Jump to content


  • Content Count

  • Joined

  • Last visited

Posts posted by xguy

  1. That is normal temp if you pushed your X2 that much I have the same setup, and I had to settle for 250 X10 (2500Mhz) to keep my temps in decent level, at 260x10 my temps started to rise to above 60C during dual prime, and would error as well .

    I found my sweet spot at 250X10 and that's what I'm staying at.

  2. Ok, I have started to overclock now, my first serious attempt ever to be honest, and so far im currently finding max cpu.


    I do have a couple of questions tho as im not sure of some stuff. What does PWM IC stand for? What part of the motherboard is that? And what kind of temperature would be max for this?


    And one last question before i go back to my idling infront of an stresstested cpu.... what kind of chipset temperature is preferable? mine is currently at 47C and its pretty much stable at that temp, should i try to lower this?




    Best regards



    EDIT: A while back one of my memory sticks broke down on me.. 512mb crucial ballistix. So I RMAd it to the store, but they didnt have any of those to replace mine so as it stands now they are gonna send me a 512mb crucial ballistix tracer memory. I want to run dual channel and i know same memory is kinda a prereq to do, will this become a problem for me running my memory, to sum up... dual channel with 1x 512mb Crucial ballistix, 1x512mb Crucial ballistix tracer?


    Thanks again.

    those temps are fine

    as far as memory I think you will have major problem running two different stick, people are having problmes with paired sticks nevertheless having two differnt types.

    hopefully some of the more experts can elaborate on that.

  3. Sounds like a ram issue to me, it happens to me when I'm playing with ram timing, it boots but then it won't go into windows just reboots again, so I have to go into bios and relax the ram.

    when you tried to installed windows seems to be corrupted, and I think that's a sign of ram errors, I would run memtest make sure your ram is fine. try one stick at a time maybe?

    good luck

  4. its got me thinking is sli even worth it .. i mean i know stasiscooling has a good 7800gtx cooler coming out which vents the hot air out of the case .. and will work in sli ..


    im currently using a lian li v1200b and its impossible to get fresh air up to the motherboard area .. because of the btx ..


    im thinking of ditching this case and switching over to like a coolermaster stacker or something .. and least with 2x120s in the front fresh air so go over the cards


    I have the same case like your however I have the plus version , which comes with the blower , so what I did I attached a zalman fan I had laying around to the blower , this way it circulates air around graphic cards and chipset (zalman blow/blower sucks out of the case) , and I get very decent temps, I manged to bring both my graphic cards down to 47C idle and 72-80C load (farcy seems to produce the most heat) chipset idles around 45C and never seen over 52C.

    I know it's not the coolest environment but I can live with that. see pic below to see the fan setup:




    If your case did not come with this blower it is available for purchase seperatly it helps a lot.


    good luck

  5. Well, I have tested with Rivatuner, and the cards clock up to 530, but even running at those speeds, my 3dmark scores dont change a bit. I havent tried downclocking them yet though. This is some wierd problem, I appreciate all of the suggestions and I am trying out as soon as I get the chance to; plz keep them coming.


    do yourself a favor and listen to me, try to compare your single and SLI in a different benchmark, try aquamark3 and see if you see performance improvment from single to SLI.

  6. That may be your card not liking those settings, not 3dMark. When a 7800 gets unhappy it downclocks itself. If I remember correctly, it has 3 clock settings. One for low intensity apps(like Windows), a medium and a high setting. When the card gets hot or has issues it is supposed to throttle itself back.


    Not saying that is what happens, but it may be.


    core slow down treshold is 115C and no sir I wasn't even near those temps, it is 3dmark and there was a post about on the nvforum as well, my cards would not exceed 75C when running 3dmark.


    The reason I know it is 3dmark and not gpu cause aquamark is very happy with those settings and I went beyond 95K score.

    I am 100% sure it is 3dmark.

  7. if you wanna OC to 2500 and above I recommand the ram in my sig, they're the sweetest things I've dealt with , they run at very low voltage you can take them up to 260mhz easily, my culprit was the CPU not the ram, I can run @2600 on air no problme but prime95 would error, therefore I would'nt call it stable, even though I can run dual superPI and OCCT no problem at those speed and I can play games for hours, with no crashing , but I think peace of mind is wrth 100Mhz set back, so I'm happy with 3500Mhz 100% stable and very cool 30C idle and 49C load (running dual prime).


    here's the ram I'm talking about:



  8. I had some interesting experience with 3dmark05, when I OC my graphic crds the scre will drop drastically , then I tested with different OC increments, and found that 3dM05 likes certain spot and dislike others , like if I go up to 480MHZ on gpu it likes that and I get higher score, then I go to 485mhz and it drops like 1000 to 2000 points, strange behavior of 3dmark, so I stopped worrying about it.

    try to downclock your cards to 450/1200 I know you the oc version, and see if 3dmark gives you around 10500, then this should be about right. And then you would know it's the 3dmark bug and not your system.

