TritonBench
Operators
Testcases
About
Login
Sign Up
Testcase Results
Press Enter or click outside to search. Database is only queried when you search.
50 per page
100 per page
250 per page
500 per page
1000 per page
Export Data
Clear
Field Chooser
All
None
Default
Kernel Fields
Name
Input Shapes
Output Shapes
Source Implementation
Operator Gigaflops
Operator Interface Size (GB)
Warmup Iterations
Test Iterations
Hardware Environment
Target Instance
CPU
GPU
Software Environment
Triton Version
CUDA Version
PyTorch Version
Performance Results
Latency (μs)
Peak GPU Memory (MB)
Test Status
Timestamp
Peak CPU Memory (MB)
Compile Time (s)
Hardware Frequency (MHz)
Best Configuration
Block M
Block N
Block K
Num Warps
Cluster Size
Software Pipeline Stages
Grid Size
Access Control
Access Group
Benchmark Data
Loading...
Loading benchmarks...