Skip to content

A100 CUDA run

GridSim  -  2026-07-02 14:55:10 UTC
code:    git 8ec809d2-dirty  |  GridSim.Core v0.1.0.0  |  build Release
runtime: .NET 10.0.9  X64  GC=Server
machine: keen-darwin  Ubuntu 24.04.3 LTS
cpu:     AMD EPYC 9554 64-Core Processor  -  28 logical cores  -  118 GB RAM
backend: gpu/cuda (NVIDIA A100 80GB PCIe [CUDA] - FP64 is fast (data-centre GPU) (driver probe: cuInit=0 driverVer=12080 cuDeviceGetCount rc=0 count=1))
run:     bench-matrix  cases=25  engines=5  runs=5  parallel=1  note="A100 80GB PCIe R570 + EPYC 28c: parallel=1 (uncontended per-case latency)"

Run note: A100 80GB PCIe R570 + EPYC 28c: parallel=1 (uncontended per-case latency)

Full solver x backend matrix over the whole archive. Every engine below is swept across every case (Q-limits off, tol 1e-8, max 100 iters). Cells: conv / iters / mismatch / mean ms. BFS is radial-only (meshed cases show n/a). Dense solves are memory-gated (few concurrent) and ordered small->large; GPU engines run one case at a time and skip cases >3000 buses (consumer-GPU FP64 is slow). Absent accelerators are simply not listed as engines.

Engines this run: cpu-dense - Dense Newton-Raphson, CPU LU; cpu-sparse - Sparse Newton-Raphson, CPU (direct LU / ILU0-BiCGSTAB); bfs - Backward-forward sweep (radial feeders only); cuda-dense - Dense NR, CUDA (gpu/cuda (NVIDIA A100 80GB PCIe [CUDA] - FP64 is fast (data-centre GPU) (driver probe: cuInit=0 driverVer=12080 cuDeviceGetCount rc=0 count=1))); cuda-sparse - Sparse NR, CUDA (gpu/cuda (NVIDIA A100 80GB PCIe [CUDA] - FP64 is fast (data-centre GPU) (driver probe: cuInit=0 driverVer=12080 cuDeviceGetCount rc=0 count=1))) ROCm/AMD: --gpu requested but no device answered.

cpu-dense - Dense Newton-Raphson, CPU LU

Case Buses Gens Branches conv / iters / mismatch / mean ms
pglib_opf_case3_lmbd 3 3 3 NO / 100 / 8.27E+000 / 1ms
pglib_opf_case5_pjm 5 5 6 yes / 3 / 3.57E-011 / 0ms
case9 (WSCC) 9 3 9 yes / 4 / 1.84E-014 / 0ms
case14 14 5 20 yes / 2 / 1.32E-010 / 0ms
cigre_mv 15 1 14 yes / 5 / 2.66E-009 / 1ms
case18 18 1 17 yes / 4 / 5.97E-014 / 1ms
case22 22 1 21 NO / 100 / 2.62E+007 / 15ms
case24_ieee_rts 24 33 38 yes / 4 / 2.31E-012 / 0ms
case30 30 6 41 yes / 3 / 9.57E-010 / 1ms
pglib_opf_case30_as 30 6 41 yes / 4 / 1.97E-014 / 1ms
case33bw 33 1 32 NO / 100 / 1.35E+003 / 24ms
case57 57 7 80 yes / 3 / 3.45E-012 / 3ms
case69 69 1 68 NO / 100 / 3.16E+003 / 99ms
pglib_opf_case73_ieee_rts 73 99 120 yes / 5 / 2.40E-014 / 3ms
case85 85 1 84 NO / 100 / 1.13E+006 / 142ms
case118 118 54 186 yes / 3 / 1.51E-012 / 5ms
case145 145 50 453 yes / 3 / 7.87E-012 / 10ms
pglib_opf_case162_ieee_dtc 162 12 284 NO / 100 / 2.21E+004 / 574ms
pglib_opf_case240_pserc 240 143 448 NO / 100 / 1.35E+004 / 1012ms
case300 300 69 411 yes / 5 / 1.38E-012 / 78ms
case533mt_hi 532 1 514 ERR:Singular matrix in LU solve.
pglib_opf_case588_sdet 588 95 686 yes / 4 / 1.47E-011 / 245ms
case1354pegase 1354 260 1991 yes / 4 / 4.10E-012 / 2281ms (x3)
case2869pegase 2869 510 4582 yes / 6 / 2.15E-009 / 46695ms (x3)
case9241pegase 9241 1445 16049 ERR:count ('-1973164928') must be a non-negative value. (Parameter 'count') Actual value was -1973164928.

