Sign in with
Sign up | Sign in
Your question
Solved

Please help! FX8350 Temp / voltage Issue!

Tags:
  • AMD
  • CPUs
  • FX8350
  • Temperature
Last response: in CPUs
Share
April 25, 2014 12:55:15 AM

Hi guys! Great forum, it has helped me a lot in the past, I hope you can help me now; I am pretty sure you will.

I built a Render node like this:

AMDFX8350, stock, CM Hyper TX3 EVO
ASUS M5A78L-M/USB
2x8GB DDR3 1866
SSD Sandisk 64GB

I want to use it just for render as a slave on a small renderfarm. After some test, i realized it was going slower than it was suposed to be, so had a look at temperatures, voltages, etc (I don't know much about this so please help me)

I noticed the core multiplier dropped from x20 to x7 on full load when CPU temp (TMPIN0 in CPUID HWmonitor) hit about 60ºC, and it has crashed sometimes at this temperature.

I took a look at voltages as I have told you and i dont like +12V to be at 8.2V, or -5V to be at -9.216V, but I really dont know what these means.

I have read some posts about this issue but can´t solve the problem yet, so I ask for your help. I will really apreciate it, Any idea? Thank you all!

I upload idle and full load HW reads.

Idle:
Full load:

More about : fx8350 temp voltage issue

a c 117 À AMD
a c 473 à CPUs
April 25, 2014 1:26:18 AM

What's the model of your PSU? There's definitely something wrong with it. The +3.3, +5 and +12V (and probably corresponding '-' values) should not be far from those values I stated.
m
0
l
April 25, 2014 1:47:31 AM

Hi Jook-D and thanks. PSU is a 500W bundled included in a B-Move Matx tower (B-Move Chill MicroATX 500W) Maybe I went cheaper than the setup needs to work properly...
m
0
l
Related resources

Best solution

April 25, 2014 1:56:00 AM

8.2v on the 12v rail is baddd, I don't like that at all, but HW is known to give some bogus ass reading too though....get a multimeter and probe the rail and see what you get
Share
April 25, 2014 2:19:33 AM

Ok guys, this makes no sense. I have new screenshots where voltages and temps dont even change, and the aren`t the same in HW monitor and Everest... I am pretty lost.

Idle

sube

Load

subir foto
m
0
l
April 25, 2014 2:31:56 AM

HW right now says i have 9.5v on +12v rail.....BIOS says I have 11.96-12.1v.....I probed every +12 rail with my meter, what did I get?- 11.96v xD
m
0
l
April 25, 2014 5:24:11 AM

Thanks JR1988, but I don´t have such a meter and no idea where to measure xD! I also tried speedFan and values were much closer to they should be... So maybe voltages are ok but core multiplier drops from time to time when full load...

Shall I have a better MB? No idea, I´m lost...
m
0
l
April 25, 2014 6:33:35 AM

your Molex 4/pin yellow wire is hot +12v rail red probe on meter goes there, black wire is ground/black probe etc, 6 pin pcie top pins are ground bottom is hot etc.
m
0
l
April 25, 2014 6:37:09 AM

Are your Fans working ok?! you got some ridiculous high temps there...
m
0
l
April 25, 2014 6:39:15 AM

nm thats HW being stupid on TMPIN2 lol
m
0
l
a c 117 À AMD
a c 473 à CPUs
April 25, 2014 6:40:13 AM

JR1988 said:
Are your Fans working ok?! you got some ridiculous high temps there...


The 128C reading is an unused or broken sensor, nothing to worry about. The core and socket temps are fine.

EDIT: saw your update. :p 
m
0
l
April 25, 2014 6:43:47 AM

if it was really at 128c think you would see some smoke and very bad smell lol
m
0
l
April 25, 2014 6:46:40 AM

I have read that when a temp is so stupidly high is due to my mb doesnt have such sensor and the software make it up. I thnik fans work propely, i have the case opened and everything is ok, or is what it looks... Cpu heatsink is not as warm as HW is telling...

I dont have multimeter but will get one... Meanwhile, any more ideas? Thanks!
m
0
l
!