Tweaks please.

botheredbothered Manchester UK
edited March 2004 in Hardware
My spec is as below but I've just added an xp3200 (400) and changed the memory for 1G pc3500 DDR433. In soft menu the cpu is down as 2500+, external clock 166MHz, multi 11.0. I've set it to turbo and it shows as 6.2.2. I had 5.2.2 before. I've just ran 3dmark 2001 and got 14307 (from 11000).
What do I tweak to get more guys?
«1

Comments

  • ShortyShorty Manchester, UK Icrontian
    edited March 2004
    Your CPU is wrong in sofmenu. A 3200+ should be set as 3200+ with clock of 200mhz. You are loosing plenty of performance there!

    If it doesn't give you the option for 3200+. Then set multiplier to 11x, clock to 200mhz. Then reboot and run 3dmark again :)
  • GHoosdumGHoosdum Icrontian
    edited March 2004
    Looks like your processor needs to be bumped up to stock speed! FSB should show as 400 for you, not 166. //edit :doh: Shortzter beat me to it!

    One tweak that might help you out is changing the 5/6 number (TRAS?) to 11 in BIOS - it's counterintuitive but for some reason it works on the nForce-2 chipset.
  • botheredbothered Manchester UK
    edited March 2004
    I've twice set it to 3200+, just as windows finishes loading it crashes. Set it back to 2500 and it's ok. When I ran 3dmark and went on system info it does say it's an xp3200, it said so on the box and the proc. I'm doing a clean install of windows in the next day or so, maybe that'll help? Still on that repair install.
  • mmonninmmonnin Centreville, VA
    edited March 2004
    Lower the timings. And then try again.
  • botheredbothered Manchester UK
    edited March 2004
    Set every page to optimal defaults except cpu speed. It's just the same.
  • botheredbothered Manchester UK
    edited March 2004
    The cpu is shown as running at 1.8GHz. And there's this.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited March 2004
    Weird. Try going into the BIOS, have it load fail-safe defaults, change the CPU FSB and Ratio to auto, reboot, and see what it does. If it still crashes, I suggest you take the heatsink off the CPU and make sure that you really got a 2500. AMD has been known to screw up on packaging retail CPUs once or twice...
  • botheredbothered Manchester UK
    edited March 2004
    Fail safe, optimal, it's all the same, if I set it to 3200+ the pc crashes. Looks like it's coming out again tomorrow. Why can I never plug something in and have it work?
  • botheredbothered Manchester UK
    edited March 2004
    Just looked at Abiteq, all the voltages look ok except the 12v is a little low @11.7v. I know that's nothing to do with the cpu but could it mean I'm running out of power? the PSU is 350w amd approved.
  • pcscustompcscustom Oklahoma
    edited March 2004
    Surely that Nvidia isnt right, 50C Idle?!?

    Trev
  • ThraxThrax 🐌 Austin, TX Icrontian
    edited March 2004
    It reads from the mobo. So yes, it's probably right.
  • SpinnerSpinner Birmingham, UK
    edited March 2004
    It does sound like you've got a 2500+ in there instead of a 3200+. Need to confirm that. You've got the latest BIOS flashed right? 22?
  • Omega65Omega65 Philadelphia, Pa
    edited March 2004
    Bothered you need a Better PSU and Heatsink. At 2.2ghz the CPU is pulling too much power for the PSU
  • leishi85leishi85 Grand Rapids, MI Icrontian
    edited March 2004
    Thrax wrote:
    It reads from the mobo. So yes, it's probably right.

    u mean it's the die temp, instead of the surface temp right??
    if that is the case, no wonder my temp has gone up since i changed from my epox 8rda+ to my nf7-s.
  • SpinnerSpinner Birmingham, UK
    edited March 2004
    bothered wrote:
    ...the PSU is 350w amd approved.
    Yeah and so are the dozens of burnt out PSU's of got lying around in my loft, but that didn't stop them from lasting less than a few months.:wink:

    Is your PSU generic, i.e it came with the case? Or did you buy it seperately
    ? if so what make and model?

    Omega's right though, whether it's a top quality 350w or an overloaded 350w standard, 400w should really be the bare minimum when running one of those bad boys. Whether the PSU is the root of your problems or not still remains to be seen, but spending an extra few bob on a new PSU would be recommended. But then I suppose you've just blown your last wad of cash showing your wife how much of a stud you are. ;D

    Let us know about that CPU though mate. :)
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited March 2004
    whether it's a top quality 350w or an overloaded 350w standard, 400w should really be the bare minimum when running one of those bad boys

    Hardly. Want to know what I've run off a 350w PS?