cpu-sparse - Sparse Newton-Raphson, CPU (direct LU / ILU0-BiCGSTAB)

Case Buses Gens Branches conv / iters / mismatch / mean ms
pglib_opf_case3_lmbd 3 3 3 NO / 100 / 6.67E+000 / 2ms
pglib_opf_case5_pjm 5 5 6 yes / 3 / 3.57E-011 / 0ms
case9 (WSCC) 9 3 9 yes / 4 / 1.68E-014 / 0ms
case14 14 5 20 yes / 2 / 1.32E-010 / 0ms
cigre_mv 15 1 14 yes / 5 / 9.45E-011 / 1ms
case18 18 1 17 yes / 4 / 5.97E-014 / 1ms
case22 22 1 21 NO / 100 / 8.21E+001 / 13ms
case24_ieee_rts 24 33 38 yes / 4 / 6.52E-012 / 1ms
case30 30 6 41 yes / 3 / 9.57E-010 / 1ms
pglib_opf_case30_as 30 6 41 yes / 4 / 8.93E-015 / 1ms
case33bw 33 1 32 NO / 100 / 6.00E+001 / 12ms
case57 57 7 80 yes / 3 / 3.45E-012 / 2ms
case69 69 1 68 NO / 100 / 1.24E+002 / 34ms
pglib_opf_case73_ieee_rts 73 99 120 yes / 7 / 2.81E-014 / 2ms
case85 85 1 84 NO / 100 / 1.14E+002 / 12ms
case118 118 54 186 yes / 3 / 1.51E-012 / 1ms
case145 145 50 453 yes / 3 / 7.88E-012 / 2ms
pglib_opf_case162_ieee_dtc 162 12 284 NO / 100 / 3.37E+000 / 77ms
pglib_opf_case240_pserc 240 143 448 NO / 100 / 1.02E+002 / 105ms
case300 300 69 411 yes / 5 / 1.38E-012 / 5ms
case533mt_hi 532 1 514 NO / 0 / 2.76E-002 / 202ms
pglib_opf_case588_sdet 588 95 686 yes / 5 / 7.27E-010 / 9ms
case1354pegase 1354 260 1991 yes / 4 / 1.45E-012 / 17ms (x3)
case2869pegase 2869 510 4582 yes / 6 / 2.15E-009 / 68ms (x3)
case9241pegase 9241 1445 16049 yes / 6 / 2.13E-009 / 405ms (x1)

bfs - Backward-forward sweep (radial feeders only)

Case Buses Gens Branches conv / iters / mismatch / mean ms
pglib_opf_case3_lmbd 3 3 3 n/a (meshed)
pglib_opf_case5_pjm 5 5 6 n/a (meshed)
case9 (WSCC) 9 3 9 n/a (meshed)
case14 14 5 20 n/a (meshed)
cigre_mv 15 1 14 yes / 11 / 6.15E-010 / 1ms
case18 18 1 17 yes / 12 / 1.11E-009 / 0ms
case22 22 1 21 NO / 100 / 9.96E+002 / 2ms
case24_ieee_rts 24 33 38 n/a (meshed)
case30 30 6 41 n/a (meshed)
pglib_opf_case30_as 30 6 41 n/a (meshed)
case33bw 33 1 32 NO / 100 / 1.01E+002 / 2ms
case57 57 7 80 n/a (meshed)
case69 69 1 68 NO / 100 / 8.51E+002 / 1ms
pglib_opf_case73_ieee_rts 73 99 120 n/a (meshed)
case85 85 1 84 NO / 100 / 2.37E+003 / 1ms
case118 118 54 186 n/a (meshed)
case145 145 50 453 n/a (meshed)
pglib_opf_case162_ieee_dtc 162 12 284 n/a (meshed)
pglib_opf_case240_pserc 240 143 448 n/a (meshed)
case300 300 69 411 n/a (meshed)
case533mt_hi 532 1 514 n/a (meshed)
pglib_opf_case588_sdet 588 95 686 n/a (meshed)
case1354pegase 1354 260 1991 n/a (meshed)
case2869pegase 2869 510 4582 n/a (meshed)
case9241pegase 9241 1445 16049 n/a (meshed)

