code large_stringlengths 193 899 | workload_type large_stringclasses 15
values | gpu_name large_stringclasses 12
values | gpu_features large_stringclasses 12
values | runtime_ms float64 0.01 4.35k | flops float64 30 21,045B | memory_bytes int64 160 13.2B | arithmetic_intensity float64 0 6.37k | dtype_bytes int64 2 4 | workload_params large_stringlengths 11 79 |
|---|---|---|---|---|---|---|---|---|---|
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.065205 | 1,073,741,824 | 17,956,864 | 59.79562 | 2 | {"M": 8192, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.035499 | 1,073,741,824 | 17,956,864 | 59.79562 | 2 | {"M": 8192, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.048255 | 1,073,741,824 | 17,956,864 | 59.79562 | 2 | {"M": 8192, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 64) x (64, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.076347 | 1,073,741,824 | 17,956,864 | 59.79562 | 2 | {"M": 8192, "N": 1024, "K": 64, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.851833 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.463759 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.19806 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.36377 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.322134 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.312936 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.111631 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.174067 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.113424 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.122012 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.162235 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.212774 | 4,294,967,296 | 42,991,616 | 99.902439 | 4 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.155054 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.10056 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.146717 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.065892 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.054379 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.115505 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.042505 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.117012 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.085493 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.051248 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.064996 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 256) x (256, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 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} | 0.109823 | 4,294,967,296 | 21,495,808 | 199.804878 | 2 | {"M": 8192, "N": 1024, "K": 256, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 1.794705 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.764818 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.450058 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.565653 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.588663 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.397795 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.182026 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.325623 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.173939 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.169891 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.231586 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float32, device='cuda')
B = torch.randn(512, 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} | 0.282299 | 8,589,934,592 | 52,428,800 | 163.84 | 4 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.219879 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.132416 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.183043 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.063693 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.089397 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.129784 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.119528 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.200784 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.116498 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.061604 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.060478 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 512) x (512, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 512, dtype=torch.float16, device='cuda')
B = torch.randn(512, 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} | 0.171562 | 8,589,934,592 | 26,214,400 | 327.68 | 2 | {"M": 8192, "N": 1024, "K": 512, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 2.324857 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 1.64372 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.696811 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 1.607342 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 1.207982 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.813549 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.269284 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.618166 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.397892 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.400601 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.552926 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float32, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.629127 | 17,179,869,184 | 71,303,168 | 240.941176 | 4 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.39588 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.311083 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.44318 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.133221 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.111946 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.232881 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.14686 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.279825 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.123831 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.043429 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.072306 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 1024) x (1024, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 1024, dtype=torch.float16, device='cuda')
