Path: csiph.com!eternal-september.org!feeder.eternal-september.org!.POSTED!not-for-mail From: John Ames Newsgroups: comp.os.linux.misc Subject: Re: Hint For GNU/Linux Progrmmers Date: Thu, 11 Sep 2025 09:30:36 -0700 Organization: A noiseless patient Spider Lines: 18 Message-ID: <20250911093036.00006328@gmail.com> References: <109uoqs$2n3c4$1@dont-email.me> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Injection-Date: Thu, 11 Sep 2025 16:30:40 +0000 (UTC) Injection-Info: dont-email.me; posting-host="3bb2d29411dda731a1fc1f1809338ba0"; logging-data="2836364"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX194ghDDUNV3eJVVb1rUOodIEkwtD4iA84E=" Cancel-Lock: sha1:cAXsXqMwT/lviF3iRxXnDi05/OA= X-Newsreader: Claws Mail 4.3.0 (GTK 3.24.42; x86_64-w64-mingw32) Xref: csiph.com comp.os.linux.misc:73898 On Thu, 11 Sep 2025 15:11:24 -0000 (UTC) John McCue wrote: > I find O1 is good enough for all programs I create. > > To, me, testing and retesting different optimizations is a > huge waste of time and at most you might save 1 second :) Even if you're an optimization freak (and there's nothing wrong with that,) the efficacy of tweaks like loop unrolling is highly dependent on machine particulars (cache size, etc.) - it's difficult if not impossible to establish a one-size-fits-all recipe for True Optimum Performance that could be handed out to non-freaks, as is being suggested here. Some level of tweaking may be warranted (e.g. unrolling loops in a way that suits the particular algorithm,) but there's little point trying to generalize deep grease-monkey fine-tuning across *all target systems* even for a single distro, let alone The World At Large.