2\n0.000000000000 2.730000000000 3.790000000000 1\n1.365000000000 4.095000000000 5.685000000000 2\n2.730000000000 0.000000000000 0.000000000000 1\n4.095000000000 1.365000000000 1.895000000000 2\n2.730000000000 0.000000000000 3.790000000000 1\n4.095000000000 1.365000000000 5.685000000000 2\n2.730000000000 2.730000000000 0.000000000000 1\n4.095000000000 4.095000000000 1.895000000000 2\n2.730000000000 2.730000000000 3.790000000000 1\n4.095000000000 4.095000000000 5.685000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.5 90.0 0.\n2 0.3 90.0 0.\n3 2.5 90.0 0.\n4 0.3 90.0 0.\n5 2.5 90.0 0.\n6 0.3 90.0 0.\n7 2.5 90.0 0.\n8 0.3 90.0 0.\n9 2.5 90.0 0.\n10 0.3 90.0 0.\n11 2.5 90.0 0.\n12 0.3 90.0 0.\n13 2.5 90.0 0.\n14 0.3 90.0 0.\n15 2.5 90.0 0.\n16 0.3 90.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "2636385ff78d16342c8d79e6ed814a52b8009ae73b594122d7fbfb620f2fb5a5", "pseudopotentials": { "Fe_fept_SOC.psf": "47a51f8ef02f4f64146a55f5445d5c7142489b1dafdeb300d25adc1ccc9f1978", "Pt_fept_SOC.psf": "2be713b0971d72ab0b47d1965b0539e494a20a42a2002a09cd9d5f3eec0d6c47" } }, "fe5pb2_32_soc_z": { "input_fdf": "SystemName REMAG regression fe5pb2_32_soc_z\nSystemLabel remag\nNumberOfAtoms 32\nNumberOfSpecies 3\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 15 P\n3 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.456 0 0\n0 5.456 0\n0 0 10.296\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.170000000000 0.670000000000 0.139000000000 1\n0.330000000000 0.170000000000 0.139000000000 1\n0.830000000000 0.330000000000 0.139000000000 1\n0.670000000000 0.830000000000 0.139000000000 1\n0.170000000000 0.670000000000 0.861000000000 1\n0.330000000000 0.170000000000 0.861000000000 1\n0.830000000000 0.330000000000 0.861000000000 1\n0.670000000000 0.830000000000 0.861000000000 1\n0.670000000000 0.170000000000 0.639000000000 1\n0.830000000000 0.670000000000 0.639000000000 1\n0.330000000000 0.830000000000 0.639000000000 1\n0.170000000000 0.330000000000 0.639000000000 1\n0.670000000000 0.170000000000 0.361000000000 1\n0.830000000000 0.670000000000 0.361000000000 1\n0.330000000000 0.830000000000 0.361000000000 1\n0.170000000000 0.330000000000 0.361000000000 1\n0.000000000000 0.000000000000 0.000000000000 1\n0.000000000000 0.000000000000 0.500000000000 1\n0.500000000000 0.500000000000 0.000000000000 1\n0.500000000000 0.500000000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n0.381000000000 0.881000000000 0.000000000000 3\n0.119000000000 0.381000000000 0.000000000000 3\n0.619000000000 0.119000000000 0.000000000000 3\n0.881000000000 0.619000000000 0.000000000000 3\n0.881000000000 0.381000000000 0.500000000000 3\n0.619000000000 0.881000000000 0.500000000000 3\n0.119000000000 0.619000000000 0.500000000000 3\n0.381000000000 0.119000000000 0.500000000000 3\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 0.0 0.\n2 2.0 0.0 0.\n3 2.0 0.0 0.\n4 2.0 0.0 0.\n5 2.0 0.0 0.\n6 2.0 0.0 0.\n7 2.0 0.0 0.\n8 2.0 0.0 0.\n9 2.0 0.0 0.\n10 2.0 0.0 0.\n11 2.0 0.0 0.\n12 2.0 0.0 0.\n13 2.0 0.0 0.\n14 2.0 0.0 0.\n15 2.0 0.0 0.\n16 2.0 0.0 0.\n17 2.0 0.0 0.\n18 2.0 0.0 0.\n19 2.0 0.0 0.\n20 2.0 0.0 0.\n21 0.0 0.0 0.\n22 0.0 0.0 0.\n23 0.0 0.0 0.\n24 0.0 0.0 0.\n25 0.0 0.0 0.\n26 0.0 0.0 0.\n27 0.0 0.0 0.\n28 0.0 0.0 0.\n29 0.0 0.0 0.\n30 0.0 0.0 0.\n31 0.0 0.0 0.\n32 0.0 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "fa042284e6a283e00cf013ef12060a1fb760e7354ce7c9189574636561b99123", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "P.psml": "56914dbba0f5c0a625efb4c0f2a1a34c50dff4100ef3dbd0a9454229ebd1c207", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } }, "fe5pb2_32_soc_x": { "input_fdf": "SystemName REMAG regression fe5pb2_32_soc_x\nSystemLabel remag\nNumberOfAtoms 32\nNumberOfSpecies 3\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 15 P\n3 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.456 0 0\n0 5.456 0\n0 0 10.296\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.170000000000 0.670000000000 0.139000000000 1\n0.330000000000 0.170000000000 0.139000000000 1\n0.830000000000 0.330000000000 0.139000000000 1\n0.670000000000 0.830000000000 0.139000000000 1\n0.170000000000 0.670000000000 0.861000000000 1\n0.330000000000 0.170000000000 0.861000000000 1\n0.830000000000 0.330000000000 0.861000000000 1\n0.670000000000 0.830000000000 0.861000000000 1\n0.670000000000 0.170000000000 0.639000000000 1\n0.830000000000 0.670000000000 0.639000000000 1\n0.330000000000 0.830000000000 0.639000000000 1\n0.170000000000 0.330000000000 0.639000000000 1\n0.670000000000 0.170000000000 0.361000000000 1\n0.830000000000 0.670000000000 0.361000000000 1\n0.330000000000 0.830000000000 0.361000000000 1\n0.170000000000 0.330000000000 0.361000000000 1\n0.000000000000 0.000000000000 0.000000000000 1\n0.000000000000 0.000000000000 0.500000000000 1\n0.500000000000 0.500000000000 0.000000000000 1\n0.500000000000 0.500000000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n0.381000000000 0.881000000000 0.000000000000 3\n0.119000000000 0.381000000000 0.000000000000 3\n0.619000000000 0.119000000000 0.000000000000 3\n0.881000000000 0.619000000000 0.000000000000 3\n0.881000000000 0.381000000000 0.500000000000 3\n0.619000000000 0.881000000000 0.500000000000 3\n0.119000000000 0.619000000000 0.500000000000 3\n0.381000000000 0.119000000000 0.500000000000 3\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 90.0 0.\n2 2.0 90.0 0.\n3 2.0 90.0 0.\n4 2.0 90.0 0.\n5 2.0 90.0 0.\n6 2.0 90.0 0.\n7 2.0 90.0 0.\n8 2.0 90.0 0.\n9 2.0 90.0 0.\n10 2.0 90.0 0.\n11 2.0 90.0 0.\n12 2.0 90.0 0.\n13 2.0 90.0 0.\n14 2.0 90.0 0.\n15 2.0 90.0 0.\n16 2.0 90.0 0.\n17 2.0 90.0 0.\n18 2.0 90.0 0.\n19 2.0 90.0 0.\n20 2.0 90.0 0.\n21 0.0 90.0 0.\n22 0.0 90.0 0.\n23 0.0 90.0 0.\n24 0.0 90.0 0.\n25 0.0 90.0 0.\n26 0.0 90.0 0.\n27 0.0 90.0 0.\n28 0.0 90.0 0.\n29 0.0 90.0 0.\n30 0.0 90.0 0.\n31 0.0 90.0 0.\n32 0.0 90.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "76d3a02f41466bcc148615565c74f1ab175171006fb90263dae2e81e2a08dbca", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "P.psml": "56914dbba0f5c0a625efb4c0f2a1a34c50dff4100ef3dbd0a9454229ebd1c207", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } } }, "checks": { "passed": false, "direct_cpu_checks": 21, "all_available_timed_checks_passed": true, "mae": [] }, "failed_trials": [ { "case": "fe3c16_collinear", "label": "screen-elsi2-r16-b8", "config_key": "elsi2-r16-b8", "path": "fe3c16_collinear/screen-elsi2-r16-b8", "passed": false, "valid": true, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [], "checks": { "energy_per_atom_ev": { "passed": true, "max_absolute_error": 6.250002115848474e-08, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-05 }, "total_moment_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": true, "max_absolute_error": 5.960464477539063e-08, "tolerance": 1e-05 }, "deformation_density_components_e_bohr3": { "passed": true, "max_absolute_error": 1.4901161193847656e-08, "tolerance": 1e-05 }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 3.4520051339999998, "candidate_seconds": 38.36134438299996, "reference_iterations": 66, "candidate_iterations": 65 }, "exports": { "passed": true, "checks": { "hsx_header": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_offsets": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_labels": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_counts": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_columns": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_precision": { "passed": true }, "hsx_basis": { "passed": true }, "hsx_kpoints": { "passed": true }, "hsx_charge": { "passed": true }, "hsx_temperature_ry": { "passed": true }, "hsx_h": { "passed": true, "max_absolute_error": 2.1107288139532443e-09, "tolerance": 1e-06 }, "hsx_overlap": { "passed": true, "max_absolute_error": 6.261723750622394e-09, "tolerance": 1e-06 }, "hsx_fermi_ry": { "passed": true, "max_absolute_error": 3.0919367066672976e-10, "tolerance": 1e-06 }, "hsx_positions": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.001 }, "hsx_cell": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-08 }, "drho_cell": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "drho": { "passed": true, "max_absolute_error": 1.4901161193847656e-08, "tolerance": 1e-05 }, "dm_header": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_counts": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_columns": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_values": { "passed": true, "max_absolute_error": 1.397340632092181e-08, "tolerance": 1e-07 } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261010T061359.231.log": "c89bf7db33bf7ef3c9dfcf96145cc9b6e05fde38620b9d5a592cc794a24c1610" } }, "effective_options": { "passed": true, "errors": [], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "2", "ELSI.OutputLevel": "3", "BlockSize": "8" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [ "elpa" ], "ELSI.ELPA.GPU": [ "1" ], "ELSI.ELPA.Flavor": [ "2" ], "ELSI.OutputLevel": [ "3" ], "BlockSize": [ "8" ] }, "runtime_parallelism": { "mpi_ranks": [ 16 ], "openmp_threads": [ 1 ], "combined_processes": [ 16 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261010T061359.231.log": "c89bf7db33bf7ef3c9dfcf96145cc9b6e05fde38620b9d5a592cc794a24c1610" }, "stdout_sha256": "a43165a1a0437a6700815f506bc6747cd22904d9ecf3fa940e8cca4db532d2f8" }, "offload": null, "wall_seconds": 