B = torch.randn(1024, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.287147 | 17,179,869,184 | 35,651,584 | 481.882353 | 2 | {"M": 8192, "N": 1024, "K": 1024, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 8.829973 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 2.925801 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 1.283453 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 2.487829 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 1.80913 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 1.914983 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.520669 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 1.59894 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.656466 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.865986 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 1.193404 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float32, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float32, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 1.23519 | 34,359,738,368 | 109,051,904 | 315.076923 | 4 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float32"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA T4 | {"gpu_name": "NVIDIA T4", "cuda_cores": 2560, "tensor_cores": 320, "memory_gb": 16, "memory_bandwidth_gbps": 320, "base_clock_mhz": 585, "boost_clock_mhz": 1590, "sm_count": 40, "fp32_tflops": 8.1, "fp16_tflops": 65, "tdp_watts": 70, "compute_capability": 7.5, "l2_cache_mb": 4} | 0.70324 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA V100 | {"gpu_name": "NVIDIA V100", "cuda_cores": 5120, "tensor_cores": 640, "memory_gb": 32, "memory_bandwidth_gbps": 900, "base_clock_mhz": 1230, "boost_clock_mhz": 1530, "sm_count": 80, "fp32_tflops": 15.7, "fp16_tflops": 125, "tdp_watts": 300, "compute_capability": 7.0, "l2_cache_mb": 6} | 0.43313 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A10G | {"gpu_name": "NVIDIA A10G", "cuda_cores": 9216, "tensor_cores": 288, "memory_gb": 24, "memory_bandwidth_gbps": 600, "base_clock_mhz": 885, "boost_clock_mhz": 1695, "sm_count": 80, "fp32_tflops": 31.2, "fp16_tflops": 62.5, "tdp_watts": 150, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.991663 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 40GB | {"gpu_name": "NVIDIA A100 40GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 40, "memory_bandwidth_gbps": 1555, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.21925 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA A100 80GB | {"gpu_name": "NVIDIA A100 80GB", "cuda_cores": 6912, "tensor_cores": 432, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 765, "boost_clock_mhz": 1410, "sm_count": 108, "fp32_tflops": 19.5, "fp16_tflops": 312, "tdp_watts": 400, "compute_capability": 8.0, "l2_cache_mb": 40} | 0.168907 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L4 | {"gpu_name": "NVIDIA L4", "cuda_cores": 7424, "tensor_cores": 232, "memory_gb": 24, "memory_bandwidth_gbps": 300, "base_clock_mhz": 795, "boost_clock_mhz": 2040, "sm_count": 58, "fp32_tflops": 30.3, "fp16_tflops": 121, "tdp_watts": 72, "compute_capability": 8.9, "l2_cache_mb": 48} | 0.505545 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA L40S | {"gpu_name": "NVIDIA L40S", "cuda_cores": 18176, "tensor_cores": 568, "memory_gb": 48, "memory_bandwidth_gbps": 864, "base_clock_mhz": 1110, "boost_clock_mhz": 2520, "sm_count": 142, "fp32_tflops": 91.6, "fp16_tflops": 183.2, "tdp_watts": 350, "compute_capability": 8.9, "l2_cache_mb": 96} | 0.319596 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 3090 | {"gpu_name": "NVIDIA RTX 3090", "cuda_cores": 10496, "tensor_cores": 328, "memory_gb": 24, "memory_bandwidth_gbps": 936, "base_clock_mhz": 1395, "boost_clock_mhz": 1695, "sm_count": 82, "fp32_tflops": 35.6, "fp16_tflops": 71, "tdp_watts": 350, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.818824 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX 4090 | {"gpu_name": "NVIDIA RTX 4090", "cuda_cores": 16384, "tensor_cores": 512, "memory_gb": 24, "memory_bandwidth_gbps": 1008, "base_clock_mhz": 2235, "boost_clock_mhz": 2520, "sm_count": 128, "fp32_tflops": 82.6, "fp16_tflops": 165.2, "tdp_watts": 450, "compute_capability": 8.9, "l2_cache_mb": 72} | 0.280955 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 SXM | {"gpu_name": "NVIDIA H100 SXM", "cuda_cores": 16896, "tensor_cores": 528, "memory_gb": 80, "memory_bandwidth_gbps": 3350, "base_clock_mhz": 1095, "boost_clock_mhz": 1830, "sm_count": 132, "fp32_tflops": 67, "fp16_tflops": 989, "tdp_watts": 700, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.119522 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA H100 PCIe | {"gpu_name": "NVIDIA H100 PCIe", "cuda_cores": 14592, "tensor_cores": 456, "memory_gb": 80, "memory_bandwidth_gbps": 2039, "base_clock_mhz": 1095, "boost_clock_mhz": 1620, "sm_count": 114, "fp32_tflops": 48, "fp16_tflops": 756, "tdp_watts": 350, "compute_capability": 9.0, "l2_cache_mb": 50} | 0.089663 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
import torch
def matmul_kernel(A, B):
# Matrix multiplication: (8192, 2048) x (2048, 1024) -> (8192, 1024)
C = torch.matmul(A, B)
return C
A = torch.randn(8192, 2048, dtype=torch.float16, device='cuda')
B = torch.randn(2048, 1024, dtype=torch.float16, device='cuda')
C = matmul_kernel(A, B)
torch.cuda.sync... | matmul | NVIDIA RTX A6000 | {"gpu_name": "NVIDIA RTX A6000", "cuda_cores": 10752, "tensor_cores": 336, "memory_gb": 48, "memory_bandwidth_gbps": 768, "base_clock_mhz": 1410, "boost_clock_mhz": 1860, "sm_count": 84, "fp32_tflops": 38.7, "fp16_tflops": 77.4, "tdp_watts": 300, "compute_capability": 8.6, "l2_cache_mb": 6} | 0.885828 | 34,359,738,368 | 54,525,952 | 630.153846 | 2 | {"M": 8192, "N": 1024, "K": 2048, "dtype": "float16"} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.