results.txt

DEFAULT PARAMETERS (10 Iterations)

CUDA (NVIDIA GeForce RTX 3090)

miniBUDE:
Using kernel parameters: <<<256,64>>> 1:65536
- Kernel time:    140.443 ms
- Average time:   14.044 ms
- Interactions/s: 113.803 billion
- GFLOP/s:        3700.988
- GFInst/s:       2845.087

miniBUDE JACC:
- Kernel time:    0.217 ms
- Average time:   0.022 ms
- Interactions/s: 73667.279 billion
- GFLOP/s:        2223029.574
- GFInst/s:       1841681.977



AMDGPU (AMD Radeon RX 6900 XT)

miniBUDE:
Using kernel parameters: <<<16384,64>>> 1:65536
- Kernel time:    2923.611 ms
- Average time:   292.361 ms
- Interactions/s: 5.467 billion
- GFLOP/s:        177.786
- GFInst/s:       136.671

miniBUDE JACC:
- Kernel time:    2683.075 ms
- Average time:   268.307 ms
- Interactions/s: 5.957 billion
- GFLOP/s:        179.760
- GFInst/s:       148.924


Threads (64 threads) (AMD EPYC 7742 64-Core Processor):

julia --project=Threaded -t 64 src/Threaded.jl

miniBUDE:
- Kernel time:    1522.385 ms
- Average time:   152.239 ms
- Interactions/s: 10.499 billion
- GFLOP/s:        316.812
- GFInst/s:       262.465



julia --project=. -e 'include("src/JACCBUDE.jl")' -t 64

miniBUDEJACC:
- Kernel time:    1342.390 ms
- Average time:   134.239 ms
- Interactions/s: 11.906 billion
- GFLOP/s:        359.292
- GFInst/s:       297.658