cuda-dense - Dense NR, CUDA (gpu/cuda (NVIDIA A100 80GB PCIe [CUDA] - FP64 is fast (data-centre GPU) (driver probe: cuInit=0 driverVer=12080 cuDeviceGetCount rc=0 count=1)))

Case Buses Gens Branches conv / iters / mismatch / mean ms
pglib_opf_case3_lmbd 3 3 3 NO / 100 / 6.71E+000 / 201ms (x3)
pglib_opf_case5_pjm 5 5 6 yes / 3 / 3.57E-011 / 1ms (x3)
case9 (WSCC) 9 3 9 yes / 4 / 1.84E-014 / 2ms (x3)
case14 14 5 20 yes / 2 / 1.32E-010 / 1ms (x3)
cigre_mv 15 1 14 yes / 5 / 2.66E-009 / 4ms (x3)
case18 18 1 17 yes / 4 / 5.27E-014 / 11ms (x3)
case22 22 1 21 NO / 100 / 1.94E+006 / 112ms (x3)
case24_ieee_rts 24 33 38 yes / 4 / 2.31E-012 / 4ms (x3)
case30 30 6 41 yes / 3 / 9.57E-010 / 4ms (x3)
pglib_opf_case30_as 30 6 41 yes / 4 / 7.32E-015 / 5ms (x3)
case33bw 33 1 32 NO / 100 / 9.12E+002 / 189ms (x3)
case57 57 7 80 yes / 3 / 3.45E-012 / 10ms (x3)
case69 69 1 68 NO / 100 / 9.38E+002 / 493ms (x3)
pglib_opf_case73_ieee_rts 73 99 120 yes / 5 / 2.09E-014 / 17ms (x3)
case85 85 1 84 NO / 100 / 2.66E+005 / 668ms (x3)
case118 118 54 186 yes / 3 / 1.51E-012 / 19ms (x3)
case145 145 50 453 yes / 3 / 7.87E-012 / 37ms (x3)
pglib_opf_case162_ieee_dtc 162 12 284 NO / 100 / 7.66E+003 / 1993ms (x3)
pglib_opf_case240_pserc 240 143 448 NO / 100 / 4.26E+005 / 4144ms (x3)
case300 300 69 411 yes / 5 / 1.38E-012 / 341ms (x3)
case533mt_hi 532 1 514 ERR:Singular matrix in GPU LU solve.
pglib_opf_case588_sdet 588 95 686 yes / 4 / 1.47E-011 / 911ms (x3)
case1354pegase 1354 260 1991 yes / 4 / 1.87E-012 / 7808ms (x3)
case2869pegase 2869 510 4582 yes / 6 / 2.15E-009 / 54886ms (x3)
case9241pegase 9241 1445 16049 n/a (>3000 buses, skipped)

cuda-sparse - Sparse NR, CUDA (gpu/cuda (NVIDIA A100 80GB PCIe [CUDA] - FP64 is fast (data-centre GPU) (driver probe: cuInit=0 driverVer=12080 cuDeviceGetCount rc=0 count=1)))

