Phenom II X4 965 BE unstable at 3.8ghz

CptMirth

Commendable
Aug 3, 2016
5
0
1,510
I cannot get my Phenom II X4 965 to run stable at 3.8ghz. I've tried voltages at 1.425V, 1.45V and 1.475V. 1.475V seems to be the most stable, but I still get occasional BSODs. The CPU temps never run above 60C. Basically, I'm not sure what to do now. Should I try 1.5V, or is that too risky?
I've done lots of research and there are few things that I am unclear on, due to conflicting information. Do I also need to OC the Northbridge, and increase its voltage?
Also, should I underclock my RAM? I've heard that if I'm just changing the multiplier and not the FSB (which is what I'm doing) that it doesn't need to be underclocked.
I would be very grateful if anyone could help shed some light on the matter.
 
Solution


A BSOD can also mean a corrupt Windows file or bad driver.It should provide you with a dump file that can be analysed.
You can check your Windows files. Run "SFC /scannow" without the quotation marks in an elevated command prompt.
This will check your system files for corruption and try to fix them.
Check Event Viewer for clues or check in Device...
Hi CptMirth :)

As you have decent temperatures at 3.8GHz, Try 1.5V which will be OK. You can even go slightly higher but not recommended above 1.550V.
Keep an eye on your temperatures as each small step in Core Voltage has an exponential increase in heat. So it will come down to your cooling capability.

Bios settings for your DIMMs should suit the SPD or XMP profile if available. List your DIMMs p/n if unsure.

Best to OC with the Multiplier and make FSB changes if needed for stability. (best left on Auto imo.)

Also stress testing is a good way to determine stability for your desired OC. Should you wish to do this and are unsure how, Advice will be given.
 

CptMirth

Commendable
Aug 3, 2016
5
0
1,510
Thank you so much for your advice MeanMachine41. I'll try 1.5V later today and let you know how it goes.
I'm not sure what SPD or XMP profiles are. The Module Part Number: BLE4G3D1608DE1TX0 (I have 2 of these). At the moment I have the memory clock set to x8.00 for 1600Mhz. Is this ok, or should it be lowered/change something else?
As for stress testing, I use Prime95.
Thanks again, I really appreciate the help :)

Edit: I probably should have mentioned this in my first post: Motherboard Model: GIGABYTE GA-MA770T-UD3
 
OK CptMirth, I checked up on your DIMMs, they are Crucial Ballistix Elite 4Gb and you have 2 modules.

Unfortunately you have two individual DIMMs of the same spec however, it is not a good idea to mix DIMMs, even of the same spec.
They may or may not work together as during Manufacture there can be variations and Crucial would match them as a working kit.
In other words, you should have bought a single 8GB kit. Anyhow testing will show and if you can't get them to work together then I suggest you return them and buy a single kit.

You can obtain the SPD info for your DIMMs Primary Timings in CPUz under the Memory and SPD tabs and for voltage settings. These will be entered in your Bios. Write them down.

If not sure how to go about this I will explain in more detail.

I don't like P95 as it is a harsh tester and will not test your sub systems. It is not a realworld tester and is too stressfull for your CPU. I only use it on small FFT for no more than 10mins initially, if at all.

Download AIDA64 trial version and HWMonitor and put them on your desktop side by side.
First run the Test for CPU and FPU in AIDA64 for 10mins and shut the test down if temps get above 80C.
In HWMonitor I want you to open up the tab to show all rail voltages, 3v,5v,12v. This will tell me the condition of your PSU under load.
You can also run the other tests separately if you wish.

At the 10min mark of AIDA test, take a screen shot and upload to IMIGUR file host. Obtain the BB url and link it here for analysis.

Please report back.
 

CptMirth

Commendable
Aug 3, 2016
5
0
1,510
I did buy the memory as an 8GB kit :)

I'm not sure what you meant by the BB url, but here is the results from running the CPU at 3.8ghz @ 1.5V

daPko1G.png

 
All rail voltages are good acceptable voltages although slightly high.
Temperatures under stress are good at max 57C. You can try for 4GH OC with same Vcore at 1.504V MAX.

Primary Timings for your modules are listed as 8-8-8-24 and your Bios settings indicate 11-11-11-29 so you can tighten them up and try a command rate of 2T.
Strange, The P/N of your DIMMs BLE4G3D1608DE1TX0 indicate a single 4GB module so don't know whats going on there. Normally it would show as BLE8GB~~

Could you now try the Memory test after changing timings in Bios.

To the left of your picture is CPUz, open the Memory and SPD tabs to verify Timings and possible XMP profile.

Extend the tests for 30mins with new settings and report back.

If you still get BSOD then we will have to analyse the Dump file for the Culprit.

The BB url is the Bulletin Board url link, However you made it work.
 

CptMirth

Commendable
Aug 3, 2016
5
0
1,510
I am tempted to try 4ghz at some point, although my BIOS only allows voltage increments of 0.025V.

After changing the memory timings to 8-8-8-24, my computer would not POST. Would just get one repeating long beep. I also tried leaving the tRC and changing according to the XMP profile (36), but both have the same result. Any idea what as to what could cause this or how to go about fixing it?

 

CptMirth

Commendable
Aug 3, 2016
5
0
1,510
I've just noticed something else that's strange. I decided to try running my CPU at stock settings (3.4Ghz @ 1.4V) to see how stable and it is, and while it is stable, I occasionally encounter BSODs during Flash video playback. That is typically when it would BSOD when at 3.8ghz @ 1.5V. I have no idea what this could possibly mean, so I would appreciate any insights into the matter :)
 


A BSOD can also mean a corrupt Windows file or bad driver.It should provide you with a dump file that can be analysed.
You can check your Windows files. Run "SFC /scannow" without the quotation marks in an elevated command prompt.
This will check your system files for corruption and try to fix them.
Check Event Viewer for clues or check in Device Manager for any yellow triangles indicating a bad or incorrect driver.
You may just need to update your ADOBE Flash player plugin.

If it continues then you may have to conduct stress testing to find what's causing it.
 
Solution