• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Odd Phenom 9850 behaviour

NightmareXX

NightmareXX

NightmareXX

Associate
Joined
12 Aug 2005
Posts
1,025
Location
Team 10
Well I was watching the core temps of my 9850 while running R@H to make sure that I had the cooling right and noticed that it occasionally drops the multiplier down to 6.25 and the volts to 1.04 resulting in the core speed drop from 2.8GHz (14x mult) to 1.25GHz. As far as I'm aware, it only drops it under full load but it's not a frequent occurance. I can't work out what's causing it do to this, any suggestions?

9850BE @ 14x multi 2.8GHz 1.264V
Asus M2N68-AM Motherboard
Corsair 400W PSU (More than enough power)
2x1GB DDR2 PC6400 OCZ RAM

I have CnQ disabled in the BIOS

EDIT: Running prime 95 on it and the decrease in CPU speed is becoming much more obvious as it's downclocking itself every 10-15 seconds. About 48-50C on full load, chipset is about 52-54C.

Just downloaded core temp and it's giving me a different reading to CPU-Z. CPUZ reports a core speed of 2800MHz where as coretemp says 178 x 14 which is 2500MHz. Which do I believe? Could CPU-Z be wrong and the reason it's throttling itself?
 
Last edited:
That chipset sounds a bit hot to me, but then again I don't know if that's normal for AMD. Have you checked Throttling/speedstep (or AMD equivalent) in the BIOS options? That sounds like the greatest cause. Also what program are you using for temperatures?
 
As I've mentioned, Cool and Quiet is disabled and I suppose it could be limited by the chipset but I've never noticed this before. The chipset is running cooler then it was (I've moved the case and added another fan). I'll set it back to stock speeds and see if that sorts it.
 
sounds like cool n quiet is enabled some how.

is the bios up to date?

it could just the program giving wrong readings.
 
Last edited:
Removed the light overclock and still getting the same problem. According to CPU-Z and to my knowledge, the BIOS is the 0317 dated 22/10/2008. Might try updating it and see if that works :)

Well the BIOS update hasn't stopped it :confused:
 
Last edited:
Reports the same as speedfan and coretemp :(

Gone in to the BIOS and disabled both CnQ and C1E. Running OCCT Linpack so I can post some pictures to illustrate the problem.

EDIT: Odd, it failed the Linpack test. What would cause that?
 
Last edited:
Yes, it's at stock settings and as far as I'm aware, everything is normal. I failed 3 of the 5 Linpack tests and that's not something I'm happy about as there's nothing wrong.

I'm pretty sure I've covered all the BIOS settings related to throttling.
 
The plot thickens.

I'd moved the case into a cupboard next to another PC and only since then I've been getting errors. I've now moved it back to where it was originally and it's working perfectly fine. Reckon it could be EMF radiation affecting it? That's the only cause I can come up with.

Just ran Linpack 20 times and it failed 8 of the 20 tests :(
 
Last edited:
Well I just ran Prime95 in RAM testing and the moment I do, I get a screen wide graphical corruption. The screen goes white and has green lines across it. What's that mean? RAM problem or broken NB/SB? It is an integrated chip so could I have burnt it out?
 
Running memtest now and although it hasn't been going for long, I'd have expected it to have failed by now. I did notice something odd though.

With the system in a case, the screen went all buggy when I ran the Prime95 RAM test. I've now moved the components on to my desk so they're open to the air. Using exactly the same configuration, I could start the Prime95 RAM test and it wouldn't fail. I didn't leave it running for long, just enough to confirm that it hadn't crashed instantly.

With the system now on my desk, I can touch the heatsink on what I presume to be the southbridge. It's roasting! Not hot enough to burn but enough to be uncomfortable to touch. If it's overheating, would that cause problems? How hot are they allowed to get?
 
Well I've done the following tests:

Swapped the memory sticks around
Swapped the memory with a completely different set
Swapped the motherboard for the exact same model, just a different one

All 3 resulted in the same behaviour. I've also noticed that if I run a Prime95 Blend test, I get a MASSIVE screen corruption on bother motherboards with different sets of RAM.

Prime95 blend fails
Linpack on 20 loops doesn't crash it but it fails 80+% of the tests

Guess this means it's the CPU?
 
It's an integrated GPU hence the change of motherboard mooting that point.

I'd be awfully surprised if it was the PSU as it's a corsair 400W model. I've tested it at the plug and it never draws more than 150W on full load. I'll see if I've got another one I can test.

EDIT: I did try the same PSU with a dual core in the same motherboard, RAM, HDD and it ran the blend test without corrupting.
 
You must log in or register to reply here.
Back
Top Bottom