code large_stringlengths 193 899 | workload_type large_stringclasses 15
values | gpu_name large_stringclasses 12
values | gpu_features large_stringclasses 12
values | runtime_ms float64 0.01 4.35k | flops float64 30 21,045B | memory_bytes int64 160 13.2B | arithmetic_intensity float64 0 6.37k | dtype_bytes int64 2 4 | workload_params large_stringlengths 11 79 |
|---|---|---|---|---|---|---|---|---|---|
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.032947 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.054493 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.053096 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.055044 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.025102 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.058346 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.060272 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize()
| matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.028238 | 67,108,864 | 1,589,248 | 42.226804 | 2 | {"M": 128, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.107607 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.068578 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.048155 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.049932 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.07596 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.070338 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.071392 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.05731 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.042275 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.040233 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.047343 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.058278 | 268,435,456 | 6,422,528 | 41.795918 | 4 | {"M": 128, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.049528 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.04338 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.053877 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.022901 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.034504 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.033678 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.026707 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.051653 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.040809 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.046608 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.035668 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.05866 | 268,435,456 | 3,211,264 | 83.591837 | 2 | {"M": 128, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.135833 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.078234 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.05107 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.096789 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.08198 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.088036 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.049687 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.07495 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.036901 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.078808 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.074606 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.049233 | 536,870,912 | 10,747,904 | 49.95122 | 4 | {"M": 128, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.062 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.052722 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.058639 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.038794 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.054761 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.080564 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.054796 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.049031 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.054168 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.038254 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.035582 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.062374 | 536,870,912 | 5,373,952 | 99.902439 | 2 | {"M": 128, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.240072 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.155618 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.102171 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.157658 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.109293 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.109577 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.047482 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.093666 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.069344 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.064391 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.070263 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.075812 | 1,073,741,824 | 19,398,656 | 55.351351 | 4 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.09895 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.050581 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.044158 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.032705 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.06357 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.074118 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.038864 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.060234 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.080578 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.040953 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.052011 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.066129 | 1,073,741,824 | 9,699,328 | 110.702703 | 2 | {"M": 128, "N": 4096, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.431416 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.299025 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.140653 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.218778 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.181099 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.119395 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.100283 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.118615 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.059552 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.059776 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.109149 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.086794 | 2,147,483,648 | 36,700,160 | 58.514286 | 4 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.120405 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.069504 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.123797 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.051393 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.035218 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.121493 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.061796 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 4096) -> (128, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.074665 | 2,147,483,648 | 18,350,080 | 117.028571 | 2 | {"M": 128, "N": 4096, "K": 2048, "dtype": "float16"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.