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: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.114403 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.167869 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.070604 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.122142 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.086886 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.049416 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.052169 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.148559 | 1,073,741,824 | 35,127,296 | 30.567164 | 4 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.128848 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.075412 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.063188 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.044508 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.047939 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.073227 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.040704 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.057998 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.084148 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.055362 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.057827 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 4096) x (4096, 2048) -> (64, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.054694 | 1,073,741,824 | 17,563,648 | 61.134328 | 2 | {"M": 64, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.074121 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.037496 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.041901 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.055323 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.040276 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.070969 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.051325 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.059862 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.051598 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.020914 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.024857 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 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} | 0.031099 | 33,554,432 | 2,113,536 | 15.875969 | 4 | {"M": 64, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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 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} | 0.04908 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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 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} | 0.04595 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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 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} | 0.028907 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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 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} | 0.052705 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.043067 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.031577 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.06169 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.028974 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.052474 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.032076 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.026562 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 64) x (64, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.033058 | 33,554,432 | 1,056,768 | 31.751938 | 2 | {"M": 64, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.062101 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.059026 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.051905 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.060012 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.058093 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.069651 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.03968 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.052128 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.027009 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.034575 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.038043 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.052675 | 134,217,728 | 5,308,416 | 25.283951 | 4 | {"M": 64, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.061178 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.056289 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.052447 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.050728 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.027242 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.077166 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.047519 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.045602 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.043481 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.034298 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.061435 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 256) x (256, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.039287 | 134,217,728 | 2,654,208 | 50.567901 | 2 | {"M": 64, "N": 4096, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.069744 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.058552 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.060337 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.077218 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.077414 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.07068 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.05284 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.04383 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.029997 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.043125 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.068153 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.077006 | 268,435,456 | 9,568,256 | 28.054795 | 4 | {"M": 64, "N": 4096, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.059174 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.044069 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.049378 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.05507 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.040753 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.046143 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.062401 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.047339 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.060898 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.033088 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.044737 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 512) x (512, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 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} | 0.067352 | 268,435,456 | 4,784,128 | 56.109589 | 2 | {"M": 64, "N": 4096, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.130848 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.114465 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.095918 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.07892 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.052796 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.123934 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.078654 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (64, 1024) x (1024, 4096) -> (64, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(64, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchroniz... | 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} | 0.084835 | 536,870,912 | 18,087,936 | 29.681159 | 4 | {"M": 64, "N": 4096, "K": 1024, "dtype": "float32"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.