cd $BASILISK/examples
CFLAGS='-DSINGLE_PRECISION -DBENCHMARK=1' make turbulence.ctst
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=8 turbulence/turbulence $i 2> /tmp/log;
done
# Title: Inteli7
# Multigrid, 2008 steps, 13.5656 CPU, 1.698 real, 1.94e+07 points.step/s, 8 var
# Multigrid, 2073 steps, 37.6603 CPU, 4.71 real, 2.88e+07 points.step/s, 8 var
# Multigrid, 2363 steps, 125.084 CPU, 15.64 real, 3.96e+07 points.step/s, 8 var
# Multigrid, 2887 steps, 566.771 CPU, 70.87 real, 4.27e+07 points.step/s, 8 var
# Multigrid, 3010 steps, 2484.88 CPU, 310.7 real, 4.06e+07 points.step/s, 8 var
# Multigrid, 3110 steps, 10083.9 CPU, 1261 real, 4.14e+07 points.step/s, 8 var
CFLAGS='-DBENCHMARK=1' make turbulence.gpu.tst
OpenGL renderer string: Mesa Intel(R) UHD Graphics (TGL GT1) (0x9a60)
Video memory: 3072MB
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu $i 2> /tmp/log;
done
# Title: IntelHUD
# Multigrid (GPU), 2009 steps, 2.59921 CPU, 5.478 real, 6.01e+06 points.step/s, 8 var
# Multigrid (GPU), 2008 steps, 1.992 CPU, 8.546 real, 1.54e+07 points.step/s, 8 var
# Multigrid (GPU), 2328 steps, 4.43931 CPU, 25.36 real, 2.41e+07 points.step/s, 8 var
# Multigrid (GPU), 2767 steps, 10.0743 CPU, 97.95 real, 2.96e+07 points.step/s, 8 var
# Multigrid (GPU), 2991 steps, 17.2899 CPU, 391.4 real, 3.21e+07 points.step/s, 8 var
OpenGL renderer string: NVIDIA GeForce RTX 3050 Ti Laptop GPU/PCIe/SSE2
Dedicated video memory: 4096 MB
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 nvidia turbulence.gpu/turbulence.gpu $i 2> /tmp/log;
done
# Title: RTX3050
# Multigrid (GPU), 2009 steps, 1.53716 CPU, 1.537 real, 2.14e+07 points.step/s, 8 var
# Multigrid (GPU), 2008 steps, 1.63853 CPU, 1.639 real, 8.03e+07 points.step/s, 8 var
# Multigrid (GPU), 2328 steps, 3.30781 CPU, 3.308 real, 1.84e+08 points.step/s, 8 var
# Multigrid (GPU), 2767 steps, 11.4863 CPU, 11.49 real, 2.53e+08 points.step/s, 8 var
# Multigrid (GPU), 2991 steps, 45.4397 CPU, 45.44 real, 2.76e+08 points.step/s, 8 var
# Multigrid (GPU), 3011 steps, 179.499 CPU, 179.5 real, 2.81e+08 points.step/s, 8 var
CFLAGS='-DBENCHMARK=1' make turbulence.cuda.tst
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.cuda/turbulence.cuda $i 2> /tmp/log;
done
# Title: RTX3050(cuda)
# Multigrid (cuda), 2009 steps, 1.19183 CPU, 1.191 real, 2.76e+07 points.step/s, 8 var
# Multigrid (cuda), 2008 steps, 1.39756 CPU, 1.397 real, 9.41e+07 points.step/s, 8 var
# Multigrid (cuda), 2328 steps, 2.872 CPU, 2.871 real, 2.12e+08 points.step/s, 8 var
# Multigrid (cuda), 2767 steps, 10.1896 CPU, 10.19 real, 2.85e+08 points.step/s, 8 var
# Multigrid (cuda), 2991 steps, 40.408 CPU, 40.39 real, 3.1e+08 points.step/s, 8 var
# Multigrid (cuda), 3011 steps, 167.324 CPU, 167.3 real, 3.02e+08 points.step/s, 8 var
OpenGL renderer string: Quadro RTX 6000/PCIe/SSE2
