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: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.032895 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.050709 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.02978 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.041334 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.026599 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.039958 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.037327 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float32, device='cuda')
B = torch.randn(64, 1024, 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} | 0.01775 | 16,777,216 | 819,200 | 20.48 | 4 | {"M": 128, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.022759 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.052882 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.053809 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.047367 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.032252 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.039622 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.034139 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.059546 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.034915 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.022258 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.052017 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 64) x (64, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 64, dtype=torch.float16, device='cuda')
B = torch.randn(64, 1024, 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} | 0.049093 | 16,777,216 | 409,600 | 40.96 | 2 | {"M": 128, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.056938 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.052767 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.072095 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.040148 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.061071 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.038374 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.024809 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.052815 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.048008 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.059732 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.053101 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, 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} | 0.031535 | 67,108,864 | 1,703,936 | 39.384615 | 4 | {"M": 128, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.04498 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.053428 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.029243 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.045027 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.059302 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.024758 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.032574 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.048191 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.053497 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.026413 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.027213 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 256) x (256, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, 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} | 0.030078 | 67,108,864 | 851,968 | 78.769231 | 2 | {"M": 128, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.075653 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.061703 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.023751 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.050081 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.049224 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.048133 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.063337 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.053597 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.02872 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.036273 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.05077 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 1024, 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} | 0.058053 | 134,217,728 | 2,883,584 | 46.545455 | 4 | {"M": 128, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.045021 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.029512 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.026931 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.043483 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.05114 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.07072 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.03136 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.02807 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.046529 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.046932 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.053659 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 512) x (512, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 1024, 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} | 0.031703 | 134,217,728 | 1,441,792 | 93.090909 | 2 | {"M": 128, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.081881 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.073692 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.057163 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.057421 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.033893 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.050127 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.047236 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.048197 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.055768 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.036163 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.047603 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.052562 | 268,435,456 | 5,242,880 | 51.2 | 4 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.039283 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.048891 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.036812 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.04014 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.044444 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.037402 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.032126 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.06551 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.053055 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.049726 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.039253 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 1024) x (1024, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, 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} | 0.063321 | 268,435,456 | 2,621,440 | 102.4 | 2 | {"M": 128, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.110771 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.086661 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.064469 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.080102 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.095337 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.069963 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.055252 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (128, 2048) x (2048, 1024) -> (128, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(128, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, 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} | 0.060423 | 536,870,912 | 9,961,472 | 53.894737 | 4 | {"M": 128, "N": 1024, "K": 2048, "dtype": "float32"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.