    ABIT NF7-S v 2.0
    1800+ JIUHB DLT3C @ 2.4GHz
    ATi Radeon 9700 Pro
    1GB PC3500 (Mushkin Level II)
    2 7200rpm HDDs
    2 Optical drives
    Floppy
    4 case fans + CPU fan, 60mm fan on video card

    A good 350w (Antec, Enermax, PC Power & Cooling, Fortron Source, etc.) is more than sufficient for any single-CPU machine that doesn't have an insane amount of drives (like, say, 6 hard drives...).

    In fact, according to APC's software, the server at habitat for humanity (A7N8X 2.0, 1800+, 2 40GB 7200RPM drives in RAID 1, GF2MX400, CD-ROM, Floppy) draws 210w running F@H, and that's with an ancient, crappy HP monitor that draws 110w by itself.
  • SpinnerSpinner Birmingham, UK
    edited March 2004
    Geeky1 wrote:
    Hardly. Want to know what I've run off a 350w PS?
    Well yes of course. I wasn't trying to put a staple on the 400w minimum thing. Hell I run plenty of top-end rigs with 350W PSU's or even just with 300W PSU's. But on a basic level of advising people, I now say 400W as a 'you should aim for' minimum, simply because it gives the cheaper PSU's room to breath, ones which can't cope very well under load (load being what? 75%). Also 400W PSU's tend to be the crossover point between good quality motherboard power feeds and bad quality. Sure 300W will do fine if you are sensible with what you put on it, and sure a 350W will do fine aswell, but when in doubt, go for more clout! Which in this case resulted in me recommending a 400W PSU for me old mate Bothered.

    But yea, fair point.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited March 2004
    But on a basic level of advising people, I now say 400W as a 'you should aim for' minimum, simply because it gives the cheaper PSU's room to breath, ones which can't cope very well under load

    Gotcha. I agree with you on that, for sure. It all depends on the quality of the PS. One of the older cad/cam machines at work (1.4GHz TBird, 512MB DDR, GF3, NIC, 40GB HD, CD, Zip, Floppy) runs quite happily on a 250w Antec power supply. BUT, it's an Antec power supply, too. I wouldn't run that system off a 250-350w (or even 400w) generic PSU if you paid me. I bought a CHEAP ($20) "550w" PSU to use as a bench/test PSU. I put it under load outside of a computer to make sure it could handle my systems. Know what? At ~300w, the +12v line was 11.3v. It lasted for all of 10s or so before going *poof*. 550w my ass. :rolleyes:
  • edited March 2004
    After sitting here ruminating on the subject all day and a goodly portion of the night; you're showing 50c at idle with the CPU running as a 2500+ which makes me wonder if what's happening is that the CPU's overheating as windows loads as it's being loaded pretty hard at boot up.
    50c at idle as a 2500+ means that it should be getting up close to 60c under load as a 2500+ and I could see it getting up to 65c or higher under load as a 3200+.
    These are just my personal thoughts on the matter.
  • botheredbothered Manchester UK
    edited March 2004
    Bios. I asked on the Abit forum and have looked at bios's. 2.0, apparently, is fine for what I have. 2.1 just updates a SATA raid bios and 2.2 has support for NF7-V.

    PSU, I don't know. It came with the case but is 350w amd approved.

    I think I'm going to take the cpu out, again!, and take it back to Aria and let them test it. They have a nf7s test rig. If it works in there I'll pick up another psu.
    Thing is, whatever I set in bios, fail safe, optimal, it's always identified as 2500. If then the only thing I change is 2500+>3200+ it crashes. On one of these settings a blue windows screen opens telling me windows has shut down to prevent damage to your pc. Once or twice @3200 windows loads but when the last few icons start to appear in the task bar everything freezes, reset button time. It has never run @3200.
  • edited March 2004
    CPU's will get flaky when they overheat...what HSF are you using? That 50C really has me worried, that's getting high for a load temp and if it's 50C at idle that's really bad news.
  • botheredbothered Manchester UK
    edited March 2004
    When I pick it up I also wanted a Thermaltake volcano 12, which they didn't have. They recommended the volcano11. It's a big copper hs with a thermaly controled fan, I probably need a case fan as well.
  • edited March 2004
    yeah both would be a good idea.
  • gibbonslgibbonsl Grand Forks AFB
    edited March 2004
    what version of the board do you have

