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, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.221903 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.319414 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.48637 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.451682 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.59287 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.881298 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.307412 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.248472 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.782363 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.363161 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.618402 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.222217 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.221953 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.378338 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.220238 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.74867 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.320318 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.1023 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.0911 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.60965 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 2048, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 12.793942 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 6.969157 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.604265 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 4.559304 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 5.117462 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.449616 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.97866 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.215786 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.181546 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.290181 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.765609 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.357332 | 68,719,476,736 | 150,994,944 | 455.111111 | 4 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.477076 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.907075 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.678841 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.40665 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.460483 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.880593 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.474432 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.119196 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.442252 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.138362 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.183875 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.954451 | 68,719,476,736 | 75,497,472 | 910.222222 | 2 | {"M": 8192, "N": 2048, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 22.746452 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 10.076322 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 5.166608 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 8.134521 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 8.312855 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 7.788945 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.659158 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 5.881018 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.82768 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.572172 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 3.102265 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 5.57999 | 137,438,953,472 | 234,881,024 | 585.142857 | 4 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 3.330035 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.137731 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.662465 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.547308 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.625604 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 1.758843 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.878952 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 2.580613 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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 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} | 1.320978 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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 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} | 0.198757 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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 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} | 0.337926 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 4096) x (4096, 2048) -> (8192, 2048)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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 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} | 2.210572 | 137,438,953,472 | 117,440,512 | 1,170.285714 | 2 | {"M": 8192, "N": 2048, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.649025 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.406316 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.289248 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.418212 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.3828 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.60096 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.421129 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.248281 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.288492 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.116689 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.128915 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.25385 | 4,294,967,296 | 137,363,456 | 31.267176 | 4 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.312552 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.15362 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.179464 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.111616 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.082287 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.375198 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.122705 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.169024 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.141101 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.077249 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.116906 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 4096, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchronize(... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.182677 | 4,294,967,296 | 68,681,728 | 62.534351 | 2 | {"M": 8192, "N": 4096, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 2.777441 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 1.759369 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 0.779028 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 1.240774 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 0.923658 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 0.89641 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 0.350803 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 4096) -> (8192, 4096)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 4096, 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} | 0.634522 | 17,179,869,184 | 146,800,640 | 117.028571 | 4 | {"M": 8192, "N": 4096, "K": 256, "dtype": "float32"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.