38.36134438299996, "iterations": 65, "cycle_seconds": null, "workspace_bytes": null }, { "case": "fe3c16_collinear", "label": "screen-elsi1-r16-b8", "config_key": "elsi1-r16-b8", "path": "fe3c16_collinear/screen-elsi1-r16-b8", "passed": false, "valid": true, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [], "checks": { "energy_per_atom_ev": { "passed": true, "max_absolute_error": 6.250002115848474e-08, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-05 }, "total_moment_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": true, "max_absolute_error": 5.960464477539063e-08, "tolerance": 1e-05 }, "deformation_density_components_e_bohr3": { "passed": true, "max_absolute_error": 1.4901161193847656e-08, "tolerance": 1e-05 }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 3.4520051339999998, "candidate_seconds": 127.34655580399999, "reference_iterations": 66, "candidate_iterations": 65 }, "exports": { "passed": true, "checks": { "hsx_header": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_offsets": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_labels": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_counts": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_columns": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "hsx_precision": { "passed": true }, "hsx_basis": { "passed": true }, "hsx_kpoints": { "passed": true }, "hsx_charge": { "passed": true }, "hsx_temperature_ry": { "passed": true }, "hsx_h": { "passed": true, "max_absolute_error": 2.1191832733080673e-09, "tolerance": 1e-06 }, "hsx_overlap": { "passed": true, "max_absolute_error": 6.261723750622394e-09, "tolerance": 1e-06 }, "hsx_fermi_ry": { "passed": true, "max_absolute_error": 3.1237201714162666e-10, "tolerance": 1e-06 }, "hsx_positions": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.001 }, "hsx_cell": { "passed": true, "max_absolute_error": 0.0, "tolerance": 1e-08 }, "drho_cell": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "drho": { "passed": true, "max_absolute_error": 1.4901161193847656e-08, "tolerance": 1e-05 }, "dm_header": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_counts": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_columns": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0 }, "dm_values": { "passed": true, "max_absolute_error": 1.3974767485575024e-08, "tolerance": 1e-07 } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261010T061442.196.log": "3622d6ee71115ea98ee35f94b794447fcca0ce6815a8e431bda63e9af85a231b" } }, "effective_options": { "passed": true, "errors": [], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "1", "ELSI.OutputLevel": "3", "BlockSize": "8" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [ "elpa" ], "ELSI.ELPA.GPU": [ "1" ], "ELSI.ELPA.Flavor": [ "1" ], "ELSI.OutputLevel": [ "3" ], "BlockSize": [ "8" ] }, "runtime_parallelism": { "mpi_ranks": [ 16 ], "openmp_threads": [ 1 ], "combined_processes": [ 16 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261010T061442.196.log": "3622d6ee71115ea98ee35f94b794447fcca0ce6815a8e431bda63e9af85a231b" }, "stdout_sha256": "0535bf050b9554224dae4980aaea27d7204eef4962d62a1cedbc3463b8954de7" }, "offload": null, "wall_seconds": 127.34655580399999, "iterations": 65, "cycle_seconds": null, "workspace_bytes": null }, { "case": "fe54_collinear", "label": "screen-elsi2-r8", "config_key": "elsi2-r8", "path": "fe54_collinear/screen-elsi2-r8", "passed": false, "valid": false, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [ "Run failed, did not complete, or did not converge" ], "checks": { "energy_per_atom_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "total_moment_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": false, "max_absolute_error": 0.2429809999999999, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-campaign-v181-20261010T060613Z/fe54_collinear/screen-elsi2-r8/remag.RHO" }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 397.44578809, "candidate_seconds": 202.782551561, "reference_iterations": 89, "candidate_iterations": 200 }, "exports": { "passed": false, "checks": { "export_read": { "passed": false, "reason": "[Errno 2] No such file or directory: '/results/easy-magnetic-campaign-v181-20261010T060613Z/fe54_collinear/screen-elsi2-r8/remag.HSX'" } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261010T062505.494.log": "ff76bf635402a041742312ff1e4e9d3f5a857909f6581af8a2185724ba6d1308" } }, "effective_options": { "passed": true, "errors": [], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "2", "ELSI.OutputLevel": "3", "BlockSize": "64" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [ "elpa" ], "ELSI.ELPA.GPU": [ "1" ], "ELSI.ELPA.Flavor": [ "2" ], "ELSI.OutputLevel": [ "3" ], "BlockSize": [ "64" ] }, "runtime_parallelism": { "mpi_ranks": [ 8 ], "openmp_threads": [ 1 ], "combined_processes": [ 8 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261010T062505.494.log": "ff76bf635402a041742312ff1e4e9d3f5a857909f6581af8a2185724ba6d1308" }, "stdout_sha256": "00e5ebb2a5b2f2ea7049fbc4fbeb05e95a09969f06e1b0a65a152e1543014848" }, "offload": null, "wall_seconds": 202.782551561, "iterations": 200, "cycle_seconds": null, "workspace_bytes": null } ], "sources": [ { "path": "results/easy-magnetic-campaign-v181-20261010T060613Z/summary.json", "sha256": "4e4a68c7d88c3e8ba435a0747cf51dce9b2260fa380f85279a7bddcda44d26f2" }, { "path": "results/easy-magnetic-campaign-v181-20261010T060613Z/workflow-results.json", "sha256": "dbbdd0b59320fd03f07d8dddb15636db32488bf0dd44847613670196ca56ca52" }, { "path": "results/easy-magnetic-campaign-v181-20261010T060613Z/metadata.json", "sha256": "8baa98049e30350f9e7f8d12338a4e60ad8dc44c5eec20cf9c127c6cfb25c6de" }, { "path": "results/easy-magnetic-campaign-v181-20261010T060613Z/upstream-comparison.json", "sha256": "49d2cd31bfcd657ea845e3647f079c2a1c3e490523a295922e42f5ff73cc20d8" }, { "path": "results/easy-magnetic-campaign-v181-20261010T060613Z/ledger.json", "sha256": "7dd5bed870f82b3c01b7b8e7313409ab267ce8edd11427f00ad0d93cd0f8766a" } ], "chart_workflows": [ { "name": "Fe54 collinear", "cases": [ "fe54_collinear" ], "required_families": [ "native", "elsi" ], "screens": [ { "config_key": "elsi2-r8", "family": "elsi", "passed": false, "wall_seconds": 202.782551561 }, { "config_key": "elsi2-r4-b32", "family": "elsi", "passed": true, "wall_seconds": 102.93532000699997 }, { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 124.33507310599998 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 270.22637843300004 } ], "finalists": [ { "family": "native", "config_key": "native2-t4", "config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 131.79213680500015, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 129.12238525899966, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 125.92732839500013, "candidate": 7.414528536000034 } } ], "case_timings": [ { "case": "fe54_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 131.79213680500015, 129.12238525899966, 125.92732839500013 ], "median_seconds": 129.12238525899966, "iterations": [ 74, 73, 71 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 695.0, 695.0, 695.0 ] }, "remag": { "seconds": [ 7.425738153000111, 7.843675305000033, 7.414528536000034 ], "median_seconds": 7.425738153000111, "iterations": [ 107, 127, 114 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3229.0, 3285.0 ] }, "speedup_over_upstream_gpu": 17.388491567916684, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 129.12238525899966, "candidate": 7.425738153000111 }, "remote_provisional_speedup": 17.388491567916684, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 17.388491567916684 }, { "family": "elsi", "config_key": "elsi2-r4-b32", "config": { "route": "elsi-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 104.26391460900004, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 103.34408651400008, "candidate": 7.414528536000034 } } ], "case_timings": [ { "case": "fe54_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 104.23722245900012, 104.26391460900004, 103.34408651400008 ], "median_seconds": 104.23722245900012, "iterations": [ 79, 79, 79 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2378.0, 2370.0, 2378.0 ] }, "remag": { "seconds": [ 7.425738153000111, 7.843675305000033, 7.414528536000034 ], "median_seconds": 7.425738153000111, "iterations": [ 107, 127, 114 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3229.0, 3285.0 ] }, "speedup_over_upstream_gpu": 14.037287648890057, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 }, "remote_provisional_speedup": 14.037287648890057, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 14.037287648890057 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 7.425738153000111 }, { "repeat": 2, "passed": true, "seconds": 7.843675305000033 }, { "repeat": 3, "passed": true, "seconds": 7.414528536000034 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 14.037287648890057, "selected_config": { "route": "elsi-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "selected_config_key": "elsi2-r4-b32", "selected_family": "elsi", "medians": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 }, "pairs": [ { "repeat": 1, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 104.23722245900012, "candidate": 7.425738153000111 } }, { "repeat": 2, "order": [ "candidate", "elsi", "native" ], "passed": true, "seconds": { "upstream": 104.26391460900004, "candidate": 7.843675305000033 } }, { "repeat": 3, "order": [ "native", "elsi", "candidate" ], "passed": true, "seconds": { "upstream": 103.34408651400008, "candidate": 7.414528536000034 } } ], "remote_provisional_speedup": 14.037287648890057, "scientific_qualified": true, "best_verified_finalist_speedup": 14.037287648890057 }, { "name": "Fe3C collinear", "cases": [ "fe3c16_collinear" ], "required_families": [ "native", "elsi" ], "screens": [ { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 144.476083417 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 294.10390538800004 }, { "config_key": "elsi2-r16-b8", "family": "elsi", "passed": false, "wall_seconds": 38.36134438299996 }, { "config_key": "elsi1-r16-b8", "family": "elsi", "passed": false, "wall_seconds": 127.34655580399999 } ], "finalists": [ { "family": "native", "config_key": "native2-t4", "config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 152.31288163600004, "candidate": 2.834449581999934 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 135.462752691, "candidate": 2.839490099000045 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 144.06467930099996, "candidate": 2.9413141740001265 } } ], "case_timings": [ { "case": "fe3c16_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 152.31288163600004, 135.462752691, 144.06467930099996 ], "median_seconds": 144.06467930099996, "iterations": [ 73, 73, 73 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 685.0, 685.0, 685.0 ] }, "remag": { "seconds": [ 2.834449581999934, 2.839490099000045, 2.9413141740001265 ], "median_seconds": 2.839490099000045, "iterations": [ 65, 65, 72 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2049.0, 2049.0, 2049.0 ] }, "speedup_over_upstream_gpu": 50.73610904709045, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 144.06467930099996, "candidate": 2.839490099000045 }, "remote_provisional_speedup": 50.73610904709045, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 50.73610904709045 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 2.834449581999934 }, { "repeat": 2, "passed": true, "seconds": 2.839490099000045 }, { "repeat": 3, "passed": true, "seconds": 2.9413141740001265 } ], "remote_qualified": false, "performance_qualified": false, "speedup_over_upstream_gpu": null, "missing_families": [ "elsi" ], "selected_config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native2-t4", "selected_family": "native", "medians": { "upstream": 144.06467930099996, "candidate": 2.839490099000045 }, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 152.31288163600004, "candidate": 2.834449581999934 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 135.462752691, "candidate": 2.839490099000045 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 144.06467930099996, "candidate": 2.9413141740001265 } } ], "remote_provisional_speedup": 50.73610904709045, "scientific_qualified": false, "best_verified_finalist_speedup": 50.73610904709045 }, { "name": "Fe3C noncollinear", "cases": [ "fe3c16_noncollinear" ], "required_families": [ "native" ], "screens": [ { "config_key": "native2-t4", "family": "native", "passed": true, "wall_seconds": 429.20419752099997 }, { "config_key": "native1-t4", "family": "native", "passed": true, "wall_seconds": 433.5560178679998 }, { "config_key": "native2-r4-b32", "family": "native", "passed": true, "wall_seconds": 370.9022678799997 } ], "finalists": [ { "family": "native", "config_key": "native2-r4-b32", "config": { "route": "native-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "gpu_execution_verified": true, "remote_qualified": true, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 360.35077029400054, "candidate": 3.3475332680000065 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 360.2013342050004, "candidate": 3.2520779189999303 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 366.6469657509997, "candidate": 3.2553646079995815 } } ], "case_timings": [ { "case": "fe3c16_noncollinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 360.35077029400054, 360.2013342050004, 366.6469657509997 ], "median_seconds": 360.35077029400054, "iterations": [ 100, 100, 100 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2182.0, 2182.0, 2180.0 ] }, "remag": { "seconds": [ 3.3475332680000065, 3.2520779189999303, 3.2553646079995815 ], "median_seconds": 3.2553646079995815, "iterations": [ 71, 65, 69 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2199.0, 2199.0, 2199.0 ] }, "speedup_over_upstream_gpu": 110.6944424622948, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 360.35077029400054, "candidate": 3.2553646079995815 }, "remote_provisional_speedup": 110.6944424622948, "scientific_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 110.6944424622948 } ], "candidate_repeats": [ { "repeat": 1, "passed": true, "seconds": 3.3475332680000065 }, { "repeat": 2, "passed": true, "seconds": 3.2520779189999303 }, { "repeat": 3, "passed": true, "seconds": 3.2553646079995815 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 110.6944424622948, "selected_config": { "route": "native-2stage", "block_size": 32, "threads": 1, "mpi_ranks": 4 }, "selected_config_key": "native2-r4-b32", "selected_family": "native", "medians": { "upstream": 360.35077029400054, "candidate": 3.2553646079995815 }, "pairs": [ { "repeat": 1, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 360.35077029400054, "candidate": 3.3475332680000065 } }, { "repeat": 2, "order": [ "candidate", "native" ], "passed": true, "seconds": { "upstream": 360.2013342050004, "candidate": 3.2520779189999303 } }, { "repeat": 3, "order": [ "native", "candidate" ], "passed": true, "seconds": { "upstream": 366.6469657509997, "candidate": 3.2553646079995815 } } ], "remote_provisional_speedup": 110.6944424622948, "scientific_qualified": true, "best_verified_finalist_speedup": 110.6944424622948 } ] }, "previous_suite": { "run": "easy-magnetic-suite-v180-20261009T233637Z", "summary": { "scientific_qualified": true, "performance_qualified": true, "local_cpu_verification_pending": false, "remote_complete": true, "remote_all_workflows_passed": true, "geometric_mean_speedup": 39.99546464961314, "total_suite_speedup": 41.248689713528165, "scope": "Single A100-80GB allocation; current v178 numerical binary, default tile modes, fresh processes and unchanged saved inputs. Results describe this suite, not all SIESTA workloads. CPU references validate science only, not performance.", "qualified_workflows": 4, "declared_workflows": 4, "direct_cpu_comparison": "upstream-comparison.json", "allocation_released": true, "aggregation_scope": "Equal-workflow geometric mean of median speedup ratios; also report ratio of sums of workflow medians and per-case range. x/z counts as one workflow. Full-suite aggregate withheld if any predeclared workflow fails.", "sum_of_workflow_medians_seconds": { "upstream": 1013.4838442439999, "candidate": 24.57008577199997 }, "minimum_workflow_speedup": 21.00900271445142, "maximum_workflow_speedup": 54.60722679583404, "workflows_faster": 4, "baseline_scope": "All ratios are against the selected passing native ELPA GPU configuration, not an exhaustively tuned best upstream GPU route.", "limitations": [ "All ratios are against the selected passing native ELPA GPU configuration, not an exhaustively tuned best upstream GPU route.", "Fe54 ELSI two-stage with 8 ranks failed to converge at the unchanged 200-SCF cap; its failed time is excluded from speedup ratios.", "Fe3C ELSI one-stage and two-stage with 8 ranks failed before SCF with idle-rank / invalid MPI decomposition errors. A suitable ELSI configuration remains unmeasured.", "Four curated previously converged magnetic regression workflows on one A100-SXM4-80GB; no global SIESTA or production-convergence claim." ], "qualified_at_utc": "2026-10-10T01:31:33.492736+00:00" }, "workflows": [ { "name": "Fe54 collinear", "cases": [ "fe54_collinear" ], "screens": [ { "config_key": "native2-t4", "passed": true, "wall_seconds": 222.32551097500001 }, { "config_key": "elsi2-r8", "passed": false, "wall_seconds": 228.669791812 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 21.00900271445142, "selected_config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native2-t4", "pairs": [ { "repeat": 1, "seconds": { "upstream": 168.48339021599998, "candidate": 8.019580582000003 }, "passed": true }, { "repeat": 2, "seconds": { "candidate": 8.83027072799996, "upstream": 169.46631304599998 }, "passed": true }, { "repeat": 3, "seconds": { "upstream": 167.11012155899994, "candidate": 7.7604819359999055 }, "passed": true } ], "case_timings": [ { "case": "fe54_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 168.48339021599998, 169.46631304599998, 167.11012155899994 ], "median_seconds": 168.48339021599998, "iterations": [ 72, 78, 76 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 695.0, 695.0, 695.0 ] }, "remag": { "seconds": [ 8.019580582000003, 8.83027072799996, 7.7604819359999055 ], "median_seconds": 8.019580582000003, "iterations": [ 99, 120, 93 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3285.0, 3285.0 ] }, "speedup_over_upstream_gpu": 21.00900271445142, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 168.48339021599998, "candidate": 8.019580582000003 }, "remote_provisional_speedup": 