[SOLVED] Upgrading stock cooler for FX-8350 ?

[patelrom1992]

I am tired of the loud noise of my stock cooler on my FX 8350.
While gaming the temperature goes to 75° C and the fan is at 6000RPM.
I am using Open Hardware Monitor to monitor the temps.
The Motherboard shows 3 temperatures T1-40-45° C ; T2- 40-45° C ; T3-75° C
CPU temp: 45-50° C.
GPU Temp: 55-60° C.
The case fans are at 1400-1600RPM.

I was thinking about upgrading the stock cooler with an aftermarket cooler.

1. Cooler master Hyper 212 RGB Black Edition 120mm.
2. Cooler master Hyper 212 Red LED 120mm.
3. Cooler master Hyper 410R RGB Edition 92mm.

I have 2 doubts:

1. Should I go with an AIO Cooler or anyone from the above will work. ? ( I am not planning to OC )
2. If I go with anyone from the above will it fit in my Motherboard without blocking the RAM Slots.?

My system config:
MB: GA 78LMT-USB3 rev6
Processor: FX 8350
GPU: Nvidia GTX 770
PSU: Cooler master 550W.
RAM: 2x 4GB DDR3. ( only 2 slots working )
Case: Gigabyte C200.
Case fans: 1 exhaust at the rear (came with the case ) / 1 inlet in the front ( Arctic P14 PWM )

 

[Phaaze88]

A)No mention of how long ago that was.

B)That's a higher wattage rating than you initially posted, but still bad - not good to play games on. Better suited for an office PC.

C)Still using apps that report temperatures in real time, which this cpu does not use, so the 55-58C result under a stress test is officially inaccurate...

D)Fortunately, you did not lose dual channel. That would've been either 1+2 or 3+4.

E)Furmark and Msi Kombustor are the same thing, essentially being the Prime95 of gpus, which is used to test how good the cooler/cooling is.
They do not represent actual in game performance; that's why it's possible for users to pass them with no trouble, but still crash in games, which are a better test of power and voltage stability.
Prime95. The settings used matter.
OCCT. The settings used matter.

 

[patelrom1992]

A)No mention of how long ago that was.

B)That's a higher wattage rating than you initially posted, but still bad - not good to play games on. Better suited for an office PC.

C)Still using apps that report temperatures in real time, which this cpu does not use, so the 55-58C result under a stress test is officially inaccurate...

D)Fortunately, you did not lose dual channel. That would've been either 1+2 or 3+4.

E)Furmark and Msi Kombustor are the same thing, essentially being the Prime95 of gpus, which is used to test how good the cooler/cooling is.
They do not represent actual in game performance; that's why it's possible for users to pass them with no trouble, but still crash in games, which are a better test of power and voltage stability.
Prime95. The settings used matter.
OCCT. The settings used matter.

A) After the cleaning and reassembly all this problems started.

B) ok. when i bought this i didn't know about the difference between this one and the gold rated one so i think that was a mistake.
its under use since last 1 year only.

C) Installed amd overdrive and will post the results by end of the day.

https://imgur.com/YdzWQzt

AMD Overdrive


E) I was comparing to the stress test because during the gameplay my max cpu and gpu usage is only 50% so i never have to go to 100% usages as its not a graphic intensive game.AMD Overdrive

 

[patelrom1992]

A)No mention of how long ago that was.

B)That's a higher wattage rating than you initially posted, but still bad - not good to play games on. Better suited for an office PC.

C)Still using apps that report temperatures in real time, which this cpu does not use, so the 55-58C result under a stress test is officially inaccurate...

D)Fortunately, you did not lose dual channel. That would've been either 1+2 or 3+4.

E)Furmark and Msi Kombustor are the same thing, essentially being the Prime95 of gpus, which is used to test how good the cooler/cooling is.
They do not represent actual in game performance; that's why it's possible for users to pass them with no trouble, but still crash in games, which are a better test of power and voltage stability.
Prime95. The settings used matter.
OCCT. The settings used matter.

