Re: [go-nuts] Re: Help with WASM performance

2023-09-23 Thread Stephen Illingworth
Heh. Yes, it is only an "aim" I suppose. Yes. My code is single threaded. I note that threading will be added to the WASM spec in the near future so maybe there will be opportunity for me there when the change is added to the Go compiler. I'll put this issue to rest for now and revisit it in a

Re: [go-nuts] Re: Help with WASM performance

2023-09-23 Thread TheDiveO
The cynic in me notes the use of "aims". IIRC is single-threaded, do you know if your code is single- threaded too, or does it use several core for decoding? On Saturday, September 23, 2023 at 7:37:13 AM UTC+2 Stephen Illingworth wrote: > Comparison to native speeds is the key attraction of W

Re: [go-nuts] Re: Help with WASM performance

2023-09-22 Thread Stephen Illingworth
Comparison to native speeds is the key attraction of WASM. From the go.org blog published 10 days ago: https://go.dev/blog/wasi "WebAssembly (Wasm) is a binary instruction format originally designed for the web. It represents a standard that allows developers to run h

Re: [go-nuts] Re: Help with WASM performance

2023-09-22 Thread Robert Engels
Typically WASM performance is compared to JavaScript - ie. faster than JavaScript usually 1.5-10x. People don’t usually compare WASM with native apps. > On Sep 22, 2023, at 7:04 PM, Robert Engels wrote: > >  > WASM goes through the browser - so it is very different. Are you using OpenGL >

Re: [go-nuts] Re: Help with WASM performance

2023-09-22 Thread Robert Engels
WASM goes through the browser - so it is very different. Are you using OpenGL or similar? > On Sep 22, 2023, at 3:44 PM, Stephen Illingworth > wrote: > >  > I'm comparing the results of a program compiled for AMD64 and WASM > architectures. The code is the same except for the change of archi

Re: [go-nuts] Re: Help with WASM performance

2023-09-22 Thread Stephen Illingworth
I'm comparing the results of a program compiled for AMD64 and WASM architectures. The code is the same except for the change of architecture. The size of the difference in performance is unexpected to me but maybe it's normal. On Friday, 22 September 2023 at 20:16:20 UTC+1 Robert Engels wrote:

Re: [go-nuts] Re: Help with WASM performance

2023-09-22 Thread Robert Engels
When you say negative performance are you talking about a native app running directly on the hardware what are you referring to exactly > On Sep 22, 2023, at 1:03 PM, Stephen Illingworth > wrote: > >  > I've been thinking some more about this problem this week. I found the > performance prof