home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.benchmarks
- Path: sparky!uunet!stanford.edu!CSD-NewsHost.Stanford.EDU!Xenon.Stanford.EDU!hoelzle
- From: hoelzle@Xenon.Stanford.EDU (Urs Hoelzle)
- Subject: Re: A little malloc speed test.
- Message-ID: <hoelzle.716589006@Xenon.Stanford.EDU>
- Sender: news@CSD-NewsHost.Stanford.EDU
- Organization: CS Department, Stanford University, California, USA
- References: <1992Sep15.190106.24620@irfu.se>
- Date: 15 Sep 92 20:30:06 GMT
- Lines: 40
-
- bt@irfu.se (Bo Thide') writes:
-
- >I was curious to see the improvement in malloc speed between HP-UX 8.05
- >and HP-UX 8.07 on our Snakes (720 and 730) and ran the test below. It
- >seems that the 8.07 malloc compares favourably to the GNU malloc on the
- >HP and to the Suns I had available. I would be interested in
- >comparisons with other platforms and malloc's. For those systems that
- >support mallopt uncommenting (and possible modification of) the mallopt
- >line in the code can be elucidating.
-
- Try these:
-
- %T Custo-Malloc: efficient synthesized memory allocators
- %A Dirk Grunwald
- %A Benjamin Zorn
- %J CU-CS-602-92
- %D July 1992
- %X trace program, then generate customized malloc for it; faster than all
- standard methods on set of 7 large C programs
-
- %T Empirical measurements of 6 allocation-intensive C programs
- %A Benjamin Zorn
- %A Dirk Grunwald
- %J CU-CS-604-92
- %D July 1992
- %X useful data (size distrib., mean time between allocs, life times)
-
- %T The measured cost of conservative garbage collection
- %A Benjamin Zorn
- %J U Col Boulder CU-CS-573-92
- %D April 1992
- %X measures Boehm 1.6 (non-generational) against various mallocs; cache misses,
- page fault measurements. Overall result: pretty comparable (but ignores
- optimization problem)
-
- -Urs
-
- ----------------------------------------------------------------------------
- Urs Hoelzle urs@cs.stanford.EDU
- Computer Systems Laboratory, CIS 57, Stanford University, Stanford, CA 94305
-