Path: csiph.com!news.mixmin.net!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Tim Rentsch Newsgroups: comp.lang.c Subject: Re: filling area by color atack safety Date: Tue, 09 Apr 2024 02:32:31 -0700 Organization: A noiseless patient Spider Lines: 31 Message-ID: <86y19m285s.fsf@linuxsc.com> References: <86h6h3nvyz.fsf@linuxsc.com> <865xxiok09.fsf@linuxsc.com> <20240319131842.00002138@yahoo.com> <20240319191859.00004bc8@yahoo.com> <86bk79m10l.fsf@linuxsc.com> <20240324202758.00001b75@yahoo.com> <86frwfjs0n.fsf@linuxsc.com> <20240325012844.0000685b@yahoo.com> <867chlk1zf.fsf@linuxsc.com> <20240329162141.00006c81@yahoo.com> <86y1a0i4tp.fsf@linuxsc.com> <20240330211506.00000b86@yahoo.com> <20240331115438.00003f3c@yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Injection-Date: Tue, 09 Apr 2024 09:32:33 +0200 (CEST) Injection-Info: dont-email.me; posting-host="47558c70f7e061adbec47d824d7ff660"; logging-data="163299"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1/OKJdBin6S3J3EjIVdmHNGhCdSCkBdMlw=" User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.4 (gnu/linux) Cancel-Lock: sha1:0AJMGrOeQJjVi0CvkcPweWxt2qU= sha1:fPfwhGsvwXV2/BLu+R9Feaxgwoo= Xref: csiph.com comp.lang.c:384230 Michael S writes: > On Sat, 30 Mar 2024 21:15:06 +0300 > Michael S wrote: > >> On Fri, 29 Mar 2024 23:58:26 -0700 >> Tim Rentsch wrote: >> >>> One, the new code is a lot more complicated than the previous >>> code. I'm not sure the performance gain is worth the cost >>> in complexity. What kind of speed improvements do you see, >>> in terms of percent? >> >> On my 11 y.o. and not top-of-the-line even then home PC for 4K >> image (3840 x 2160) with cross-in-cross shape that I took from one of >> your previous post, it is 2.43 times faster. >> I don't remember how it compares on more modern systems. Anyway, right >> now I have no test systems more modern than 3 y.o. Zen3. > > I tested on newer hardware - Intel Coffee Lake (Xeon-E 2176G) and AMD > Zen3 (EPYC 7543P). > Here I no longer see significant drop in speed of the 1x1 variant at 4K > size, but I still see that more complicated variant provides nice speed > up. Up to 1.56x on Coffee Lake and up to 3x on Zen3. On my test system the numbers are closer and also more evenly balanced: ratios range from about 0.70 to about 1.40, roughly evenly split with the 2x2 version somewhat better. There was one outlier at approximately 1.48. More precisely, the ratios have an average of 1.06 (which means the 1x1 version is about 6 percent slower on average), with a standard deviation of 0.21.