having some trouble

TheBaronTheBaron Austin, TX
edited February 2004 in Hardware
so i've been having some difficulty getting a good oc out of my system, for the most part i can't get any kind of oc out of the chip without massive voltages (is7, 2.8c). the ram isn't a problem as i have some geil pc4000 that i know for a fact runs rock solid all the way up to 250 as long as i keep with their timings (surprising isn't it?). so the question is where is my problem here? do i need to look into replacing the board or the chip, cause i mean i dont really have a problem with getting a new cpu... im always ready for an upgrade.

the cpu is 2xFAH stable at stock voltages / speeds, anything above that and the best i can do is prime stable, 2 instances of folding is a no go.

is there anyway i can test out the chip / mobo to find out where my problems lie?

Comments

  • madmatmadmat
    edited February 2004
    I was having problems with my IS7 in oc's with the ram at 1-1 but it was ok at 5-4 so I bagged the IS7 and went back to the Chaintech, the new board is fine at 245fsb with the ram at 1-1 and it's PC3700.
  • TheLostSwedeTheLostSwede Trondheim, Norway Icrontian
    edited February 2004
    TheBaron,

    Test the chip by using a divider. Try 6:5 since you know the memory runs good at 250 1:1
    300/6X5=250. Start at 240 in 6:5 mode. The cpu might need voltage.
  • madmatmadmat
    edited February 2004
    The problems I was having was that if I tried to run the ram 1-1 at 233fsb up I'd get artifacts in 3DMark '03 and in certain games but if I ran the ram at 5-4 I could get up to 259fsb easily with no artifacting so it led me to suspect the memory controller was the issue.
    I also found that increasing the Vdimm made memory errors pile up as more voltage was applied to the dimms and reducing the voltage made it go away.
    I found the optimal Vdimm for error control was 2.55v but I think the lower voltage was what was causing the memory controller to act so oddly.
    I quereyed the Ram Guy at Corsair about this and he said that in his mind it was the board that was at fault, the ram is tested at 2.75 Vdimm to be certified.
    I've been looking at the Abit forums and have seen other cases of this problem as well.
    I hope this helps.
  • TheLostSwedeTheLostSwede Trondheim, Norway Icrontian
    edited February 2004
    madmat wrote:
    The problems I was having was that if I tried to run the ram 1-1 at 233fsb up I'd get artifacts in 3DMark '03 and in certain games but if I ran the ram at 5-4 I could get up to 259fsb easily with no artifacting so it led me to suspect the memory controller was the issue.
    I also found that increasing the Vdimm made memory errors pile up as more voltage was applied to the dimms and reducing the voltage made it go away.
    I found the optimal Vdimm for error control was 2.55v but I think the lower voltage was what was causing the memory controller to act so oddly.
    I quereyed the Ram Guy at Corsair about this and he said that in his mind it was the board that was at fault, the ram is tested at 2.75 Vdimm to be certified.
    I've been looking at the Abit forums and have seen other cases of this problem as well.
    I hope this helps.
    Ohhboy, i was gonna suggest a vdimm mod and 3.55 volts. Good luck TheBaron.
  • muddocktormuddocktor
    edited February 2004
    Sounds like you might want to rma that board, TheBaron. Maybe something's fubar on the memory controller.
  • TheBaronTheBaron Austin, TX
    edited February 2004
    yeah i've been having a lot of trouble with the memory divider. but to respond to mackanz, using 3:2 or 5:4 doesn't help at all... the cpu is still the limiting factor. (the reason i know this is because 1:1 with just the geil and using any divider gets me the same max stable oc (210 @ 1.575v , which seems obscenely bad). i say "stable" in that i can pass prime95 at those settings indefinitely, but 2xFAH still causes lockups. memtest is fine at any fsb i can get a successful boot with. if it turns out i need a new board i'll buy a max3, screw waiting on an RMA
Sign In or Register to comment.