Path: ...!2.eu.feeder.erje.net!3.eu.feeder.erje.net!feeder.erje.net!weretis.net!feeder8.news.weretis.net!reader5.news.weretis.net!news.solani.org!.POSTED!not-for-mail
From: Mild Shock <janburse@fastmail.fm>
Newsgroups: comp.lang.prolog
Subject: Re: Will WASM be fast some time? (Was: thank you for the FLOPs)
Date: Tue, 4 Mar 2025 10:13:40 +0100
Message-ID: <vq6g84$11g3t$1@solani.org>
References: <b406aa35-c39b-46f3-862f-1cc4b75143ae@googlegroups.com>
 <1b7ce2bd-722b-4c2e-b853-12fc2232752bn@googlegroups.com>
 <vb2jba$202pl$1@solani.org> <vc7mjb$9ptr$1@solani.org>
 <vcelid$b5f1$3@solani.org> <vihokt$lrea$1@solani.org>
 <vq6g0v$11g11$1@solani.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Tue, 4 Mar 2025 09:13:40 -0000 (UTC)
Injection-Info: solani.org;
	logging-data="1097853"; mail-complaints-to="abuse@news.solani.org"
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:128.0) Gecko/20100101
 Firefox/128.0 SeaMonkey/2.53.20
Cancel-Lock: sha1:zsKw3gdmEl1vsyS33BdShaji02E=
X-User-ID: eJwNy8EBwCAIA8CVakIAx0Eq+4/Q3v9EX95hLjeNJp8dBz19bhbdri52q0Db4IPElLEmctYbtSqPZfl/0CL9A1NEFTE=
In-Reply-To: <vq6g0v$11g11$1@solani.org>
Bytes: 3939
Lines: 65

Hi,

I made a test, you can run it from here, its CORS enabled:

:- ['https://www.dogelog.ch/module/bench/suite.pl'].

:- suite.

First a WASM Prolog, showing warm run:

nrev  % 2,995,946 inferences, 1.417 CPU in 1.417 seconds
crypt  % 4,169,360 inferences, 2.776 CPU in 2.776 seconds
deriv  % 2,100,600 inferences, 1.555 CPU in 1.555 seconds
poly  % 2,088,942 inferences, 1.681 CPU in 1.681 seconds
sortq  % 3,627,136 inferences, 1.923 CPU in 1.923 seconds
tictac  % 1,013,322 inferences, 2.659 CPU in 2.659 seconds
queens  % 4,599,283 inferences, 2.496 CPU in 2.496 seconds
query  % 8,645,933 inferences, 5.189 CPU in 5.189 seconds
mtak  % 3,946,569 inferences, 1.642 CPU in 1.642 seconds
perfect  % 3,243,474 inferences, 1.498 CPU in 1.498 seconds
calc  % 3,062,293 inferences, 1.791 CPU in 1.791 seconds
https://wasm.swi-prolog.org/wasm/tinker

Then a JavaScript Prolog, showing warm run:

nrev % Zeit 1413 ms, GC 0 ms, Lips 2127790, Uhr 04.03.2025 10:00
crypt % Zeit 1204 ms, GC 0 ms, Lips 3461555, Uhr 04.03.2025 10:00
deriv % Zeit 1270 ms, GC 0 ms, Lips 3094644, Uhr 04.03.2025 10:00
poly % Zeit 1068 ms, GC 0 ms, Lips 2336682, Uhr 04.03.2025 10:00
sortq % Zeit 1595 ms, GC 0 ms, Lips 2667571, Uhr 04.03.2025 10:00
tictac % Zeit 1596 ms, GC 0 ms, Lips 1087015, Uhr 04.03.2025 10:00
queens % Zeit 1718 ms, GC 0 ms, Lips 3322441, Uhr 04.03.2025 10:00
query % Zeit 2764 ms, GC 0 ms, Lips 3132399, Uhr 04.03.2025 10:00
mtak % Zeit 2129 ms, GC 8 ms, Lips 3936434, Uhr 04.03.2025 10:00
perfect % Zeit 1415 ms, GC 0 ms, Lips 3317768, Uhr 04.03.2025 10:00
calc % Zeit 1314 ms, GC 0 ms, Lips 2922571, Uhr 04.03.2025 10:00
https://www.xlog.ch/runtab/doclet/docs/04_tutor/basic/example01/package.html

Whats the cause of the discrepancy? 32-bit? Auto-Yield?

Bye

Mild Shock schrieb:
> Hi,
> 
> I wonder wether WASM will be fast some time.
> I found this paper which draws a rather dim picture,
> but the paper is already a little bit old:
> 
> Understanding the Performance of WebAssembly Application
> Weihang Wang et al. - 2021
> https://weihang-wang.github.io/papers/imc21.pdf
> 
> The take away is more or less that jitted JavaScript has
> same speed as WASM. With sometimes WASM more favorable
> outcome for Firefox than for Chrome and Edge:
> 
> Java Script     Chrome    Firefox    Edge
> D.1 Exec. Time (ms)    45.57    48.26    63.62
> M.2 Exec. Time (ms)    249.60    167.03    201.68
> 
> WASM    Chrome    Firefox    Edge
> D.1 Exec. Time (ms)    65.23    39.65    83.53
> M.2 Exec. Time (ms)    233.08    345.98    192.87
> 
> Bye