![]() |
|
|
#1 |
|
Dec 2002
2×3 Posts |
Hiya, Xyzzy refered me here to see what I can do to fix the errors I get in prime95.
First, here's my system specs: Antec True430 AthlonXP 2000+ (AX7 w/ Panaflo L1A cooling) Asus A7N266-C 512mb Mushkin PC2100 HiPerf CAS2 Geforce4 ti4600 LSI Ultra160 SCSI controller A little info on the system: WinXP Pro SP1 (fresh build less than a month old) not overclocked in any way MBM reports CPU temp below 50C at max load Passed all tests in memtest86 for 24 hrs. Prime95 v22.12 So, I've run Prime95 three times now and each time I get the same error at approximately the same time: little over 4 hours. This is the error i get: FATAL ERROR: Rounding was 0.4999xxxxxx, expected less than 0.4 Each time, the "x"'s represent different numbers. So what could it be? =( |
|
|
|
|
|
#2 |
|
Aug 2002
207228 Posts |
Have you tried with just one stick of mem, no SCSI and with a PCI video card? I always try to start as simple as possible to minimise the variables... You might also try CAS 3 and you might check for a BIOS update... You can get an old PCI video card for free if you look a little... People just throw them away...
Although... In *my* experience, you either get errors right quick (cooling) or later randomly (memory)... You *are* running the torture test, right? |
|
|
|
|
|
#3 |
|
Dec 2002
2·3 Posts |
I'm running only 1 DIMM. Everything in my system is using the latest WHQL drivers and BIOS's are up to date. It's my primary system so I'd rather not replace my AGP video card with a PCI one. I could try lowering my memory to CAS3 but again, I ran memtest86 v3 on all tests (not standard) for 24 hours without any errors.
Yes, I'm running the torture test. I just find it really odd that I get the same error at approximately the same time each time but the test it errors on is different each time. |
|
|
|
|
|
#4 |
|
Aug 2002
101 Posts |
again i would say the cpu is at fault.
call amd. they will give you an rma number right away. just tell them fpu fails. you ship it back. they vverify it. new cpu sent! |
|
|
|
|
|
#5 |
|
Dec 2002
616 Posts |
It's an OEM cpu that's a few months old. Will I still be able to RMA it?
|
|
|
|
|
|
#6 | |
|
Dec 2002
1102 Posts |
Quote:
|
|
|
|
|
|
|
#7 |
|
Aug 2002
2×32×13×37 Posts |
Well, to isolate the CPU just borrow a different one and see if it fails too... If it does, then the CPU is most likely not the culprit... You can do the same with the memory...
|
|
|
|
|
|
#8 |
|
Dec 2002
2·3 Posts |
If i had a spare athlon laying around, Id do it :) But, I dont and my non-enthusiast roommate wouldn't be happy with me ripping out his CPU. He wouldn't understand if I tried to explain. So, no DC for me i guess.
Thanks for all your help. |
|
|
|
|
|
#9 |
|
Aug 2002
11001012 Posts |
I actually bought another one before RMA. The new xp1700+ runs way better than the old core. Cooler and more overclockable and factory unlocked. Although I ended up running it at stack spec.
It came bundled with ECS board for $79 retail, cpu worked out to be $45, from Fry's. Now I cant let that motherboard and cpu sit idle ... I can understand your roommate. :D |
|
|
|
|
|
#10 |
|
Aug 2002
2×32×13×37 Posts |
79 bux? Wow that is a good deal... Was it a retail CPU?
|
|
|
|
|
|
#11 |
|
Oct 2002
Lost in the hills of Iowa
7008 Posts |
Sounds like the Thuroughbred core might be finally bubbling down to the slower grade XPs - that 1700+ might be a TBred-A, if it runs THAT much cooler and lower-power than the original Palomino XP1700+
|
|
|
|
![]() |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Prime95 fatal error | maris | Software | 2 | 2014-07-09 18:52 |
| Fatal Error | Cameron2384 | Information & Answers | 3 | 2009-04-26 13:37 |
| Fatal Error | zPrime | Information & Answers | 21 | 2008-10-11 20:33 |
| Prime95 v.25.5 FATAL ERROR | Keepcase | PrimeNet | 1 | 2008-03-07 11:05 |
| Fatal error on new PC. What to do? | Unregistered | Hardware | 2 | 2005-09-05 21:14 |