21.00900271445142, "scientific_qualified": true }, { "name": "Fe16 noncollinear", "cases": [ "fe16_noncollinear" ], "screens": [ { "config_key": "native1-t4", "passed": true, "wall_seconds": 244.23288456 }, { "config_key": "native2-t4", "passed": true, "wall_seconds": 228.9640159129999 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 48.440372642563474, "selected_config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native2-t4", "pairs": [ { "repeat": 1, "seconds": { "upstream": 224.89342920100012, "candidate": 4.889874321999969 }, "passed": true }, { "repeat": 2, "seconds": { "candidate": 4.879863900000146, "upstream": 238.32062738800005 }, "passed": true }, { "repeat": 3, "seconds": { "upstream": 236.38242576100015, "candidate": 4.589188325000123 }, "passed": true } ], "case_timings": [ { "case": "fe16_noncollinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 224.89342920100012, 238.32062738800005, 236.38242576100015 ], "median_seconds": 236.38242576100015, "iterations": [ 83, 83, 83 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 693.0, 693.0, 693.0 ] }, "remag": { "seconds": [ 4.889874321999969, 4.879863900000146, 4.589188325000123 ], "median_seconds": 4.879863900000146, "iterations": [ 76, 76, 76 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2291.0, 2291.0, 2291.0 ] }, "speedup_over_upstream_gpu": 48.440372642563474, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 236.38242576100015, "candidate": 4.879863900000146 }, "remote_provisional_speedup": 48.440372642563474, "scientific_qualified": true }, { "name": "Fe3C collinear", "cases": [ "fe3c16_collinear" ], "screens": [ { "config_key": "native2-t4", "passed": true, "wall_seconds": 144.18449145400018 }, { "config_key": "native1-t4", "passed": true, "wall_seconds": 213.3624469350002 }, { "config_key": "elsi2-r8", "passed": false, "wall_seconds": 1.1189374069999758 }, { "config_key": "elsi1-r8", "passed": false, "wall_seconds": 1.1223090919997958 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 46.044732275349006, "selected_config": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native2-t4", "pairs": [ { "repeat": 1, "seconds": { "upstream": 153.39608893500008, "candidate": 3.2665748730000814 }, "passed": true }, { "repeat": 2, "seconds": { "candidate": 3.5619918669999606, "upstream": 154.24431896000033 }, "passed": true }, { "repeat": 3, "seconds": { "upstream": 179.69235119299992, "candidate": 3.349879808999958 }, "passed": true } ], "case_timings": [ { "case": "fe3c16_collinear", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 153.39608893500008, 154.24431896000033, 179.69235119299992 ], "median_seconds": 154.24431896000033, "iterations": [ 73, 73, 73 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 685.0, 685.0, 685.0 ] }, "remag": { "seconds": [ 3.2665748730000814, 3.5619918669999606, 3.349879808999958 ], "median_seconds": 3.349879808999958, "iterations": [ 63, 65, 71 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2051.0, 2051.0, 2051.0 ] }, "speedup_over_upstream_gpu": 46.044732275349006, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 154.24431896000033, "candidate": 3.349879808999958 }, "remote_provisional_speedup": 46.044732275349006, "scientific_qualified": true }, { "name": "Fe2B full-SOC x/z", "cases": [ "fe2b12_soc_z", "fe2b12_soc_x" ], "screens": [ { "config_key": "native2-t1", "passed": true, "wall_seconds": 525.0334891769999 }, { "config_key": "native2-t4", "passed": true, "wall_seconds": 388.49010499299993 }, { "config_key": "native1-t4", "passed": true, "wall_seconds": 386.9074142600002 } ], "remote_qualified": true, "performance_qualified": true, "speedup_over_upstream_gpu": 54.60722679583404, "selected_config": { "route": "native-1stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "selected_config_key": "native1-t4", "pairs": [ { "repeat": 1, "seconds": { "upstream": 372.56578745799925, "candidate": 8.224499795998781 }, "passed": true }, { "repeat": 2, "seconds": { "candidate": 8.320761480999863, "upstream": 454.37370930699944 }, "passed": true }, { "repeat": 3, "seconds": { "upstream": 462.725297424, "candidate": 8.929390126999351 }, "passed": true } ], "case_timings": [ { "case": "fe2b12_soc_z", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 173.254533327, 187.64450156099974, 203.83354330199927 ], "median_seconds": 187.64450156099974, "iterations": [ 54, 52, 46 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 1153.0, 1137.0, 1073.0 ] }, "remag": { "seconds": [ 3.754743118999613, 3.7520800740003324, 3.8581685629997082 ], "median_seconds": 3.754743118999613, "iterations": [ 44, 44, 44 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2395.0, 2395.0, 2395.0 ] }, "speedup_over_upstream_gpu": 49.97532337471715, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" }, { "case": "fe2b12_soc_x", "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 199.31125413099926, 266.7292077459997, 258.89175412200075 ], "median_seconds": 258.89175412200075, "iterations": [ 65, 62, 63 ], "blas_threads": [ 4, 4, 4 ], "sampled_gpu_memory_peak_mib": [ 1231.0, 1223.0, 1209.0 ] }, "remag": { "seconds": [ 4.469756676999168, 4.568681406999531, 5.071221563999643 ], "median_seconds": 4.568681406999531, "iterations": [ 88, 85, 89 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2395.0, 2395.0, 2395.0 ] }, "speedup_over_upstream_gpu": 56.66662458129844, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "medians": { "upstream": 454.37370930699944, "candidate": 8.320761480999863 }, "remote_provisional_speedup": 54.60722679583404, "scientific_qualified": true } ], "hardware": [ "NVIDIA A100-SXM4-80GB, GPU-4e39a966-3ced-1584-45ca-530b319102ed, 610.57.04" ], "cpu_cores_requested": 8, "candidate_settings": { "REMAG_NONLOCAL_H0_POISON_HOST": "0", "REMAG_NONLOCAL_H0": "1", "REMAG_NONLOCAL_H0_SHADOW": "0", "OMP_NUM_THREADS": "1", "OPENBLAS_NUM_THREADS": "1", "MKL_NUM_THREADS": "1", "REMAG_GRID": "1", "REMAG_DM": "1", "REMAG_GRID_EPOCH": "1", "REMAG_GRID_VERIFY_EPOCH": "0", "REMAG_NORMALIZE": "1", "REMAG_OCCUPATIONS": "1", "REMAG_RESIDENT_SPECTRA": "1", "REMAG_HS": "1", "REMAG_HARTREE": "1", "REMAG_RESIDENT_GRID": "1", "REMAG_RESIDENT_SCF": "1", "REMAG_DEVICE_SCF": "1", "REMAG_DEVICE_VECTOR_CACHE": "1", "REMAG_BLOCK_SOLVER": "0", "REMAG_BLOCK_WIDTH": "32", "REMAG_BLOCK_LOCAL_SOLVER": "jacobi", "REMAG_FORCE_SHADOW": "0", "REMAG_FORCE_SHADOW_RESIDENT": "0", "REMAG_GRID_FORCE": "1", "REMAG_ATOMIC_GRID_FORCE": "1", "REMAG_ATOMIC_FORCE_SHADOW": "0", "REMAG_FINAL_GRID": "1", "REMAG_FINAL_GRID_SHADOW": "0", "REMAG_FINAL_DENSITY": "1", "REMAG_FINAL_DENSITY_SHADOW": "0", "REMAG_ORBITAL_BASIS_SHADOW": "0", "REMAG_ORBITAL_BASIS": "1", "REMAG_ATOMIC_BASIS_SHADOW": "0", "REMAG_ATOMIC_BASIS": "1", "REMAG_TWO_CENTER_SHADOW": "0", "REMAG_PAIR_FORCE_SHADOW": "0", "REMAG_NONLOCAL_FORCE_SHADOW": "0", "REMAG_INTEGRAL_FORCE": "1", "REMAG_POPULATION": "1", "REMAG_TRACE_IO": "0", "REMAG_TRACE_CUDA": "0", "REMAG_GPU_INITIALIZATION": "1", "REMAG_GPU_HISTORY": "1", "REMAG_FINAL_SPECTRA": "1", "REMAG_GATE_H_EXPORT": "1", "REMAG_DEFER_SCF_LOG": "1", "GFORTRAN_FORMATTED_BUFFER_SIZE": "1048576", "GFORTRAN_UNBUFFERED_ALL": "0", "GFORTRAN_UNBUFFERED_PRECONNECTED": "0", "REMAG_BLOCK_SHARED_JACOBI": "1", "REMAG_INITIAL_MESH": "1", "REMAG_DENSITY_UNIQUE_PAIRS": "1", "REMAG_SCF_SOLVER_LANES": "8", "REMAG_SCF_WARM_START": "1", "REMAG_RADIAL_BATCH": "1", "REMAG_GRID_POTENTIAL_SPINS": "1", "REMAG_SCF_STAGE_PROFILE": "0", "REMAG_HOST_ID": "4fb5335dfc144d188938b38bfcdfd7e9", "LD_LIBRARY_PATH": "/opt/remag-candidate/remag-native:/opt/elpa-gpu/lib:/usr/local/cuda/lib64", "REMAG_BLOCK_FUSED_MATRIX": "0", "REMAG_BLOCK_FP64_VECTORS": "0", "REMAG_BLOCK_FP64_MATRIX": "0", "OMPI_ALLOW_RUN_AS_ROOT": "1", "OMPI_ALLOW_RUN_AS_ROOT_CONFIRM": "1" }, "upstream_settings": { "LD_LIBRARY_PATH": "/opt/elpa-gpu/lib:/usr/local/cuda/lib64", "REMAG_TRACE_CUDA": "0", "REMAG_TRACE_IO": "0", "REMAG_SCF_STAGE_PROFILE": "0", "REMAG_HOST_ID": "4fb5335dfc144d188938b38bfcdfd7e9", "OMPI_ALLOW_RUN_AS_ROOT": "1", "OMPI_ALLOW_RUN_AS_ROOT_CONFIRM": "1" }, "binaries": { "/opt/remag-candidate/Src/siesta": "e9c444827228c6b680dc5f9a62bf54f422a0220e5a6add27766ecbc0a2da40d1", "/opt/remag-candidate/remag-native/libremag_gpu.so": "479381a3d20051a2540c12c823d2143bee091cd53c51d8ee4bfb2068dcaa50b6", "/opt/upstream-gpu/Src/siesta": "59781a7c103181e30ca77cced4c5c1256c101c5fbbb88033263e8e5a83d44c66" }, "plan": { "created_at_utc": "2026-10-09T23:35:06.799436+00:00", "suite": "Four previously converged magnetic workflows; selected before new timing results", "selection": "Existing pristine CPU and candidate convergence with unchanged physics and no case-specific recovery: CPU 47\u201389 iterations. This does not assert universal convergence.", "candidate_run": "fp64-matrix-qualification-v178-20261009T215709Z", "candidate_image": "im-1U2ZnxmQQlDqw1VEGqVh5x", "workflows": [ { "name": "Fe54 collinear", "cases": [ "fe54_collinear" ], "configs": [ "native2-t4", "elsi2-r8" ] }, { "name": "Fe16 noncollinear", "cases": [ "fe16_noncollinear" ], "configs": [ "native1-t4", "native2-t4" ] }, { "name": "Fe3C collinear", "cases": [ "fe3c16_collinear" ], "configs": [ "native2-t4", "native1-t4", "elsi2-r8", "elsi1-r8" ] }, { "name": "Fe2B full-SOC x/z", "cases": [ "fe2b12_soc_z", "fe2b12_soc_x" ], "configs": [ "native2-t1", "native2-t4", "native1-t4" ] } ], "cases": { "fe54_collinear": { "remote_reference_run": "scf-solver-lanes-regression-v116b-20261008T174638Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/scf-solver-lanes-regression-v116b-20261008T174638Z/fe54_collinear/cpu", "input_sha256": "705a76444e50a1096898f71b64f769c5575128c4ab93acce486f2149784e1005", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" }, "prior_cpu_iterations": 89, "atoms": 54, "spin": "polarized" }, "fe16_noncollinear": { "remote_reference_run": "scf-seeded-regression-v119b-20261008T191317Z", "remote_reference_role": "cpu", "local_cpu_reference": "results/scf-seeded-regression-v119b-20261008T191317Z/fe16_noncollinear/cpu", "input_sha256": "3602afc93642d3aaf393530d0c440b3dc65b57af72360e4a390fdcfe28f732ec", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" }, "prior_cpu_iterations": 77, "atoms": 16, "spin": "non-collinear" }, "fe3c16_collinear": { "remote_reference_run": "carbide-gpu-v169-20261009T170851Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/carbide-local-reference-v168-20261009T164908Z/fe3c16_collinear/cpu", "input_sha256": "833c95fbd300f7ee745f87027f613379f47fbcff7d5c12f8abd930b0c1e42cbf", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" }, "prior_cpu_iterations": 68, "atoms": 16, "spin": "polarized" }, "fe2b12_soc_z": { "remote_reference_run": "boride-gpu-v173-20261009T202117Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/boride-local-reference-v171-20261009T172616Z/fe2b12_soc_z/cpu", "input_sha256": "4ed2ed5b40f6cd2fc689061fcefbfa31fd6ed935e3ff8544ffb870a307935ddf", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 47, "atoms": 12, "spin": "spin-orbit" }, "fe2b12_soc_x": { "remote_reference_run": "boride-gpu-v173-20261009T202117Z", "remote_reference_role": "candidate", "local_cpu_reference": "results/boride-local-reference-v171-20261009T172616Z/fe2b12_soc_x/cpu", "input_sha256": "1987a287dd32acdf7fe7e190e9363673b31b257ba2d3722a62674d69f3e031ac", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" }, "prior_cpu_iterations": 75, "atoms": 12, "spin": "spin-orbit" } }, "configs": { "native2-t4": { "route": "native-2stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "native1-t4": { "route": "native-1stage", "block_size": 64, "threads": 4, "mpi_ranks": 1 }, "native2-t1": { "route": "native-2stage", "block_size": 64, "threads": 1, "mpi_ranks": 1 }, "elsi2-r8": { "route": "elsi-2stage", "block_size": 64, "threads": 1, "mpi_ranks": 8 }, "elsi1-r8": { "route": "elsi-1stage", "block_size": 64, "threads": 1, "mpi_ranks": 8 } }, "repetitions": 3, "candidate_tile_flags": { "REMAG_BLOCK_FUSED_MATRIX": "0", "REMAG_BLOCK_FP64_VECTORS": "0", "REMAG_BLOCK_FP64_MATRIX": "0" }, "trial_timeout_seconds": 600, "allocation_timeout_seconds": 7200, "gpu_limit": 1, "convergence": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "baseline_selection": "Screen established native/ELSI settings sequentially; choose lowest complete workflow wall time among scientifically passing settings; then three fresh repetitions. No claim of exhaustive tuning.", "aggregate": "Equal-workflow geometric mean of median speedup ratios; also report ratio of sums of workflow medians and per-case range. x/z counts as one workflow. Full-suite aggregate withheld if any predeclared workflow fails.", "scope": "Single A100-80GB allocation; current v178 numerical binary, default tile modes, fresh processes and unchanged saved inputs. Results describe this suite, not all SIESTA workloads. CPU references validate science only, not performance." }, "inputs": { "fe54_collinear": { "input_fdf": "SystemName REMAG regression fe54_collinear\nSystemLabel remag\nNumberOfAtoms 54\nNumberOfSpecies 1\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n8.61 0.0 0.0\n0.0 8.61 0.0\n0.0 0.0 8.61\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.435000000000 1.435000000000 1.435000000000 1\n0.000000000000 0.000000000000 2.870000000000 1\n1.435000000000 1.435000000000 4.305000000000 1\n0.000000000000 0.000000000000 5.740000000000 1\n1.435000000000 1.435000000000 7.175000000000 1\n0.000000000000 2.870000000000 0.000000000000 1\n1.435000000000 4.305000000000 1.435000000000 1\n0.000000000000 2.870000000000 2.870000000000 1\n1.435000000000 4.305000000000 4.305000000000 1\n0.000000000000 2.870000000000 5.740000000000 1\n1.435000000000 4.305000000000 7.175000000000 1\n0.000000000000 5.740000000000 0.000000000000 1\n1.435000000000 7.175000000000 1.435000000000 1\n0.000000000000 5.740000000000 2.870000000000 1\n1.435000000000 7.175000000000 4.305000000000 1\n0.000000000000 5.740000000000 5.740000000000 1\n1.435000000000 7.175000000000 7.175000000000 1\n2.870000000000 0.000000000000 0.000000000000 1\n4.305000000000 1.435000000000 1.435000000000 1\n2.870000000000 0.000000000000 2.870000000000 1\n4.305000000000 1.435000000000 4.305000000000 1\n2.870000000000 0.000000000000 5.740000000000 1\n4.305000000000 1.435000000000 7.175000000000 1\n2.870000000000 2.870000000000 0.000000000000 1\n4.305000000000 4.305000000000 1.435000000000 1\n2.870000000000 2.870000000000 2.870000000000 1\n4.305000000000 4.305000000000 4.305000000000 1\n2.870000000000 2.870000000000 5.740000000000 1\n4.305000000000 4.305000000000 7.175000000000 1\n2.870000000000 5.740000000000 0.000000000000 1\n4.305000000000 7.175000000000 1.435000000000 1\n2.870000000000 5.740000000000 2.870000000000 1\n4.305000000000 7.175000000000 4.305000000000 1\n2.870000000000 5.740000000000 5.740000000000 1\n4.305000000000 7.175000000000 7.175000000000 1\n5.740000000000 0.000000000000 0.000000000000 1\n7.175000000000 1.435000000000 1.435000000000 1\n5.740000000000 0.000000000000 2.870000000000 1\n7.175000000000 1.435000000000 4.305000000000 1\n5.740000000000 0.000000000000 5.740000000000 1\n7.175000000000 1.435000000000 7.175000000000 1\n5.740000000000 2.870000000000 0.000000000000 1\n7.175000000000 4.305000000000 1.435000000000 1\n5.740000000000 2.870000000000 2.870000000000 1\n7.175000000000 4.305000000000 4.305000000000 1\n5.740000000000 2.870000000000 5.740000000000 1\n7.175000000000 4.305000000000 7.175000000000 1\n5.740000000000 5.740000000000 0.000000000000 1\n7.175000000000 7.175000000000 1.435000000000 1\n5.740000000000 5.740000000000 2.870000000000 1\n7.175000000000 7.175000000000 4.305000000000 1\n5.740000000000 5.740000000000 5.740000000000 1\n7.175000000000 7.175000000000 7.175000000000 1\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n1 0 0 0.0\n0 1 0 0.0\n0 0 1 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin polarized\n%block DM.InitSpin\n1 2.5 0.0 0.\n2 2.5 0.0 0.\n3 2.5 0.0 0.\n4 2.5 0.0 0.\n5 2.5 0.0 0.\n6 2.5 0.0 0.\n7 2.5 0.0 0.\n8 2.5 0.0 0.\n9 2.5 0.0 0.\n10 2.5 0.0 0.\n11 2.5 0.0 0.\n12 2.5 0.0 0.\n13 2.5 0.0 0.\n14 2.5 0.0 0.\n15 2.5 0.0 0.\n16 2.5 0.0 0.\n17 2.5 0.0 0.\n18 2.5 0.0 0.\n19 2.5 0.0 0.\n20 2.5 0.0 0.\n21 2.5 0.0 0.\n22 2.5 0.0 0.\n23 2.5 0.0 0.\n24 2.5 0.0 0.\n25 2.5 0.0 0.\n26 2.5 0.0 0.\n27 2.5 0.0 0.\n28 2.5 0.0 0.\n29 2.5 0.0 0.\n30 2.5 0.0 0.\n31 2.5 0.0 0.\n32 2.5 0.0 0.\n33 2.5 0.0 0.\n34 2.5 0.0 0.\n35 2.5 0.0 0.\n36 2.5 0.0 0.\n37 2.5 0.0 0.\n38 2.5 0.0 0.\n39 2.5 0.0 0.\n40 2.5 0.0 0.\n41 2.5 0.0 0.\n42 2.5 0.0 0.\n43 2.5 0.0 0.\n44 2.5 0.0 0.\n45 2.5 0.0 0.\n46 2.5 0.0 0.\n47 2.5 0.0 0.\n48 2.5 0.0 0.\n49 2.5 0.0 0.\n50 2.5 0.0 0.\n51 2.5 0.0 0.\n52 2.5 0.0 0.\n53 2.5 0.0 0.\n54 2.5 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "705a76444e50a1096898f71b64f769c5575128c4ab93acce486f2149784e1005", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" } }, "fe16_noncollinear": { "input_fdf": "SystemName REMAG regression fe16_noncollinear\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 1\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.74 0.0 0.0\n0.0 5.74 0.0\n0.0 0.0 5.74\n%endblock LatticeVectors\nAtomicCoordinatesFormat NotScaledCartesianAng\n%block AtomicCoordinatesAndAtomicSpecies\n0.000000000000 0.000000000000 0.000000000000 1\n1.435000000000 1.435000000000 1.435000000000 1\n0.000000000000 0.000000000000 2.870000000000 1\n1.435000000000 1.435000000000 4.305000000000 1\n0.000000000000 2.870000000000 0.000000000000 1\n1.435000000000 4.305000000000 1.435000000000 1\n0.000000000000 2.870000000000 2.870000000000 1\n1.435000000000 4.305000000000 4.305000000000 1\n2.870000000000 0.000000000000 0.000000000000 1\n4.305000000000 1.435000000000 1.435000000000 1\n2.870000000000 0.000000000000 2.870000000000 1\n4.305000000000 1.435000000000 4.305000000000 1\n2.870000000000 2.870000000000 0.000000000000 1\n4.305000000000 4.305000000000 1.435000000000 1\n2.870000000000 2.870000000000 2.870000000000 1\n4.305000000000 4.305000000000 4.305000000000 1\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin non-collinear\n%block DM.InitSpin\n1 2.5 30.0 0.\n2 2.5 30.0 0.\n3 2.5 30.0 0.\n4 2.5 30.0 0.\n5 2.5 30.0 0.\n6 2.5 30.0 0.\n7 2.5 30.0 0.\n8 2.5 30.0 0.\n9 2.5 30.0 0.\n10 2.5 30.0 0.\n11 2.5 30.0 0.\n12 2.5 30.0 0.\n13 2.5 30.0 0.\n14 2.5 30.0 0.\n15 2.5 30.0 0.\n16 2.5 30.