home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!sun-barr!cs.utexas.edu!swrinde!gatech!hubcap!ncrcae!ncrlnk!torynews!martins
- From: martins@TorreyPinesCA.ncr.com (Martin Sandman 3562)
- Newsgroups: comp.benchmarks
- Subject: Looking for MSYNCS
- Summary: Millions of Synchronization Per Second Benchmark
- Message-ID: <1992Sep1.181700.5501@TorreyPinesCA.ncr.com>
- Date: 1 Sep 92 18:17:00 GMT
- Organization: NCR (Torrey Pines Development Center)
- Lines: 21
- Disclaimer: This posting does not neccessarily reflect the opinions of NCR.
-
- I have seen a vidio tape in which Harld Stone proposes that a useful
- standard measure of multiprocessing computer performance is MSYNS per
- second, that is, Millions of Synchronizations per second. He explains
- why MSYNCS establish the speed up limit for the number of processing
- elements. He makes a case for why test-and-set-lock would be slow
- and cause memory contentions. Then he introduces the Fetch-And-Add and
- explains how FAA can be executed in parallel in a combining interprocessor
- connection network.
-
- Does anyone know of a benchmark designed to measure synchronizations
- per second? Please send me email and I will summarize. Thanks.
-
- -- Martin
- --
- ======================================================================
- = Martin D. Sandman ......... =
- = NCR Torrey Pines ..x...... =
- = 11010 Torreyana Road martins@sandiego.ncr.com ....x.... =
- = San Diego, CA 92121 Phone: 619-597-3562 .xx..xxx. =
- = U. S. A. FAX: 619-597-3501 ......... ants =
- ======================================================================
-