Closed
Description
Spawned off of #48247
There have been instances where we have injected performance issues when compiling winapi
It has been suggested that we add winapi
as a standard benchmark.
(In general I would like to revisit our benchmarks as a whole, which I guess is a task covered by #48750.)
This issue is more targeted: We should add winapi
, or a variant that captures its workload but works on Linux hosts, to the set of rustc
benchmarks that live on perf.rust-lang.org.