2 CPUs vs. 1...

Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
edited December 2003 in Folding@Home
Would 2 750MHz P3s fold faster (since they're doing 2 WUs at once) than 1 1500MHz Celeron, or would it make any difference?

Comments

  • shwaipshwaip bluffin' with my muffin Icrontian
    edited December 2003
    Is the celeron a p4 celeron?
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited December 2003
    Overclocked 1.3GHz Celeron, so it's a P3-Tualatin on a 100MHz bus, basically.

    The Tualatin P3s (not the -S ones) and the Tualatin Celerons are basically identical, excpet for the bus speed.
  • shwaipshwaip bluffin' with my muffin Icrontian
    edited December 2003
    They would fold at the same rate.
  • qparadoxqparadox Vancouver, BC
    edited December 2003
    yeah should be pretty darn close, you'll also save on the power bill a bit with the single IIR my p3 power disappation tables correctly ;). Should be about 30 Watts less, which does add up.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited December 2003
    You guys are taking into account the fact that F@H is NOT smp-capable, right?

    The P3-750s would take about 2x as long, but they'd be doing 2 WUs at once, too...
  • SimGuySimGuy Ottawa, Canada
    edited December 2003
    Folding is SSE-enhanced. When running an SSE-enabled application, SSE advantage scales with faster clock speed.

    Therefore, the Celeron 1.3 @ 1.5 should fold slightly more efficiently and faster than the dual 750s.

    One thing to keep in mind though... the Celeron only has 128 KB of L2 while the P3's have 256 KB (AFAIK), so that speed/efficiency advantage that the Celeron 1.5 has may be lost to the Dual P3's.

    Straight answer? Probably in and around the same area, give or take a wee bit.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited December 2003
    Ok. I'll leave it alone for now, then. I'm getting restless... I have nothing to tweak (all my computers work *gasp*) so I'm trying to find something I can screw around with... :rolleyes:

    BTW, simguy, the Tualatin-core Celerons were P3s on a 100MHz bus; they had a full 256k of L2 cache.
  • SimGuySimGuy Ottawa, Canada
    edited December 2003
    Geeky1 had this to say
    Ok. I'll leave it alone for now, then. I'm getting restless... I have nothing to tweak (all my computers work *gasp*) so I'm trying to find something I can screw around with... :rolleyes:

    BTW, simguy, the Tualatin-core Celerons were P3s on a 100MHz bus; they had a full 256k of L2 cache.

    I know the restlessness feeling. Everything here is working too good. I need something to explode/fail so I can troubleshoot. I'm getting ancy :D

    Humm... then no doubt, the 1.5 Tualatin Celeron with the 256 KB L2 will fold faster than the 2 750's.
  • GargGarg Purveyor of Lincoln Nightmares Icrontian
    edited December 2003
    The Tualatin is a better core, too. That should give it a slight edge per clock.
  • mmonninmmonnin Centreville, VA
    edited December 2003
    After reading the last few posts I guess the Celly would fold a lil bit faster. Not much tho. Plus it would take less power to run it like qparadox mentioned.
Sign In or Register to comment.