Hi all, I've set up a box to run the simple benchmark of compiling librustc with the stage2 rustc. The data is landing in http://hnn.mrsd.org/~cmr/, I don't know what to do with it, hopefully someone else does! (dbaupp expressed interest in integrating it into IRFY). Each commit is its own directory, mem.json is the output of my memory profiling script (https://github.com/cmr/rust-bench/blob/master/mem-bench.py), and time.txt is the output of GNU time on a separate run of it (to avoid the skew caused by the memory profiling script). I'll be going through src/tests/bench for good benchmarks to also profile, and I will start storing generated IR for some of them too.
However, we have few good, solid benchmarks that stress a single part of the codegen or libraries to be able to easily spot regressions. Writing these should be easy, and would be a wonderful project for anyone looking to contribute to Rust. I've opened https://github.com/mozilla/rust/issues/7532 about it. _______________________________________________ Rust-dev mailing list [email protected] https://mail.mozilla.org/listinfo/rust-dev