@Phaaze88 I've uploaded the AMD Overdrive data.
Could you suggest what can I do further ?

 

[Phaaze88]

A)Last week, last month, last year... I'm curious for how long you've been wrestling with this.

C)Heaven Benchmark is a dated gpu focused benchmark. It does not appear to have anything in common with the game(s) you play.
The 50% cpu usage could simply be the game(s) are not optimized to use more than 2 cores, and if those cores are maxed, then the gpu can't stretch its legs either.
Slow storage, or maxed disk usage can also slow things down in the rendering pipeline.

@Phaaze88 I've uploaded the AMD Overdrive data.
Could you suggest what can I do further ?

Monitor temperatures in the actual game, as well as disk usage(Task Manager).

 

[patelrom1992]

A)Last week, last month, last year... I'm curious for how long you've been wrestling with this.

C)Heaven Benchmark is a dated gpu focused benchmark. It does not appear to have anything in common with the game(s) you play.
The 50% cpu usage could simply be the game(s) are not optimized to use more than 2 cores, and if those cores are maxed, then the gpu can't stretch its legs either.
Slow storage, or maxed disk usage can also slow things down in the rendering pipeline.


Monitor temperatures in the actual game, as well as disk usage(Task Manager).

A) I did that around 7 or 8th May. That day after cleaning and reassembly I faced 2 direct restarts while in-game so I decided to re-check everything and also updated the drivers. So I dont know the problem is related to cleaning or just appeared after driver update.

B) For benchmark after the Heaven benchmark I also did the 3dmark Firestrike and faced no issue in that. ( it was considerably short 5-7 mintues ).

One thing that I did yesterday was to change my GTX 770 2GB with an old Radeon HD6670 1GB.
Was able to play without any issues and also the GPU was running at 90-95% usage ( memory usage was 600MB ) while the CPU was at 40-50% only similar to the GTX 770.

Also as you asked to monitor the temperatures while gaming, the GPU doesn't go above 60C and the CPU works around 53-56C.
I'm using MSI Afterburner, Riva Tuner Statistics ( for on screen display ) and compared it to AMD Overdrive also. All the readings are same.

Regarding the disk usage I am able to play with HD6670 with the same disk and CPU but facing issues only with the GTX 770.
Also if you will notice since last 1-2 weeks lots of people are coming up to different forums with this problem after updating drivers.

 

[Phaaze88]

A)So we can't rule out the possibility that something was damaged while cleaning, because everything was fine before that...

B)HD 6670 is a fair bit weaker than GTX 770. It also can't pull as much power(66w Vs 230w).

Gpu isn't hitting throttle limit. Cpu:

I'm using MSI Afterburner, Riva Tuner Statistics ( for on screen display ) and compared it to AMD Overdrive also. All the readings are same.

Surprised that doing so worked without sensor conflicts, thus giving incorrect readings on both...
I also wasn't aware Afterburner and Rivatuner read Thermal Margin either...

Also if you will notice since last 1-2 weeks lots of people are coming up to different forums with this problem after updating drivers.

There's so many different system configs out there... there's bound to be some that don't agree with the latest driver(s).
Roll back to a previous driver until the next one comes out... but Kepler driver support ends come the 470 drivers, so there will be no need to keep updating when those come.



Ok, so rereading back through the entire thread, the highlights appear to be:
1)Noisy.
This one's simple. A new cooler, but better chassis airflow would help whole system temperatures too.
The problem here is the dated AM3+ platform; that might not be worth trying to fix.

2)Possible damage incurred during cleaning.
We don't know exactly what may have been damaged. Gpu and motherboard are the prime suspects, I think - maybe the psu, though it might be a coincidence...
Or maybe you didn't plug one of the connectors all the way...

I was careless during cleaning once - knocked off some part of the motherboard's onboard audio, causing sound to only display through the right speaker. That was easily remedied by using the gpu's onboard audio instead.

