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.