AMD FX-6100 with Gigabyte 990FXA-ud3

[g2nightmare]

Hey guys,

I'm looking to overclock this processor. I found a nice guide (http://www.overclockers.com/forums/showthread.php?t=688663) but I appeared to have screwed up somewhere. Performance seemed to have gotten worse.

I figured I would solve my problems easier if I asked this:
Has anybody overclocked the AMD FX-6100 cpu with the Gigabyte 990FXA-ud3 (or similar) motherboard?

If so, can you type me your bios settings (or take a picture) so I can model them? I also want to check up on my settings.

Much appreciated!!

 

[RGone]

Hello and ...

and welcome to OCF. Got some stuff that will get you started in helping us help you. Good luck with your journey.

My how to put up Sig at OCF.


New Shortcut method for putting a Signature with your system information following your every post so people can know what is in the case that they are trying to assist with. You can use something like what is shown in my signature as a good template of needed information Thank you.

 

[gnusounduave]

If you have good air cooling you're looking at around 4.5 - 4.7 ghz max; I haven't seen much on water cooled 6100's. You've got a decent board that can handle some good top end OC's.

Now put your config in your sig and let's see what you're workin with.

 

[g2nightmare]

Ok. I updated my signature.

Really, I'm looking for someone who has overclocked the same processor on a 990fxa motherboard. I would prefer if somebody could share their settings so I have something to base off of.

 

[Psykoikonov]

I have the same motherboard as you and processor is of the same family. I have cool&quiet and APM enabled (save some power) however C1E and C6 are disabled.

It's been shown that running the HT Link at the same speed as CPU Northbridge (2200Mhz or 11) has benefits. I would set CPU clock ratio to 20, Core performance boost disabled and the remainder on auto.

Test with something like Prime95 (3 hours or more of blend), Intelburntest or LinX (5+ passes on "high"), all free programs. Try to keep temps below ~60C.

If the system passes the test and temps are good, then go up one multiplier and test again. If it fails add voltage (CPU voltage control), also CPU NB VID control sometimes helps stabilize things.

When the temps are too high, I usually drop back a full multiplier or a least a half and it should be solid for day to day use.

Processors are not created equal, that's part of the interest in overclocking.

 

[g2nightmare]

I have the same motherboard as you and processor is of the same family. I have cool&quiet and APM enabled (save some power) however C1E and C6 are disabled.

It's been shown that running the HT Link at the same speed as CPU Northbridge (2200Mhz or 11) has benefits. I would set CPU clock ratio to 20, Core performance boost disabled and the remainder on auto.

Test with something like Prime95 (3 hours or more of blend), Intelburntest or LinX (5+ passes on "high"), all free programs. Try to keep temps below ~60C.

If the system passes the test and temps are good, then go up one multiplier and test again. If it fails add voltage (CPU voltage control), also CPU NB VID control sometimes helps stabilize things.

When the temps are too high, I usually drop back a full multiplier or a least a half and it should be solid for day to day use.

Processors are not created equal, that's part of the interest in overclocking.

Thanks for your quick reply!

Could you take a picture (with a cell phone or something) of your bios screen? I want to make sure I am changing the correct things. Also, how do I determine test results with Prime95? Do I just run it and only monitor temperature?

I'll be able to try this tomorrow after work.

 

[Psykoikonov]

Starting on page 33 of your mobo manual are descriptions and pictures of the BIOS layout.

When Prime95 is run it creates a "worker" for each core (in your case 6). The "workers" run indefinitely or until an error is detected, any errors detected is bad news. When a worker fails you have 3 basic options 1. Increase the voltage provided the last load temp was below ~60C, 2. Lower the temperature (better cooling) or 3. Decrease the speed. With how fast your processor will pass tests I would say somewhere between 3-6 hours of no errors, before increasing the multiplier, rinse and repeat.

 

[trents]

Failing Prime95 can take several forms: Blue Screen, spontaneous restart, program aborts/locks up or as Psykoikonov mentioned, one or more if the core workers can drop out. If the latter happens, that usually is a sign that you are very close to being stable.

 

[BeepBeep2]

With FX it is almost always a core dropping out, even 300-400 MHz from stability.

The UD3 isn't a very good board for overclocking AMD FX, the boards have a horrid vdroop AFAIK under load.

 

[g2nightmare]

Starting on page 33 of your mobo manual are descriptions and pictures of the BIOS layout.

When Prime95 is run it creates a "worker" for each core (in your case 6). The "workers" run indefinitely or until an error is detected, any errors detected is bad news. When a worker fails you have 3 basic options 1. Increase the voltage provided the last load temp was below ~60C, 2. Lower the temperature (better cooling) or 3. Decrease the speed. With how fast your processor will pass tests I would say somewhere between 3-6 hours of no errors, before increasing the multiplier, rinse and repeat.