Dedicated video memory: 24576 MB
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu $i 2> /tmp/log;
done
# Title: RTX6000
# Multigrid (GPU), 2009 steps, 1.32091 CPU, 1.321 real, 2.49e+07 points.step/s, 8 var
# Multigrid (GPU), 2008 steps, 1.45009 CPU, 1.45 real, 9.07e+07 points.step/s, 8 var
# Multigrid (GPU), 2328 steps, 2.27639 CPU, 2.276 real, 2.68e+08 points.step/s, 8 var
# Multigrid (GPU), 2767 steps, 4.9143 CPU, 4.915 real, 5.9e+08 points.step/s, 8 var
# Multigrid (GPU), 2991 steps, 14.5549 CPU, 14.56 real, 8.62e+08 points.step/s, 8 var
# Multigrid (GPU), 3011 steps, 53.9424 CPU, 53.95 real, 9.36e+08 points.step/s, 8 var
OpenGL renderer string: NVIDIA GeForce RTX 4090/PCIe/SSE2
Dedicated video memory: 24564 MB
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu $i 2> /tmp/log;
done
# Title: RTX4090
# Multigrid (gpu), 2009 steps, 1.2733 CPU, 1.273 real, 2.58e+07 points.step/s, 8 var
# Multigrid (gpu), 2008 steps, 1.42316 CPU, 1.423 real, 9.24e+07 points.step/s, 8 var
# Multigrid (gpu), 2328 steps, 1.96559 CPU, 1.966 real, 3.1e+08 points.step/s, 8 var
# Multigrid (gpu), 2767 steps, 3.07085 CPU, 3.071 real, 9.44e+08 points.step/s, 8 var
# Multigrid (gpu), 2991 steps, 6.72074 CPU, 6.721 real, 1.87e+09 points.step/s, 8 var
# Multigrid (gpu), 3011 steps, 30.9292 CPU, 30.93 real, 1.63e+09 points.step/s, 8 var
CFLAGS='-DBENCHMARK=1' make turbulence.cuda.tst
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.cuda/turbulence.cuda $i 2> /tmp/log;
done
# Title: RTX4090(cuda)
# Multigrid (cuda), 2009 steps, 1.03011 CPU, 1.03 real, 3.19e+07 points.step/s, 8 var
# Multigrid (cuda), 2008 steps, 1.19515 CPU, 1.195 real, 1.1e+08 points.step/s, 8 var
# Multigrid (cuda), 2328 steps, 1.61682 CPU, 1.617 real, 3.77e+08 points.step/s, 8 var
# Multigrid (cuda), 2767 steps, 2.66406 CPU, 2.663 real, 1.09e+09 points.step/s, 8 var
# Multigrid (cuda), 2991 steps, 5.8545 CPU, 5.853 real, 2.14e+09 points.step/s, 8 var
# Multigrid (cuda), 3011 steps, 26.6636 CPU, 26.66 real, 1.89e+09 points.step/s, 8 var
CFLAGS='-DBENCHMARK=1' make turbulence.ocl.tst
for i in 128 256 512 1024 2048 4096; do
OMP_NUM_THREADS=1 turbulence.ocl/turbulence.ocl $i 2> /tmp/log;
done
# Title: RTX4090(opencl)
# Multigrid (opencl), 2009 steps, 2.29921 CPU, 2.299 real, 1.43e+07 points.step/s, 8 var
# Multigrid (opencl), 2008 steps, 2.62439 CPU, 2.624 real, 5.01e+07 points.step/s, 8 var
# Multigrid (opencl), 2329 steps, 3.56261 CPU, 3.562 real, 1.71e+08 points.step/s, 8 var
# Multigrid (opencl), 2765 steps, 5.39095 CPU, 5.39 real, 5.38e+08 points.step/s, 8 var
# Multigrid (opencl), 2991 steps, 9.12213 CPU, 9.12 real, 1.37e+09 points.step/s, 8 var
# Multigrid (opencl), 3011 steps, 30.6107 CPU, 30.6 real, 1.65e+09 points.step/s, 8 var
OpenGL renderer string: NVIDIA GeForce RTX 4090 D/PCIe/SSE2 (stokes.lmm.jussieu.fr)