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "3602afc93642d3aaf393530d0c440b3dc65b57af72360e4a390fdcfe28f732ec", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895" } }, "fe3c16_collinear": { "input_fdf": "SystemName REMAG regression fe3c16_collinear\nSystemLabel remag\nNumberOfAtoms 16\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe_nc\n2 6 C\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.081 0.0 0.0\n0.0 6.754 0.0\n0.0 0.0 4.515\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.184600000000 0.059400000000 0.334000000000 1\n0.184600000000 0.440600000000 0.334000000000 1\n0.815400000000 0.559400000000 0.666000000000 1\n0.315400000000 0.559400000000 0.834000000000 1\n0.684600000000 0.440600000000 0.166000000000 1\n0.684600000000 0.059400000000 0.166000000000 1\n0.315400000000 0.940600000000 0.834000000000 1\n0.815400000000 0.940600000000 0.666000000000 1\n0.034600000000 0.250000000000 0.837700000000 1\n0.965400000000 0.750000000000 0.162300000000 1\n0.465400000000 0.750000000000 0.337700000000 1\n0.534600000000 0.250000000000 0.662300000000 1\n0.898400000000 0.250000000000 0.446700000000 2\n0.101600000000 0.750000000000 0.553300000000 2\n0.601600000000 0.750000000000 0.946700000000 2\n0.398400000000 0.250000000000 0.053300000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin polarized\n%block DM.InitSpin\n1 2.5 0.0 0.0\n2 2.5 0.0 0.0\n3 2.5 0.0 0.0\n4 2.5 0.0 0.0\n5 2.5 0.0 0.0\n6 2.5 0.0 0.0\n7 2.5 0.0 0.0\n8 2.5 0.0 0.0\n9 2.5 0.0 0.0\n10 2.5 0.0 0.0\n11 2.5 0.0 0.0\n12 2.5 0.0 0.0\n13 0.0 0.0 0.0\n14 0.0 0.0 0.0\n15 0.0 0.0 0.0\n16 0.0 0.0 0.0\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nPAO.EnergyShift 300 meV\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 300 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 6\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "833c95fbd300f7ee745f87027f613379f47fbcff7d5c12f8abd930b0c1e42cbf", "pseudopotentials": { "Fe_nc.psf": "43829da4a9189b60bb7ebb43c0fbdb3de49519ab6c6c48487fb50c6f7f031895", "C.psf": "835287a894d851d30c2c613edcc4124c493a08dad74a2aa1a982266c15d1a0e6" } }, "fe2b12_soc_z": { "input_fdf": "SystemName REMAG regression fe2b12_soc_z\nSystemLabel remag\nNumberOfAtoms 12\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.077 0 0\n0 5.077 0\n0 0 4.254\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.167300000000 0.667300000000 0.000000000000 1\n0.332700000000 0.167300000000 0.000000000000 1\n0.832700000000 0.332700000000 0.000000000000 1\n0.667300000000 0.832700000000 0.000000000000 1\n0.667300000000 0.167300000000 0.500000000000 1\n0.832700000000 0.667300000000 0.500000000000 1\n0.332700000000 0.832700000000 0.500000000000 1\n0.167300000000 0.332700000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 0.0 0.\n2 2.0 0.0 0.\n3 2.0 0.0 0.\n4 2.0 0.0 0.\n5 2.0 0.0 0.\n6 2.0 0.0 0.\n7 2.0 0.0 0.\n8 2.0 0.0 0.\n9 0.0 0.0 0.\n10 0.0 0.0 0.\n11 0.0 0.0 0.\n12 0.0 0.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "4ed2ed5b40f6cd2fc689061fcefbfa31fd6ed935e3ff8544ffb870a307935ddf", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } }, "fe2b12_soc_x": { "input_fdf": "SystemName REMAG regression fe2b12_soc_x\nSystemLabel remag\nNumberOfAtoms 12\nNumberOfSpecies 2\n%block ChemicalSpeciesLabel\n1 26 Fe\n2 5 B\n%endblock ChemicalSpeciesLabel\nLatticeConstant 1.0 Ang\n%block LatticeVectors\n5.077 0 0\n0 5.077 0\n0 0 4.254\n%endblock LatticeVectors\nAtomicCoordinatesFormat Fractional\n%block AtomicCoordinatesAndAtomicSpecies\n0.167300000000 0.667300000000 0.000000000000 1\n0.332700000000 0.167300000000 0.000000000000 1\n0.832700000000 0.332700000000 0.000000000000 1\n0.667300000000 0.832700000000 0.000000000000 1\n0.667300000000 0.167300000000 0.500000000000 1\n0.832700000000 0.667300000000 0.500000000000 1\n0.332700000000 0.832700000000 0.500000000000 1\n0.167300000000 0.332700000000 0.500000000000 1\n0.000000000000 0.000000000000 0.250000000000 2\n0.000000000000 0.000000000000 0.750000000000 2\n0.500000000000 0.500000000000 0.250000000000 2\n0.500000000000 0.500000000000 0.750000000000 2\n%endblock AtomicCoordinatesAndAtomicSpecies\n%block kgrid_Monkhorst_Pack\n2 0 0 0.0\n0 2 0 0.0\n0 0 2 0.0\n%endblock kgrid_Monkhorst_Pack\nSpin spin-orbit\n%block DM.InitSpin\n1 2.0 90.0 0.\n2 2.0 90.0 0.\n3 2.0 90.0 0.\n4 2.0 90.0 0.\n5 2.0 90.0 0.\n6 2.0 90.0 0.\n7 2.0 90.0 0.\n8 2.0 90.0 0.\n9 0.0 90.0 0.\n10 0.0 90.0 0.\n11 0.0 90.0 0.\n12 0.0 90.0 0.\n%endblock DM.InitSpin\nPAO.BasisSize SZP\nMeshCutoff 100 Ry\nXC.functional GGA\nXC.authors PBE\nElectronicTemperature 1000 K\nSolutionMethod diagon\nDiag.Algorithm divide-and-conquer\nMaxSCFIterations 200\nSCF.MustConverge true\nDM.Tolerance 1e-9\nSCF.H.Tolerance 1e-8 eV\nDM.MixingWeight 0.05\nDM.NumberPulay 8\nSCF.Mix.Spin all\nDM.UseSaveDM false\nWriteForces true\nWriteMullikenPop 1\nSaveRho true\nSaveDeltaRho true\nSaveHS true\nUseTreeTimer true\nUseParallelTimer false\nUseWallTime true\nTiming.Split.SCF.Steps false\nWrite.DM false\nWrite.DM.End.Of.Cycle true\nTS.DE.Save false\nCDF.Save false\nCharge.Mulliken init+geometry+end\nCharge.Mulliken.Format 1\nCharge.Spin init+scf+geometry+end\n", "sha256": "1987a287dd32acdf7fe7e190e9363673b31b257ba2d3722a62674d69f3e031ac", "pseudopotentials": { "Fe.psml": "959e0d8bd121540933877209affcb5787f15cafcb7ac05dcc984f85580acdd9f", "B.psml": "53ca4a54f08f4de4fad41b706067bcee662f849d245dbf5f04be92334cefdcef" } } }, "checks": { "passed": true, "direct_cpu_checks": 30, "all_available_timed_checks_passed": true, "mae": [ { "workflow": "Fe2B full-SOC x/z", "role": "upstream", "repeat": 1, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" }, { "workflow": "Fe2B full-SOC x/z", "role": "upstream", "repeat": 2, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" }, { "workflow": "Fe2B full-SOC x/z", "role": "upstream", "repeat": 3, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" }, { "workflow": "Fe2B full-SOC x/z", "role": "candidate", "repeat": 1, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" }, { "workflow": "Fe2B full-SOC x/z", "role": "candidate", "repeat": 2, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" }, { "workflow": "Fe2B full-SOC x/z", "role": "candidate", "repeat": 3, "passed": true, "max_absolute_error": 0.0, "tolerance": 5e-06, "reference_mae_ev": -0.0016219999997701962, "candidate_mae_ev": -0.0016219999997701962, "convention": "E(x)-E(z)" } ] }, "failed_trials": [ { "case": "fe54_collinear", "label": "screen-elsi2-r8", "config_key": "elsi2-r8", "path": "fe54_collinear/screen-elsi2-r8", "passed": false, "valid": false, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [ "Run failed, did not complete, or did not converge" ], "checks": { "energy_per_atom_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "total_moment_muB": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": false, "max_absolute_error": 0.2429809999999999, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": true, "max_absolute_error": 0.0, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-suite-v180-20261009T233637Z/fe54_collinear/screen-elsi2-r8/remag.RHO" }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 397.44578809, "candidate_seconds": 228.669791812, "reference_iterations": 89, "candidate_iterations": 200 }, "exports": { "passed": false, "checks": { "export_read": { "passed": false, "reason": "[Errno 2] No such file or directory: '/results/easy-magnetic-suite-v180-20261009T233637Z/fe54_collinear/screen-elsi2-r8/remag.HSX'" } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261009T234058.539.log": "ff76bf635402a041742312ff1e4e9d3f5a857909f6581af8a2185724ba6d1308" } }, "effective_options": { "passed": true, "errors": [], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "2", "ELSI.OutputLevel": "3", "BlockSize": "64" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [ "elpa" ], "ELSI.ELPA.GPU": [ "1" ], "ELSI.ELPA.Flavor": [ "2" ], "ELSI.OutputLevel": [ "3" ], "BlockSize": [ "64" ] }, "runtime_parallelism": { "mpi_ranks": [ 8 ], "openmp_threads": [ 1 ], "combined_processes": [ 8 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261009T234058.539.log": "ff76bf635402a041742312ff1e4e9d3f5a857909f6581af8a2185724ba6d1308" }, "stdout_sha256": "9afd34bf21dda0f2aad5dac9c6f83b62972601a368e779f5277aa9b37001708c" }, "offload": null, "wall_seconds": 228.669791812, "iterations": 200, "cycle_seconds": null, "workspace_bytes": null }, { "case": "fe3c16_collinear", "label": "screen-elsi2-r8", "config_key": "elsi2-r8", "path": "fe3c16_collinear/screen-elsi2-r8", "passed": false, "valid": false, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [ "Run failed, did not complete, or did not converge" ], "checks": { "energy_per_atom_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "total_moment_muB": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi2-r8/remag.RHO" }, "deformation_density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi2-r8/remag.DRHO" }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 3.4520051339999998, "candidate_seconds": 1.1189374069999758, "reference_iterations": 66, "candidate_iterations": 0 }, "exports": { "passed": false, "checks": { "export_read": { "passed": false, "reason": "[Errno 2] No such file or directory: '/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi2-r8/remag.HSX'" } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261010T002126.037.log": "477ec748c3e5fddfdfd4c4ff39e3b87d4d3f667ccfe69847a9fdb9a6e8d2d085" } }, "effective_options": { "passed": false, "errors": [ "Missing, defaulted or inconsistent effective option: ELSI.Solver", "Missing, defaulted or inconsistent effective option: ELSI.ELPA.GPU", "Missing, defaulted or inconsistent effective option: ELSI.ELPA.Flavor", "Missing, defaulted or inconsistent effective option: