From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00
	autolearn=unavailable autolearn_force=no version=3.4.4
Path: 
 eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: Paul Rubin <no.email@nospam.invalid>
Newsgroups: comp.lang.ada
Subject: Re: How to get Ada to ?cross the chasm??
Date: Tue, 08 May 2018 23:49:10 -0700
Organization: A noiseless patient Spider
Message-ID: <87fu31coih.fsf@nightsong.com>
References: <1c73f159-eae4-4ae7-a348-03964b007197@googlegroups.com>
	<pcdagp$7pn$1@franka.jacob-sparre.dk> <87lgd1heva.fsf@nightsong.com>
	<pcg54t$gul$1@franka.jacob-sparre.dk> <87zi1gz3kl.fsf@nightsong.com>
	<pciqcd$ou1$1@franka.jacob-sparre.dk> <878t8x7k1j.fsf@nightsong.com>
	<fl7t0uF410rU1@mid.individual.net> <pcmjao$1soj$1@gioia.aioe.org>
	<fl8lgcF9dqgU1@mid.individual.net> <pcnem2$19mn$1@gioia.aioe.org>
	<fl938cFchk8U1@mid.individual.net> <pcnqgm$1tb4$1@gioia.aioe.org>
	<87k1sg2qux.fsf@nightsong.com> <pcov1b$1cvt$1@gioia.aioe.org>
	<87h8njmk4r.fsf@nightsong.com> <pcq81a$1n85$1@gioia.aioe.org>
	<87po27fbv9.fsf@nightsong.com> <pcrjv1$1jsj$1@gioia.aioe.org>
	<87h8nhwhef.fsf@nightsong.com> <flfjh7Fq18hU1@mid.individual.net>
Mime-Version: 1.0
Content-Type: text/plain
Injection-Info: reader02.eternal-september.org;
 posting-host="a15f8e13c869a4517dec92123a650d01";
	logging-data="28771"; mail-complaints-to="abuse@eternal-september.org";
	posting-account="U2FsdGVkX1+Rf+18nFGLMw0Y0xo7Zgf+"
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/25.3 (gnu/linux)
Cancel-Lock: sha1:BDpyS95/9OO/an77P/gzyhs/qJ8=
	sha1:O5E0D959tGI5fcxBMOzChKleH0Y=
Xref: reader02.eternal-september.org comp.lang.ada:52146
Date: 2018-05-08T23:49:10-07:00
List-Id: <comp.lang.ada>

Niklas Holsti <niklas.holsti@tidorum.invalid> writes:
> But the atomic instructions are usually much faster than a kernel lock.

They're not as bad as I thought.  I remembered timings like 100 cycles,
which on a superscalar machine is enough for several hundred
instructions.  But Agner Fog's optimization guide gives figures like 10
cycles for recent Intel cpus.  I'm quite impressed by that and wonder
what happens when multiple cpu sockets are involved.