Re: [go-nuts] Re: Some confusion about escape analysis

2022-08-26 Thread Sean Foley
Perhaps it will have a small benefit. However, according to the benchmark, the performance is slightly better in the case where there is 1 alloc/op, Normally people avoid the heap allocation for better performance and there is no evidence of that here. On Fri, Aug 26, 2022 at 3:32 AM Lee Chou wr

[go-nuts] Re: Some confusion about escape analysis

2022-08-26 Thread Lee Chou
Thanks for taking the time for the clue. It seems that this is an open issue. Actually, It work for this case. aovid the alloc will have more benefit for gc. 在2022年8月26日星期五 UTC+8 02:49:08 写道: > Actually, looking more closely at your benchmark numbers, the trick does > not work in your case.

[go-nuts] Re: Some confusion about escape analysis

2022-08-25 Thread Sean Foley
Actually, looking more closely at your benchmark numbers, the trick does not work in your case. Which makes sense to me, I don't see why it would work. On Wednesday, August 24, 2022 at 2:59:34 AM UTC-4 mozart...@gmail.com wrote: > I was try using net.Buffers for writev. But I meet a unexptecte

[go-nuts] Re: Some confusion about escape analysis

2022-08-25 Thread Sean Foley
That trick was copied from here: https://cs.opensource.google/go/go/+/refs/tags/go1.19:src/strings/builder.go;l=26 See https://github.com/golang/go/issues/23382 and https://github.com/golang/go/issues/7921 I don't know why the trick works in your case. It appears to be safe. On Wednesday, Aug

[go-nuts] Re: Some confusion about escape analysis

2022-08-25 Thread Sean Foley
That trick was copied from here: https://cs.opensource.google/go/go/+/refs/tags/go1.19:src/strings/builder.go;l=26 See https://github.com/golang/go/issues/23382 and https://github.com/golang/go/issues/23382 and https://github.com/golang/go/issues/7921 I don't know why the trick works in your