ELSI.OutputLevel", "Missing, defaulted or inconsistent effective option: BlockSize" ], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "2", "ELSI.OutputLevel": "3", "BlockSize": "64" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [], "ELSI.ELPA.GPU": [], "ELSI.ELPA.Flavor": [], "ELSI.OutputLevel": [], "BlockSize": [] }, "runtime_parallelism": { "mpi_ranks": [ 8 ], "openmp_threads": [ 1 ], "combined_processes": [ 8 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261010T002126.037.log": "477ec748c3e5fddfdfd4c4ff39e3b87d4d3f667ccfe69847a9fdb9a6e8d2d085" }, "stdout_sha256": "3d27580a4c0d7a2e0c7a926733abb8c236f130f6c9484318ff5ac390a108e830" }, "offload": null, "wall_seconds": 1.1189374069999758, "iterations": 0, "cycle_seconds": null, "workspace_bytes": null }, { "case": "fe3c16_collinear", "label": "screen-elsi1-r8", "config_key": "elsi1-r8", "path": "fe3c16_collinear/screen-elsi1-r8", "passed": false, "valid": false, "diagnostic_only": false, "scientific": { "passed": false, "reasons": [ "Run failed, did not complete, or did not converge" ], "checks": { "energy_per_atom_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "fermi_energy_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "forces_ev_ang": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "stress_ev_ang3": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 1e-05 }, "total_moment_muB": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "atom_moments_muB": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "atom_net_charges_e": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "eigenvalues_ev": { "passed": false, "reason": "Required observable missing", "max_absolute_error": null, "tolerance": 0.0001 }, "density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi1-r8/remag.RHO" }, "deformation_density_components_e_bohr3": { "passed": false, "reason": "/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi1-r8/remag.DRHO" }, "tsde_not_requested": { "passed": true }, "netcdf_not_requested": { "passed": true } }, "end_to_end_speedup": null, "performance_comparable": false, "performance_reasons": [ "Hardware or runtime allocation differs" ], "reference_seconds": 3.4520051339999998, "candidate_seconds": 1.1223090919997958, "reference_iterations": 66, "candidate_iterations": 0 }, "exports": { "passed": false, "checks": { "export_read": { "passed": false, "reason": "[Errno 2] No such file or directory: '/results/easy-magnetic-suite-v180-20261009T233637Z/fe3c16_collinear/screen-elsi1-r8/remag.HSX'" } } }, "convergence": { "passed": true, "errors": [], "observed": { "DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.DM.Tolerance": [ [ "0.1000000000E-08" ] ], "SCF.H.Tolerance": [ [ "0.7349864435E-09", "Ry" ] ], "SCF.DM.Converge": [ [ "T" ] ], "SCF.H.Converge": [ [ "T" ] ] }, "expected_tolerances": { "DM.Tolerance": 1e-09, "SCF.DM.Tolerance": 1e-09, "SCF.H.Tolerance": 1e-08 }, "evidence_sha256": { "fdf.20261010T002129.371.log": "477ec748c3e5fddfdfd4c4ff39e3b87d4d3f667ccfe69847a9fdb9a6e8d2d085" } }, "effective_options": { "passed": false, "errors": [ "Missing, defaulted or inconsistent effective option: ELSI.Solver", "Missing, defaulted or inconsistent effective option: ELSI.ELPA.GPU", "Missing, defaulted or inconsistent effective option: ELSI.ELPA.Flavor", "Missing, defaulted or inconsistent effective option: ELSI.OutputLevel", "Missing, defaulted or inconsistent effective option: BlockSize" ], "expected": { "SolutionMethod": "ELSI", "ELSI.Solver": "ELPA", "ELSI.ELPA.GPU": "1", "ELSI.ELPA.Flavor": "1", "ELSI.OutputLevel": "3", "BlockSize": "64" }, "observed": { "SolutionMethod": [ "elsi" ], "ELSI.Solver": [], "ELSI.ELPA.GPU": [], "ELSI.ELPA.Flavor": [], "ELSI.OutputLevel": [], "BlockSize": [] }, "runtime_parallelism": { "mpi_ranks": [ 8 ], "openmp_threads": [ 1 ], "combined_processes": [ 8 ], "expected_threads": 1 }, "actual_gpu_execution_verified": false, "fdf_log_sha256": { "fdf.20261010T002129.371.log": "477ec748c3e5fddfdfd4c4ff39e3b87d4d3f667ccfe69847a9fdb9a6e8d2d085" }, "stdout_sha256": "178af92d671c890d2eafe91cd4ac08afd971311fd1f4543b6e0ab4a201047899" }, "offload": null, "wall_seconds": 1.1223090919997958, "iterations": 0, "cycle_seconds": null, "workspace_bytes": null } ], "sources": [ { "path": "results/easy-magnetic-suite-v180-20261009T233637Z/summary.json", "sha256": "55c0425d6b4042decd66a29282a84e1bf9ab3f725dd1073f461f2369afa5e1eb" }, { "path": "results/easy-magnetic-suite-v180-20261009T233637Z/workflow-results.json", "sha256": "d4b5bcf013fcb3d4f8eb3195e988347b3d69711f616c29635115b791a61c109e" }, { "path": "results/easy-magnetic-suite-v180-20261009T233637Z/metadata.json", "sha256": "5046e53861a7ee80f4c2b5fcf2e8f4f1aa2983a8402fcf09d5aa5f78f771c1b9" }, { "path": "results/easy-magnetic-suite-v180-20261009T233637Z/upstream-comparison.json", "sha256": "d304ee7ce090883b9d5b310f0b9577808336497a6df67292568d8621de5be9e9" }, { "path": "results/easy-magnetic-suite-v180-20261009T233637Z/ledger.json", "sha256": "3862f837a2854c7662c1b58b86f2545a2d00f2e94046c8d1c8a91dad45eabd69" } ] }, "incremental_comparisons": [ { "change": "Eight solver workspaces", "run": "scf-solver-lanes-v116-20261008T173132Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/scf-solver-lanes-v116-20261008T173132Z/summary.json", "sha256": "796229cd0b492a952f4234dbf9a0c5b4e24856c00f37b5fd8c226de8515c48b6" }, "gpu": [ "NVIDIA A100 80GB PCIe, GPU-7e35ee3c-3004-93d7-370c-a1c98ac24daf" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 89.32561377400003, "candidate": 24.436990805999983 }, "iterations": { "accepted": { "fept16_soc_z": 66, "fept16_soc_x": 60 }, "candidate": { "fept16_soc_z": 60, "fept16_soc_x": 57 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 25.548513635999996, "accepted": 91.79026674200003 }, "iterations": { "accepted": { "fept16_soc_z": 70, "fept16_soc_x": 60 }, "candidate": { "fept16_soc_z": 68, "fept16_soc_x": 57 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 84.19261077500005, "candidate": 23.729646129000002 }, "iterations": { "accepted": { "fept16_soc_z": 63, "fept16_soc_x": 55 }, "candidate": { "fept16_soc_z": 58, "fept16_soc_x": 54 } } } ], "median_pair_seconds": { "accepted": 89.32561377400003, "candidate": 24.436990805999983 }, "incremental_speedup": 3.6553442477077835, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." }, { "change": "Reuse the previous eigenbasis", "run": "scf-seeded-eigensolver-v119-20261008T190153Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/scf-seeded-eigensolver-v119-20261008T190153Z/summary.json", "sha256": "c86d5f25c5f577eedc36bbdbe9b7e095493487fa5c9080bfe3d1aa1d5415938f" }, "gpu": [ "NVIDIA A100-SXM4-80GB, GPU-4cc9fe19-0b7f-cb57-1d47-b1166336f8f6" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 28.741767529000008, "candidate": 16.61437615 }, "iterations": { "accepted": { "fept16_soc_z": 88, "fept16_soc_x": 61 }, "candidate": { "fept16_soc_z": 57, "fept16_soc_x": 71 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 16.415854770999985, "accepted": 25.73200926399997 }, "iterations": { "accepted": { "fept16_soc_z": 67, "fept16_soc_x": 61 }, "candidate": { "fept16_soc_z": 63, "fept16_soc_x": 60 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 24.502486832000017, "candidate": 16.397587982999966 }, "iterations": { "accepted": { "fept16_soc_z": 58, "fept16_soc_x": 59 }, "candidate": { "fept16_soc_z": 67, "fept16_soc_x": 61 } } } ], "median_pair_seconds": { "accepted": 25.73200926399997, "candidate": 16.415854770999985 }, "incremental_speedup": 1.5675095584701304, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." }, { "change": "Share density work across grid clusters", "run": "grid-cluster-density-v120-20261008T193307Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/grid-cluster-density-v120-20261008T193307Z/summary.json", "sha256": "6533ad4c3d41af062a63b96626b9840a5e859623667f57647d85ce091c7e857a" }, "gpu": [ "NVIDIA A100 80GB PCIe, GPU-b83cf95e-fbba-7a86-c69c-5c4d3764ca35" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 18.525981955000002, "candidate": 16.42298690300001 }, "iterations": { "accepted": { "fept16_soc_z": 78, "fept16_soc_x": 64 }, "candidate": { "fept16_soc_z": 74, "fept16_soc_x": 65 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 16.754018367, "accepted": 18.141055383999984 }, "iterations": { "accepted": { "fept16_soc_z": 71, "fept16_soc_x": 63 }, "candidate": { "fept16_soc_z": 75, "fept16_soc_x": 68 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 18.359591746000007, "candidate": 16.68191127100002 }, "iterations": { "accepted": { "fept16_soc_z": 75, "fept16_soc_x": 66 }, "candidate": { "fept16_soc_z": 78, "fept16_soc_x": 68 } } } ], "median_pair_seconds": { "accepted": 18.359591746000007, "candidate": 16.68191127100002 }, "incremental_speedup": 1.100568840569035, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." }, { "change": "Contract unique orbital pairs", "run": "unique-density-pairs-v140-20261009T091807Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/unique-density-pairs-v140-20261009T091807Z/summary.json", "sha256": "97409b7401add66be67d8f6a0f1d6aa00ce98d76c15c79aca90f60c5d9e6a1ad" }, "gpu": [ "NVIDIA A100-SXM4-80GB, GPU-9879642f-604f-0cdd-76c7-fa5c2bff716d" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 11.264639922, "candidate": 10.468355120000005 }, "iterations": { "accepted": { "fept16_soc_z": 