Re: [go-nuts] Re: Question on 'cmd/compile: handle partially overlapping assignments' behavior on ARM64.

I am going to open an issue for this regression on ARM64 unless there are objections or a view that this is the intended behavior. Josh On Tuesday, September 20, 2022 at 3:04:00 PM UTC-4 ccie...@gmail.com wrote: > Hi Peter, > > Please note that versions devel *go1.20-bcd44b61d3*, *1.19.1*, and

Re: [go-nuts] Re: Question on 'cmd/compile: handle partially overlapping assignments' behavior on ARM64.

Hi Peter, Please note that versions devel *go1.20-bcd44b61d3*, *1.19.1*, and *1.18.6* show times in the *4xxx ns/op* range for *BenchmarkSliceOfArray* while all the versions of *1.19* and *1.18* show times in the *2xxx ns/op* range. As noted earlier in this thread, the regression is seen on *A

[go-nuts] Question on 'cmd/compile: handle partially overlapping assignments' behavior on ARM64.

I stumbled across a performance regression on ARM64 for go 1.18.6 and 1.19.1 that wasn't present in earlier releases. In the following benchmark, you can see `BenchmarkSliceOfArray` experiences a 70% increase in execution time while `BenchmarkSliceOfInt` remains unchanged. BenchmarkSliceOf