Can memory run async on P4s?

GHoosdumGHoosdum Icrontian
edited July 2004 in Hardware
Is it possible to run PC2700 DDR in dual-channel mode on a 2.8GHz P4 with the 800MHz FSB?

How about running PC3200 in single channel mode?

Comments

  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited July 2004
    Yes and yes.
  • JimboraeJimborae Newbury, Berks, UK New
    edited July 2004
    On my IS-7 I can have the memory runing at 166fsb & the cpu at 250fsb, it still seems to fold fine. :)
  • GHoosdumGHoosdum Icrontian
    edited July 2004
    Sweet. Thanks guys.
  • Omega65Omega65 Philadelphia, Pa
    edited July 2004
    Run the Mem:CPU at 5:6 (166:200) or lower

    5:6 (166:200) (PC2700)
    4:5 (160:200)
    3:4 (150:200)
    2:3 (133:200)
    1:2 (100:200)
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited July 2004
    I am running PC3200 Corsair L2PT on the Prescott, and PC2700 RAM on the Northwood right now. The L2PT is running at DDR450 settings, synchronous (If I had not pumped the nominal 2.8 GHz Prescott to a 225 base, I did and could run the RAM FASTER than the CPU). The Non-Dual Channel (not stated as Dual Channel) is at async settings, but actively used as dual channel (timings and OCability of RAM determine what you can do, not pure JEDEC timings).

    Both boxes have IC7-Max3 boards in them, and the Prescott, to OC where I have it, runs fine-- but, on the IC7-Max3 I had to strap the NB to DSP800 and not tie it to CPU base. Strapping NB to DSP800 took a box that kept LINCS erroring folding in one of two consoles to STABLE folding-- with two folding consoles, on XP O\S base, on a very heravily used Windows box. Since NB fan than also ran slower, I conclude the BRIDGE was OHing, NB or SB or both. If you have an unstable mobo, try strapping the bridges to an async that is equal to nominal base of CPU, see if it goes stable suddenly and stays stable for a week or so. Leave NB strapped, play with CPU timings further, see if still stable, if so then play with RAM timings.

    I guess what I am saying, is RAM might not be breaker for an unstable P4 box, might be bridges OHing instead. But RAM that can handle the resulting speed at Low Latency can indeed be OC'd async to both CPU and bridge rates, AND ASYNC to PCI and AGP also (both boards have PCI and AGP set to fixed base rates for those busses, PCI at 33 MHz, AGP at 66 MHz). CPU can be async to AGP, CPU can be async to PCI, CPU can be async to RAM, CPU can be async to bridge hyperbus to a degree also (I do not know the limits, right now the bridge bus is async to CPU like this on Prescott box:

    CPU base:NB strap 225:200

    On Northwood, the base is very close to nominal, I used multiplier to OC the CPU. On Prescott I used BUS BASE OCing almost totally, the Prescott is limited more as to how much multiplier you can push at it, but can be BASE RATE OC'd.

    Note, I am running BIOS 15 on both IC7-Max3's right now.
Sign In or Register to comment.