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: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.219931 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.150304 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.182943 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.090006 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.09728 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.305173 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.09114 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.18708 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.120441 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.042586 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.05003 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.205336 | 8,589,934,592 | 38,797,312 | 221.405405 | 2 | {"M": 8192, "N": 256, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 3.172373 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 1.604271 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.902203 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.991901 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 1.142371 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.820126 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.27085 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.679942 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.289717 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.410548 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.623735 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.647744 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.448054 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.212632 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.360652 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.137091 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.148437 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.374683 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.150083 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.409549 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.154478 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.052109 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.103875 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 256) -> (8192, 256)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 256, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.350848 | 17,179,869,184 | 73,400,320 | 234.057143 | 2 | {"M": 8192, "N": 256, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.140102 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.093336 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.064477 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.068684 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.086007 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.082459 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.046183 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.053827 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.067045 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.05521 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.049036 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 512, 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} | 0.073316 | 536,870,912 | 19,005,440 | 28.248276 | 4 | {"M": 8192, "N": 512, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.078143 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.070083 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.061884 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.062443 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.061312 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.071162 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.072439 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.077617 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.056451 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.059353 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.038774 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 512, 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} | 0.064713 | 536,870,912 | 9,502,720 | 56.496552 | 2 | {"M": 8192, "N": 512, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.434839 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.224955 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.136941 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.181766 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.261175 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.114984 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.066829 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.127746 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.089874 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.077791 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.090205 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 512, 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} | 0.102392 | 2,147,483,648 | 25,690,112 | 83.591837 | 4 | {"M": 8192, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.12485 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.047707 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.087438 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.070978 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.047883 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.086338 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.052227 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.098633 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.056909 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.048787 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.033582 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 512, 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} | 0.10318 | 2,147,483,648 | 12,845,056 | 167.183673 | 2 | {"M": 8192, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.689485 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.440167 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.206105 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.295164 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.332633 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.21819 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.099358 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.193035 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.090212 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.120635 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.18356 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 512, 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} | 0.149394 | 4,294,967,296 | 34,603,008 | 124.121212 | 4 | {"M": 8192, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 512, 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} | 0.139048 | 4,294,967,296 | 17,301,504 | 248.242424 | 2 | {"M": 8192, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 512, 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} | 0.084682 | 4,294,967,296 | 17,301,504 | 248.242424 | 2 | {"M": 8192, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 512, 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} | 0.148747 | 4,294,967,296 | 17,301,504 | 248.242424 | 2 | {"M": 8192, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 512) -> (8192, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 512, 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} | 0.049132 | 4,294,967,296 | 17,301,504 | 248.242424 | 2 | {"M": 8192, "N": 512, "K": 512, "dtype": "float16"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.