Case Buses Gens Branches conv / iters / mismatch / mean ms
pglib_opf_case3_lmbd 3 3 3 NO / 100 / 6.67E+000 / 91ms (x3)
pglib_opf_case5_pjm 5 5 6 yes / 3 / 3.57E-011 / 3ms (x3)
case9 (WSCC) 9 3 9 yes / 4 / 1.84E-014 / 9ms (x3)
case14 14 5 20 yes / 2 / 1.32E-010 / 6ms (x3)
cigre_mv 15 1 14 yes / 5 / 9.45E-011 / 32ms (x3)
case18 18 1 17 yes / 4 / 1.19E-013 / 31ms (x3)
case22 22 1 21 NO / 100 / 8.21E+001 / 3736ms (x3)
case24_ieee_rts 24 33 38 yes / 4 / 6.52E-012 / 16ms (x3)
case30 30 6 41 yes / 3 / 9.57E-010 / 20ms (x3)
pglib_opf_case30_as 30 6 41 yes / 4 / 1.61E-014 / 26ms (x3)
case33bw 33 1 32 NO / 100 / 6.00E+001 / 3589ms (x3)
case57 57 7 80 yes / 3 / 3.45E-012 / 65ms (x3)
case69 69 1 68 NO / 100 / 1.24E+002 / 8701ms (x3)
pglib_opf_case73_ieee_rts 73 99 120 yes / 7 / 2.84E-014 / 58ms (x3)
case85 85 1 84 NO / 100 / 1.14E+002 / 7707ms (x3)
case118 118 54 186 yes / 3 / 1.50E-012 / 38ms (x3)
case145 145 50 453 yes / 3 / 7.87E-012 / 45ms (x3)
pglib_opf_case162_ieee_dtc 162 12 284 NO / 100 / 3.37E+000 / 1761ms (x3)
pglib_opf_case240_pserc 240 143 448 NO / 100 / 9.07E+001 / 13202ms (x3)
case300 300 69 411 yes / 5 / 1.38E-012 / 179ms (x3)
case533mt_hi 532 1 514 NO / 0 / 2.76E-002 / 178ms (x3)
pglib_opf_case588_sdet 588 95 686 yes / 5 / 7.27E-010 / 1567ms (x3)
case1354pegase 1354 260 1991 yes / 4 / 3.73E-012 / 923ms (x3)
case2869pegase 2869 510 4582 yes / 6 / 2.15E-009 / 4851ms (x3)
case9241pegase 9241 1445 16049 n/a (>3000 buses, skipped)

Comparison - mean solve time (ms), converged unless noted

Case Buses cpu-dense cpu-sparse bfs cuda-dense cuda-sparse
pglib_opf_case3_lmbd 3 NO NO n/a NO NO
pglib_opf_case5_pjm 5 0ms 0ms n/a 1ms 3ms
case9 (WSCC) 9 0ms 0ms n/a 2ms 9ms
case14 14 0ms 0ms n/a 1ms 6ms
cigre_mv 15 1ms 1ms 1ms 4ms 32ms
case18 18 1ms 1ms 0ms 11ms 31ms
case22 22 NO NO NO NO NO
case24_ieee_rts 24 0ms 1ms n/a 4ms 16ms
case30 30 1ms 1ms n/a 4ms 20ms
pglib_opf_case30_as 30 1ms 1ms n/a 5ms 26ms
case33bw 33 NO NO NO NO NO
case57 57 3ms 2ms n/a 10ms 65ms
case69 69 NO NO NO NO NO
pglib_opf_case73_ieee_rts 73 3ms 2ms n/a 17ms 58ms
case85 85 NO NO NO NO NO
case118 118 5ms 1ms n/a 19ms 38ms
case145 145 10ms 2ms n/a 37ms 45ms
pglib_opf_case162_ieee_dtc 162 NO NO n/a NO NO
pglib_opf_case240_pserc 240 NO NO n/a NO NO
case300 300 78ms 5ms n/a 341ms 179ms
case533mt_hi 532 ERR:Singular matrix in LU solve. NO n/a ERR:Singular matrix in GPU LU solve. NO
pglib_opf_case588_sdet 588 245ms 9ms n/a 911ms 1567ms
case1354pegase 1354 2281ms 17ms n/a 7808ms 923ms
case2869pegase 2869 46695ms 68ms n/a 54886ms 4851ms
case9241pegase 9241 ERR:count ('-1973164928') must be a non-negative value. (Parameter 'count') Actual value was -1973164928. 405ms n/a skip skip