code large_stringlengths 193 899 | workload_type large_stringclasses 15
values | gpu_name large_stringclasses 12
values | gpu_features large_stringclasses 12
values | runtime_ms float64 0.01 4.35k | flops float64 30 21,045B | memory_bytes int64 160 13.2B | arithmetic_intensity float64 0 6.37k | dtype_bytes int64 2 4 | workload_params large_stringlengths 11 79 |
|---|---|---|---|---|---|---|---|---|---|
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.102048 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.068885 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.051217 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.086009 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.047189 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.062894 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.063225 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.084817 | 1,073,741,824 | 13,107,200 | 81.92 | 4 | {"M": 4096, "N": 512, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.070218 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.043677 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.07188 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.035877 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.060104 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.056958 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.067864 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.042446 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.057143 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.04206 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.050669 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 256) x (256, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.076896 | 1,073,741,824 | 6,553,600 | 163.84 | 2 | {"M": 4096, "N": 512, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.365456 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.203548 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.140785 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.170155 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.200161 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.096813 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.072693 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.110047 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.064526 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.080184 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.121785 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.089723 | 2,147,483,648 | 17,825,792 | 120.470588 | 4 | {"M": 4096, "N": 512, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.066401 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.087358 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.122757 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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} | 0.051643 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.045335 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.086928 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.044963 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.110343 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.069502 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.032002 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.055348 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 512) x (512, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 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 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} | 0.075616 | 2,147,483,648 | 8,912,896 | 240.941176 | 2 | {"M": 4096, "N": 512, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.622966 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.354582 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.252393 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.309455 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.415953 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.250927 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.092414 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.169949 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.122193 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.143377 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.153548 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 512, 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} | 0.161507 | 4,294,967,296 | 27,262,976 | 157.538462 | 4 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.112732 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.081058 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.17197 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.047588 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.062583 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.063623 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.077367 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.122695 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.089379 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.068172 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.05932 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 1024) x (1024, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 512, 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} | 0.111364 | 4,294,967,296 | 13,631,488 | 315.076923 | 2 | {"M": 4096, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 1.715273 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 1.018418 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.378033 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.762431 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchro... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.81183 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.436253 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.196413 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.340642 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.198976 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.216872 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.25161 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 512, 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} | 0.369733 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.259851 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.114168 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.21963 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.060279 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.082833 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.158554 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.083114 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.236002 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.093055 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.05752 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.056678 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 2048) x (2048, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 512, 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} | 0.200745 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 4096, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 3.3967 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 1.297739 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 0.629734 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 1.035783 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 1.086518 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 0.993958 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 0.286018 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (4096, 4096) x (4096, 512) -> (4096, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(4096, 4096, dtype=torch.float32, device='cuda')
B = torch.randn(4096, 512, 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} | 0.804281 | 17,179,869,184 | 83,886,080 | 204.8 | 4 | {"M": 4096, "N": 512, "K": 4096, "dtype": "float32"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.