[Beowulf] RASM : random memory latency test
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
Greg Lindahl lindahl at pathscale.comMon Jul 4 13:40:52 PDT 2005
- Previous message: [Beowulf] RASM : random memory latency test
- Next message: [Beowulf] Bonding-alb and NAS performance
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Mon, Jul 04, 2005 at 09:04:20PM +0200, Vincent Diepeveen wrote: > Latencies are NOT dependant upon memory frequency, just dependant upon the > speed the RAM works at (which sometimes IS related to cpu frequency in case > of opteron as the faster opterons allow 400Mhz ram. This is simply wrong -- on the Opteron, if you hold the memory frequency constant, the memory bandwidth and latency improve by a modest amount as you increase the CPU clock. This test does have the advantage that it uses all the cpus, which is more realistic than using just one. -- greg
- Previous message: [Beowulf] RASM : random memory latency test
- Next message: [Beowulf] Bonding-alb and NAS performance
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list