79, "fept16_soc_x": 75 }, "candidate": { "fept16_soc_z": 75, "fept16_soc_x": 95 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 10.756390878000005, "accepted": 11.365360855000006 }, "iterations": { "accepted": { "fept16_soc_z": 75, "fept16_soc_x": 83 }, "candidate": { "fept16_soc_z": 78, "fept16_soc_x": 99 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 12.06630226899999, "candidate": 10.644179659999992 }, "iterations": { "accepted": { "fept16_soc_z": 75, "fept16_soc_x": 109 }, "candidate": { "fept16_soc_z": 75, "fept16_soc_x": 88 } } } ], "median_pair_seconds": { "accepted": 11.365360855000006, "candidate": 10.644179659999992 }, "incremental_speedup": 1.06775357219027, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." }, { "change": "Generate initial mesh fields on GPU", "run": "initial-mesh-integration-v141b-20261009T094041Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/initial-mesh-integration-v141b-20261009T094041Z/summary.json", "sha256": "a5d27a98cd8d3f5c83f42f741806e4f04b2c3c7aea437adbf9ad81d980254209" }, "gpu": [ "NVIDIA A100-SXM4-80GB, GPU-25c4985f-0762-6dd7-56e7-f191fcd4b585" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 10.958768618999997, "candidate": 10.144988468999998 }, "iterations": { "accepted": { "fept16_soc_z": 82, "fept16_soc_x": 71 }, "candidate": { "fept16_soc_z": 80, "fept16_soc_x": 73 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 10.554089947999998, "accepted": 11.178392633999977 }, "iterations": { "accepted": { "fept16_soc_z": 85, "fept16_soc_x": 73 }, "candidate": { "fept16_soc_z": 88, "fept16_soc_x": 77 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 11.064519689999997, "candidate": 10.254100898000019 }, "iterations": { "accepted": { "fept16_soc_z": 79, "fept16_soc_x": 73 }, "candidate": { "fept16_soc_z": 84, "fept16_soc_x": 70 } } } ], "median_pair_seconds": { "accepted": 11.064519689999997, "candidate": 10.254100898000019 }, "incremental_speedup": 1.0790336276248311, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." }, { "change": "Shared-memory local eigensolver", "run": "shared-local-eigensolver-v142-20261009T095832Z", "workload": "FePt16 full-SOC x/z pair", "source": { "path": "results/shared-local-eigensolver-v142-20261009T095832Z/summary.json", "sha256": "f554db3aa0be7ef10a7d16f09e4c2eb5a169b0a780473b1a8425c62477a424f8" }, "gpu": [ "NVIDIA A100-SXM4-80GB, GPU-cb32c6de-2452-c141-70a1-dbdd0115300c" ], "pairs": [ { "repeat": 1, "passed": true, "seconds": { "accepted": 11.511638806000008, "candidate": 11.236511054000005 }, "iterations": { "accepted": { "fept16_soc_z": 75, "fept16_soc_x": 91 }, "candidate": { "fept16_soc_z": 75, "fept16_soc_x": 97 } } }, { "repeat": 2, "passed": true, "seconds": { "candidate": 9.643438170999985, "accepted": 11.06192907700003 }, "iterations": { "accepted": { "fept16_soc_z": 80, "fept16_soc_x": 105 }, "candidate": { "fept16_soc_z": 80, "fept16_soc_x": 73 } } }, { "repeat": 3, "passed": true, "seconds": { "accepted": 11.057816454999966, "candidate": 9.630004498000005 }, "iterations": { "accepted": { "fept16_soc_z": 75, "fept16_soc_x": 102 }, "candidate": { "fept16_soc_z": 80, "fept16_soc_x": 73 } } } ], "median_pair_seconds": { "accepted": 11.06192907700003, "candidate": 9.643438170999985 }, "incremental_speedup": 1.1470938975132097, "scope": "Matched end-to-end comparison with the preceding REMAG implementation; not an upstream-GPU ratio." } ], "fp64_matrix_rotation": { "run": "fp64-matrix-qualification-v178-20261009T215709Z", "source": { "path": "results/fp64-matrix-qualification-v178-20261009T215709Z/production-integration.json", "sha256": "3642d476ff757056e09c0017ca0ef92c23f51916b531550d895accb47874b0fc" }, "statistics": { "complete": true, "medians": { "accepted": 203.27527914899997, "candidate": 190.832317299 }, "means": { "accepted": 203.47249893733334, "candidate": 190.92423687300004 }, "faster_pairs": 3, "performance_improved": true, "incremental_speedup": 1.0652036406941707 }, "pairs": [ { "repeat": 1, "seconds": { "accepted": 204.13548844200005, "candidate": 190.832317299 } }, { "repeat": 2, "seconds": { "candidate": 189.67925770600004, "accepted": 203.006729221 } }, { "repeat": 3, "seconds": { "accepted": 203.27527914899997, "candidate": 192.26113561400007 } } ], "small_statistics": { "complete": true, "medians": { "accepted": 11.011168807999994, "candidate": 11.02457983299999 }, "means": { "accepted": 11.029370896666654, "candidate": 11.021230029333339 }, "faster_pairs": 2, "provisional_incremental_speedup": 0.9987835341388837, "performance_improved": false }, "scope": "FePt128 SOC z; matched incremental REMAG comparison. Both contenders enable FP64 vector rotation; only the candidate enables FP64 matrix rotation." }, "earlier_fept16_upstream_gpu": { "run": "upstream-gpu-refresh-v143-20261009T101214Z", "source": { "path": "results/upstream-gpu-refresh-v143-20261009T101214Z/qualification.json", "sha256": "604a204ee5176c42781f87fe1e38cac102784ff6c62cfbc1a529ab557733078d" }, "measured_build": "v142", "directional_pair": { "config": { "block_size": 64, "route": "native-2stage", "threads": 1 }, "cases": [ { "case": "fept16_soc_z", "config": { "block_size": 64, "route": "native-2stage", "threads": 1 }, "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 320.618512494, 321.50422765200005, 319.3178219720003 ], "median_seconds": 320.618512494, "iterations": [ 106, 106, 106 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 557.0, 557.0, 557.0 ] }, "remag": { "seconds": [ 5.082426274999989, 4.967734279000069, 4.964099434999753 ], "median_seconds": 4.967734279000069, "iterations": [ 82, 85, 89 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2545.0, 2545.0, 2545.0 ] }, "speedup_over_upstream_gpu": 64.5401896493014, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" }, { "case": "fept16_soc_x", "config": { "block_size": 64, "route": "native-2stage", "threads": 1 }, "performance_qualified": true, "errors": [], "upstream_gpu": { "seconds": [ 257.697900987, 251.80690825200008, 253.02751403599996 ], "median_seconds": 253.02751403599996, "iterations": [ 82, 82, 82 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 557.0, 557.0, 557.0 ] }, "remag": { "seconds": [ 4.971673142999862, 4.776544880000074, 4.873999238000124 ], "median_seconds": 4.873999238000124, "iterations": [ 81, 79, 80 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 2545.0, 2545.0, 2545.0 ] }, "speedup_over_upstream_gpu": 51.913736888440916, "scope": "Process start through initialization, SCF, final observables and requested exports; excludes build, provisioning and evidence transfer", "tuning": "Repeated timing for this qualified configuration; final selection uses a complete directional pair" } ], "upstream_pair_seconds": [ 578.316413481, 573.3111359040001, 572.3453360080002 ], "remag_pair_seconds": [ 10.05409941799985, 9.744279159000143, 9.838098672999877 ], "upstream_median_pair_seconds": 573.3111359040001, "remag_median_pair_seconds": 9.838098672999877, "speedup_over_upstream_gpu": 58.27458688510831, "scope": "Two separately initialized directional calculations per repetition; one identical solver configuration for x and z" }, "scope": "Historical standalone comparison; excluded from the v180 suite average." }, "earlier_fe54_elsi_gpu": { "run": "spin-layout-performance-v127-20261008T220559Z-interrupted-c7b6eddcebea4576b95627dfced200ef", "source": { "path": "results/spin-layout-performance-v127-20261008T220559Z-interrupted-c7b6eddcebea4576b95627dfced200ef/summary.json", "sha256": "8b27c8ccff02bd3b71821364ac85b8cf82ddcf531647d7e371e91b8ac7daabba" }, "hardware": [ "NVIDIA A100 80GB PCIe, GPU-cca4805e-4cb0-a491-413a-3db96c0e0013" ], "input_sha256": "705a76444e50a1096898f71b64f769c5575128c4ab93acce486f2149784e1005", "config": { "route": "elsi-2stage", "block_size": 64, "threads": 1, "mpi_ranks": 8 }, "upstream_gpu": { "seconds": [ 79.62677819800001, 82.91530027600015, 79.10352472900013 ], "median_seconds": 79.62677819800001, "iterations": [ 71, 71, 71 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 4722.0, 4722.0, 4722.0 ] }, "remag": { "seconds": [ 9.221504380000056, 10.308567315000118, 9.511502805000191 ], "median_seconds": 9.511502805000191, "iterations": [ 95, 117, 96 ], "blas_threads": [ 1, 1, 1 ], "sampled_gpu_memory_peak_mib": [ 3229.0, 3281.0, 3229.0 ] }, "speedup_over_upstream_gpu": 8.371629576363082, "scope": "Completed Fe54 segment of a subsequently interrupted v127 attempt. Same physical input as v180, earlier REMAG build, separate A100 PCIe allocation. ELSI passed here; this historical result is not replaced by v180 native-only timing and is excluded from the v180 aggregate." }, "notes": [ "All timing data comes from existing completed calculations. No benchmark was added to create this article.", "SCF iteration counts may differ; timing measures time to equivalent converged solutions.", "Individual incremental ratios are never multiplied into an upstream-GPU comparison.", "The current campaign is incomplete: no full-suite aggregate. Only qualified finalists are charted.", "Fe54 native and ELSI comparisons share the same three REMAG repetitions.", "The earlier four-workflow native-only aggregate is historical and separate from the current campaign.", "Failed baseline screens remain in each suite record; full raw outputs remain in the separate results archive." ] }