Files
asmbench/README.rst
2018-09-25 10:37:53 +02:00

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.