Dedicated video memory: 24564 MB
# Title: RTX4090D
# Multigrid (GPU), 2009 steps, 1.39246 CPU, 1.392 real, 2.36e+07 points.step/s, 8 var
# Multigrid (GPU), 2008 steps, 1.5809 CPU, 1.581 real, 8.32e+07 points.step/s, 8 var
# Multigrid (GPU), 2328 steps, 2.20537 CPU, 2.205 real, 2.77e+08 points.step/s, 8 var
# Multigrid (GPU), 2767 steps, 3.45719 CPU, 3.457 real, 8.39e+08 points.step/s, 8 var
# Multigrid (GPU), 2991 steps, 7.41217 CPU, 7.413 real, 1.69e+09 points.step/s, 8 var
# Multigrid (GPU), 3011 steps, 31.3258 CPU, 31.33 real, 1.61e+09 points.step/s, 8 var
CFLAGS='-DTRACE=2 -DBENCHMARK=1' make turbulence.gpu.tst
__NV_PRIME_RENDER_OFFLOAD=1 __GLX_VENDOR_LIBRARY_NAME=nvidia OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu 256
# Multigrid (GPU), 2008 steps, 5.22215 CPU, 5.223 real, 2.52e+07 points.step/s, 8 var
calls total self % total function
4025 4.64 4.11 78.6% mg_cycle():/src/poisson.h:92
336353 0.56 0.55 10.5% setup_shader():/src/grid/gpu/grid.h:1721
6034 0.14 0.14 2.7% gpu_reduction():/src/poisson.h:381
2009 4.97 0.14 2.6% mg_solve():/src/poisson.h:230
2009 0.09 0.08 1.6% tracer_fluxes():/src/bcg.h:61
__NV_PRIME_RENDER_OFFLOAD=1 __GLX_VENDOR_LIBRARY_NAME=nvidia OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu 512
# Multigrid (GPU), 2328 steps, 8.17009 CPU, 8.17 real, 7.47e+07 points.step/s, 8 var
calls total self % total function
4670 6.88 6.17 75.6% mg_cycle():/src/poisson.h:92
440039 0.75 0.73 9.0% setup_shader():/src/grid/gpu/grid.h:1721
2329 7.58 0.33 4.1% mg_solve():/src/poisson.h:230
6999 0.26 0.26 3.1% gpu_reduction():/src/poisson.h:381
2329 0.25 0.23 2.8% tracer_fluxes():/src/bcg.h:61
2329 0.34 0.09 1.1% advection():/src/bcg.h:99
2329 0.08 0.08 1.0% gpu_reduction():/src/timestep.h:6
2329 0.08 0.08 1.0% gpu_reduction():/src/poisson.h:163
2329 0.17 0.08 1.0% velocity():/src/advection.h:54
__NV_PRIME_RENDER_OFFLOAD=1 __GLX_VENDOR_LIBRARY_NAME=nvidia OMP_NUM_THREADS=1 turbulence.gpu/turbulence.gpu 1024
# Multigrid (GPU), 2767 steps, 17.2732 CPU, 17.27 real, 1.68e+08 points.step/s, 8 var
calls total self % total function
5559 13.17 12.23 70.8% mg_cycle():/src/poisson.h:92
2768 15.34 1.15 6.7% mg_solve():/src/poisson.h:230
574414 0.99 0.97 5.6% setup_shader():/src/grid/gpu/grid.h:1721
2768 0.85 0.83 4.8% tracer_fluxes():/src/bcg.h:61
8327 0.76 0.76 4.4% gpu_reduction():/src/poisson.h:381
2768 1.17 0.32 1.8% advection():/src/bcg.h:99
2768 0.50 0.26 1.5% velocity():/src/advection.h:54
2768 15.59 0.25 1.4% velocity_0():/src/navier-stokes/stream.h:87
2768 0.24 0.24 1.4% gpu_reduction():/src/timestep.h:6
2768 0.23 0.23 1.3% gpu_reduction():/src/poisson.h:163
