mirror of
https://github.com/RRZE-HPC/asmbench.git
synced 2025-09-05 00:20:06 +02:00
18 lines
451 B
ReStructuredText
18 lines
451 B
ReStructuredText
asmbench
|
|
========
|
|
|
|
A benchmark toolkit for assembly instructions using the LLVM JIT.
|
|
|
|
Usage
|
|
=====
|
|
|
|
To benchmark latency and throughput of a 64bit integer add use the following command:
|
|
|
|
``python -m asmbench 'add {src:i64:r}, {srcdst:i64:r}'``
|
|
|
|
To benchmark two instructions interleaved use this:
|
|
|
|
``python -m asmbench 'add {src:i64:r}, {srcdst:i64:r}' 'sub {src:i64:r}, {srcdst:i64:r}'``
|
|
|
|
To find out more add `-h` for help and `-v` for verbose mode.
|