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: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.065738 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.124793 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.10496 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.124008 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.076711 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.064309 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.059873 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.115914 | 4,294,967,296 | 19,922,944 | 215.578947 | 2 | {"M": 4096, "N": 2048, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 1.739229 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.890587 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.467421 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.585164 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.782005 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.377114 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.173324 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.323975 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.153327 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.197764 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.364405 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.352739 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.239094 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.129906 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.186521 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.06517 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.0962 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.109984 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.117236 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.209071 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.096514 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.053595 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.052319 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | 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} | 0.194953 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 2048, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.638334 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.541487 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.735795 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.997551 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.045014 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.631272 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.331586 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.580678 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.362909 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.372129 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.558307 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.672685 | 17,179,869,184 | 58,720,256 | 292.571429 | 4 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.394508 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.212273 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.422314 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.11802 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.141597 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.242254 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.163643 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.313764 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.198725 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.053718 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.067502 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.335856 | 17,179,869,184 | 29,360,128 | 585.142857 | 2 | {"M": 4096, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 5.697466 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.866432 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.585138 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.139381 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.793542 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.695086 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.714324 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.619184 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.583474 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.848332 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.810873 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.931578 | 34,359,738,368 | 83,886,080 | 409.6 | 4 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.783795 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.483961 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.84903 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.215392 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.174456 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.415681 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.33725 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.616778 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.276535 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.136683 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.118344 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.672443 | 34,359,738,368 | 41,943,040 | 819.2 | 2 | {"M": 4096, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 12.768591 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 5.28707 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.647997 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 4.762708 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 3.80745 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 3.036134 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.992125 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.089525 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.234796 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.734622 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.782613 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 2.339515 | 68,719,476,736 | 134,217,728 | 512 | 4 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.906819 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.992579 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.175186 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.410348 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.361997 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.892146 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 0.594319 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 2048) -> (4096, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | 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} | 1.489193 | 68,719,476,736 | 67,108,864 | 1,024 | 2 | {"M": 4096, "N": 2048, "K": 4096, "dtype": "float16"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.