Front, Kernel, Security, Distributions, Development. See your byline here on LWN.net.

Create an account

Subscribe to LWN

LWN.net Weekly Edition for February 16, 2012

Book review: Open Advice

Linux support for ARM big.LITTLE

LWN.net Weekly Edition for February 9, 2012

XBMC 11 "Eden"

Printable page

Weekly edition	Kernel	Security	Distributions	Contact Us	Search
Archives	Calendar	Subscribe	Write for LWN	LWN.net FAQ	Sponsors

Thanks Ulrich and Jon

Posted Oct 1, 2007 19:31 UTC (Mon) by nix (subscriber, #2304)
In reply to: Thanks Ulrich and Jon by bwarren
Parent article: Memory part 2: CPU caches

I found the simplicity of the necessary testcases particularly
interesting. Over and over again I was kicking myself: I've often wanted
to know the sizes of the various levels of cache, for instance, and never
once thought of running a trivial little benchmark.

(Of course *interpreting* the results of that benchmark requires all sorts
of knowledge which I didn't have. Thanks to this article I am not so
clueless now.)

(Log in to post comments)

Such benchmarks were known for years!

Posted Oct 1, 2007 21:02 UTC (Mon) by khim (subscriber, #9252) [Link]

For example on Digit-Life:

So detailed investigations of CPU internals with simple benchmarks are not so interesting - the detailed explanations related to programming should be...

Such benchmarks were known for years!

Posted Oct 2, 2007 20:48 UTC (Tue) by nix (subscriber, #2304) [Link]

Yeah, they were known for years (everything Ulrich talks about here was
known for years; it's a human creation, after all, not a discovery.)

But *I* didn't know it. Thanks to this article, now I do. (The later
chapters promise to get even more interesting, as you said.)

(And those links look really cool. Thanks.)