  9. I'll be interested to hear this as well. I'm reading it as actual value determined through monitoring, approximately base times percentage multiplier.


    I am trying to zero in on my max OC for my 4400 at the moment (on the slow train averaging about 1-2 tests a night). Memory timings at the moment are at 2.5-3-3-7 @ 2.7V not at the ones indicated below; I wanted to loosen things up during CPU testing. Other setting are as specified in the Overclocking Guide sticky.


    I was using SuperPI with stock voltage until I started hitting errors at around 250MHz x 11. I then bumped voltage up to 1.3 and +123% which still led to errors. I also tried 1.4 and +113% which also had errors. Both cases, the CPU hit 51C at its highest temp. I then scaled back to 244Mhz x 11 with no errors at 1.4 and +113%. I'm wondering if I should be running two simultaneous instances of SuperPI, and if doing so will cause execution of one instance on each core. I'm guessing that SuperPI is not programmed to use both cores simultaneously so likely is only executing on a single core. I hope it was running on the core with the lower OC potential; otherwise, I may have to spend quite a bit of time with dual instances of prime95 to get to a dual core stable OC.


    In any case, I ran OCCT at the 244Mhz x 11. OCCT hit a driver issue running under XP Pro 64-bit, which makes me wonder (only a little :)) about its accuracy on testing stability under this OS. Does anyone know if there is a 64-bit version of this software available? Anyway, OCCT indicated instability at about 10 minutes into the 30 minute stress test.


    I'm wondering if I should bump the voltage base or multiplier at this point to see if I can take it higher or reach stability at the current settings. What are your views on the max Vcore for these 4400's?




    I dunno about OCCT and XP64, but I know you can run 2 instances of superPI, but you have to create two different directories for superPI files, and launch them from their directories accordingly, then before you hit the calculate buttons, set the affinity using windows task manager, set one of them to CPU0 and the other to CPU1.


    you will aslo see that your temp will go much higher that way on air. Also you should test with prime95 running two instances as well, from my experience @2600MHZ I passed OCCT and superPI(2 inst) with flying colors, but prime95 would fail unless I pushed the voltage too high, which resluted to temps nearing and trying to pass 60C, forget it I'm not ready to fry my cpu, and that's why I settled with 250X10 @2500MHZ stable in all tests at sweet low voltage (see my sig) running nice and cool 30C idle and 49C full load with 2 primes running.

  10. this is interesting i tried a different PSU and it just gave 4 led lights and no i cant even enter the bios

    I had the same problem happened to me, and when I reset cmos , nothing happens at all, I get to one light and it freezes at first screen showing cpu info and ram info, and I can't even turn it off with button I had to switch PSU off. If that's the smae thing happening to you do this, but do it at your own risk, I've done it many times and it works usually this happens to me when I'm messing around with ram timing.


    shot system off

    disconnect power supply and wait for the led light to go off.

    take out battery

    switch the cmos jumper to 2-3 position

    with battery off and jumper in 2-3 position start the system


    it will start up and even load the window that way, but you can enter bios and set default settings, save and exit then you can power down your system, put your battery back on and switch the jumper back to normal.


    Hope that works for you, it works for me every single time. I actually made a post about it asking how come this works but I got no feed back from the guru's here:




    Good luck

  11. Well first I would test the ram seperatly at that speed , to eliminate the ram factor, second maybe your processor is not ocing well, clock it down and test again, or you may need more voltage but then that's gonna run too hot on air.

    I ran into the same problems ocing my 4400+ at 2600mhz evertyhing runs fine except prime95 sometimes gives me errors sometimes it doesn't I had to up the voltage way high to run it stables, so I decided to clock it down to 250x10 which run stable and cool on my XP-90. What I mean by stable is running two instances of prime95 for hours, two instances of superPI and OCCT. prime95 is your ultimate test cause @2600mhz I had no problem with OCCT or superPI yet prime95 would error after an hour or so.

    So if you're not passing OCCT, my guess is you won't pass prime95 for sure, and you want to do that before you determine that your PC is stable.

    In other words you need to find your sweet spot, whether you up your voltage or clock down a little, it took me two months to do so, just keep trying till you get stable.

    Happy OC :)

  • Create New...