 

[patelrom1992]

A)So we can't rule out the possibility that something was damaged while cleaning, because everything was fine before that...

B)HD 6670 is a fair bit weaker than GTX 770. It also can't pull as much power(66w Vs 230w).

Gpu isn't hitting throttle limit. Cpu:

Surprised that doing so worked without sensor conflicts, thus giving incorrect readings on both...
I also wasn't aware Afterburner and Rivatuner read Thermal Margin either...


There's so many different system configs out there... there's bound to be some that don't agree with the latest driver(s).
Roll back to a previous driver until the next one comes out... but Kepler driver support ends come the 470 drivers, so there will be no need to keep updating when those come.



Ok, so rereading back through the entire thread, the highlights appear to be:
1)Noisy.
This one's simple. A new cooler, but better chassis airflow would help whole system temperatures too.
The problem here is the dated AM3+ platform; that might not be worth trying to fix.

2)Possible damage incurred during cleaning.
We don't know exactly what may have been damaged. Gpu and motherboard are the prime suspects, I think - maybe the psu, though it might be a coincidence...
Or maybe you didn't plug one of the connectors all the way...

I was careless during cleaning once - knocked off some part of the motherboard's onboard audio, causing sound to only display through the right speaker. That was easily remedied by using the gpu's onboard audio instead.

I had posted this issue on a Nvidia Geforce Forum too and someone had asked me to upload the voltage data and he did notice that at idle condition the 12V is coming at 11.736V but as soon as the GPU is put under load the voltage drops to 11.664 and even 11.592 and then it will keep changing between 11.592 - 11.664 - 11.736 every few seconds.

Does it point out that the PSU is causing problems ?
https://docs.google.com/spreadsheets/d/1JwlkpnIH7NuGuFWB9upjHNC27rQJyFUSwqBF7fH6fk4/edit?usp=sharing

Here is the link to the file if you would like to check it out.
Going to try out a spare PSU to see if that gives proper readings.

I am a bit unsure regarding the cleaning because if I would have damaged anything related to the hardware then I would not be able to play any games with any GPU, and if by chance I have damaged the GPU then I am already planning to give it to a service center for testing.

 

[Phaaze88]

The max acceptable voltage tolerance for the ATX 12v spec appears to be 5%: 11.4v - 12.6v.
The problem here is whether the software can be trusted to be reporting the voltage correctly... I wish I could tell you more about that one besides the following...
-Try a different power supply
-Try the GTX 770 in a friend's PC
-Try the GTX 770 with lower power limits in Msi Afterburner

I am a bit unsure regarding the cleaning because if I would have damaged anything related to the hardware then I would not be able to play any games with any GPU, and if by chance I have damaged the GPU then I am already planning to give it to a service center for testing.

That's why I said it could've been a coincidence that it started after the cleaning.

 

[Karadjgne]

Open Hardware Monitor or ANY other software used for cpu temp monitoring on an FX cpu is a fail. Can't be done reliably in any way.

The FX do not have thermal strips in the cores. There's physically no way to read core temps the same as Intel are read.

There's only 2 programs that can 'judge' temps and that's amd Overdrive and HWInfo and those Must be used in Thermal Margin mode.

Thermal Margins are Not a temperature reading, they are a load monitor. They use a complex algorithm comprised of cpu voltages, load %, core use, package temps and other things to come up with a range of heat output. The closer that number is to 0, the warmer the cpu is. So idles in the 40's are good, use in the 20's is good, but hitting the single digits means you don't have much room left thermally at all.

The number itself is not important, what is important is what that number represents. There's not really any difference between 20 and 29, you still have plenty of room before 0.

As to cooling that motherboard, I'd suggest a downdraft cooler design such as Noctua NH-C14S, BeQuiet DarkRock TF etc, which can use a secondary fan underneath. Not only will that allow greater heat tolerances, but has a direct airflow onto the board, unlike tower coolers that have indirect airflow.