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: (2048, 1024) x (1024, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.056776 | 2,147,483,648 | 7,340,032 | 292.571429 | 2 | {"M": 2048, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 1024) x (1024, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.028509 | 2,147,483,648 | 7,340,032 | 292.571429 | 2 | {"M": 2048, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 1024) x (1024, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.049603 | 2,147,483,648 | 7,340,032 | 292.571429 | 2 | {"M": 2048, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 1024) x (1024, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.093828 | 2,147,483,648 | 7,340,032 | 292.571429 | 2 | {"M": 2048, "N": 512, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.57086 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.391145 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.178001 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.390155 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.339401 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.240943 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.108397 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.23224 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.130414 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.131752 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.198306 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.216622 | 4,294,967,296 | 25,165,824 | 170.666667 | 4 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.132657 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.098642 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.122528 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.047038 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.048646 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.093941 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.065829 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.151207 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.081977 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.035832 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.030328 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 2048) x (2048, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.07519 | 4,294,967,296 | 12,582,912 | 341.333333 | 2 | {"M": 2048, "N": 512, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 1.46439 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.772602 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.376851 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.598892 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.525046 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.418798 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.189513 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.426708 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.198696 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.211574 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.234329 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.348486 | 8,589,934,592 | 46,137,344 | 186.181818 | 4 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.252927 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.151845 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.208403 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.084403 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.087184 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.091363 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.091715 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.22895 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.087277 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.04419 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.038955 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 4096) x (4096, 512) -> (2048, 512)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 4096, dtype=torch.float16, device='cuda')
B = torch.randn(4096, 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} | 0.16297 | 8,589,934,592 | 23,068,672 | 372.363636 | 2 | {"M": 2048, "N": 512, "K": 4096, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.086913 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.077741 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.055813 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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 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} | 0.064271 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.047634 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.064453 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.036601 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.049561 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.052846 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.048656 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.049304 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.049043 | 268,435,456 | 9,175,040 | 29.257143 | 4 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.057328 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.036153 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.061238 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.049245 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.047816 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.038551 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.038132 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.052076 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.050282 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.04664 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.027537 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 64) x (64, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 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} | 0.055272 | 268,435,456 | 4,587,520 | 58.514286 | 2 | {"M": 2048, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.177116 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.12735 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.079455 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.139209 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.114777 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.08368 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.062163 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.056629 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.06241 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.077475 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.081374 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float32, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.084427 | 1,073,741,824 | 11,534,336 | 93.090909 | 4 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.051639 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.035505 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.091336 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.038811 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.05867 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.04852 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.046264 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.070701 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.039509 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.045303 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.056568 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (2048, 256) x (256, 1024) -> (2048, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(2048, 256, dtype=torch.float16, device='cuda')
B = torch.randn(256, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.synchron... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.080394 | 1,073,741,824 | 5,767,168 | 186.181818 | 2 | {"M": 2048, "N": 1024, "K": 256, "dtype": "float16"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.