    the 2.0 is the only version that OFFICALY supports the 400 fsb CPU

    http://www.abit-usa.com/products/mb/products.php?categories=1&model=6

    if you have the older version the chipset is not the NV 2 ultra which does 400 in it sleep

    so of the older chips can do it and some cannot
  • gibbonslgibbonsl Grand Forks AFB
    edited March 2004
    try colling down the northbrige chipset

    that might help
  • mmonninmmonnin Centreville, VA
    edited March 2004
    Stock cooling on the northbridge can go past its rated frequency which he cant even get to atm.
  • GHoosdumGHoosdum Icrontian
    edited March 2004
    Hmm... What does the VCore read from the sensor? If you're dropping below 1.65 actual volts, that might be why your CPU can't reach 3200+ speed... the high temps aren't helping much, either!
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited March 2004
    bothered wrote:
    The cpu is shown as running at 1.8GHz. And there's this.

    Case FANS, plural. My Barton runs 12-13 C above case under load NORMALLY, when case gets to (or hotter than) 31 C it goes to 47-50 C. At 32 C CPU runs 46-47 C. At 26-27 C, it runs at 40-41 C. In other words, when heat is not vented out of case fast enough, and cool air drawn in, the temps rise drastically at CPU. It would die if case got to 36 with the temp increase curves I see NOW. In fact, the motherboard monitor is set to kill the box at 36 C case temp.

    Couple notes: I have about 145 CFM air flow in just case fans-- two 80 mm in fans at about 32 CFM each, two 80 mm output rear fans at about 40 CFM each (all Panaflos). CPU fan adds about 40 CFM (39.6 exactly, it is a Panaflo). HS is a SLK-900 with a very thin layer of Ceramique HS compound. Too much HS compound will cause CPU temps to rise too high relative the HS. Note, AS5 STINKS. AS4 or Ceramique are better, much better, for compound. PSU id a dual fanned PSU, it runs about 50 CFM out of the case as the fans are not high-flow on it but is stable so I have not R&R'd the fans. I am discounting the one IDE Cold-Swap bay fan that is running all the time, it is about 15 CFM and is a very high flow fan for a 40 mm fan. I am discounting the NB fan also, about same flow.

    Definitely a case Overheat thing as primary culprit here. 99% sure. Southern climate here gives me a natural heat curve and delta analysis lab.... :D

    John D-- who actually thinks that unless you have the computer right next to heatduct or wood stove or fireplace or some heat source, where you are, case should be 10 C lower with proper venting and cooling and then CPU would be VERY happy. One last idea, seen this done (BIG MISTAKE for first, second better)-- take an old desk with solid back panel, put case under desk, you get a massive heat trap effect and computer will not cool right. Put case NEXT to desk, wastebasket instead UNDER desk, you get a very happy computer as heat is not trapped under desk.
  • SpinnerSpinner Birmingham, UK
    edited March 2004
    Okay okay, first things first Bothered mate, you need to confirm you've actually got a 3200+. If you've got a 2500+, then that would explain everything.

    Heat may be the problem, but the CPU itself won't cause stability problems at stock untill well into the 80's. The problem you may be encountering, is that the BIOS is configured to turn the PC off when it reaches a certain temp, that may be set too low currently for the temps your CPU is running at. So find those settings in the BIOS and raise them and see if it makes any difference. Don't worry even the most relaxed settings in this department will shut your PC down before it goes nova.

    I actually had a similar problem recently with a stock heatsink on a 2800+. It just wasn't cooling the CPU properly, couldn't figure out why, I swapped it out for another one, and it worked fine. Are you using the stock thermal paste that comes on the bottom of the heatsink? You should.

    However, reading one of your above posts, you shouldn't experience freezing or BSOD's in the event of an safety shutdown. If it were an overheating related stability problem, the computer should shut it self down before the the system gets to the point of instability. Which still makes me think your CPU isn't what you think it is. Nevertheless...

    Either way Bothered please confirm what CPU you've got. It will be written on the CPU.
  • botheredbothered Manchester UK
    edited March 2004
    Long story but the end result was it seems it may have been a motherboard fault. I got a replacement and blew that up so I got another. The cpu is running correctly now but the memory is suspect. I tried both the memory tests in the download section, memtest31 is too complicated, I don't want to screw anything else up, memdoc won't run, it copies to a floppy, I restart and the pc tells me to remove the disk and press any key, the pc then reboots as normal.
Sign In or Register to comment.