Thanks! Great explanation.. I'll be trying this shortly. Hopefully nothing breaks.

 

[Psykoikonov]

The UD3 isn't a very good board for overclocking AMD FX, the boards have a horrid vdroop AFAIK under load.

I've read that in a number of posts too. Mine doesn't though all the way up to 1.475V (highest I've taken it), Vcore goes up slightly under load. SSs below is load and no load, 1.30V set and "regular" LLC. My mobo is rev 1.1, perhaps that is the difference.

 

[BeepBeep2]

Yeah only Rev 1.1 and up have LLC, also it isn't completely effective on Gigabyte boards for some reason, if you try LinX you will still see vdroop.

EDIT:
Whoops

 

[RGone]

@ "Psykoikonov", glad your board clocks good with FX cpu. You got any idea what the Revision of that board is? The Rev is usually in the lower left corner as it is mounted upright in the case. RGone...

 

[g2nightmare]

@ "Psykoikonov", glad your board clocks good with FX cpu. You got any idea what the Revision of that board is? The Rev is usually in the lower left corner as it is mounted upright in the case. RGone...

I have REV 1.1 .. what are you guys even talking about?

BTW, I set north bridge and ht link to 2200 mhz and cpu to 4 ghz. I'm currently running 6 blend tests.. how long does this go?

Also is it normal for Windows Aero to turn off?

HWMonitor reports my cores to be 50 degrees celsius. I set my cpu voltage and north bridge voltage to 1.3V is that too much?

 

[Psykoikonov]

@ "Psykoikonov", glad your board clocks good with FX cpu. You got any idea what the Revision of that board is? The Rev is usually in the lower left corner as it is mounted upright in the case. RGone...

My mobo is rev 1.1.

I have REV 1.1 .. what are you guys even talking about?

BTW, I set north bridge and ht link to 2200 mhz and cpu to 4 ghz. I'm currently running 6 blend tests.. how long does this go?

Also is it normal for Windows Aero to turn off?

HWMonitor reports my cores to be 50 degrees celsius. I set my cpu voltage and north bridge voltage to 1.3V is that too much?

Vdroop refers to a drop in voltage from the set amount, for instance when stress testing the voltage will drop slightly, some motherboards do more than others.

 

[g2nightmare]

I just ran Prime95 for 2 hours.. 0 errors at 4 ghz and 2.2 ghz cpu north bridge and ht link.

Where do I go from here? Should I step it up by .5x or higher?

edit: temperatures were at 52 max.

 

[RGone]

I just ran Prime95 for 2 hours.. 0 errors at 4 ghz and 2.2 ghz cpu north bridge and ht link.

Where do I go from here? Should I step it up by .5x or higher?

edit: temperatures were at 52 max.

1. When you say temps, we normally 'see' a capture of HWMonitor with CPU CORE TEMP visible and HWMonitor was open on the desktop logging min/max temps during the P95 blend testing. That is what we like to 'see' when talking about temps to decide on going to more Mhz.

2. Going up by 0.5 multiplier is okay, but 'how far' do you really think you want to go?

3. If you plan on going to 4.2/4.4Ghz, then I would jump multipliers by 1.0 and test.

4. Currently even if you do not see it since you are not going to run an Intel rig right beside your AMD rig, the AMD is giving away power to do work to Intel, so most want to clock the AMD to gain back what they feel they have lost. The other side of that coin is that with my knowledge and yet I have an AMD system, then I must not care. So I will push upward to the point I don't feel like temps are just too much and cannot throttle my fans down and call the job done before I run some high MHz number that I will never know is better than somethinig 300Mhz lower but cooler and quieter. You got to make your own up on that sort of thing.

 

[g2nightmare]

1. When you say temps, we normally 'see' a capture of HWMonitor with CPU CORE TEMP visible and HWMonitor was open on the desktop logging min/max temps during the P95 blend testing. That is what we like to 'see' when talking about temps to decide on going to more Mhz.

2. Going up by 0.5 multiplier is okay, but 'how far' do you really think you want to go?

3. If you plan on going to 4.2/4.4Ghz, then I would jump multipliers by 1.0 and test.

4. Currently even if you do not see it since you are not going to run an Intel rig right beside your AMD rig, the AMD is giving away power to do work to Intel, so most want to clock the AMD to gain back what they feel they have lost. The other side of that coin is that with my knowledge and yet I have an AMD system, then I must not care. So I will push upward to the point I don't feel like temps are just too much and cannot throttle my fans down and call the job done before I run some high MHz number that I will never know is better than somethinig 300Mhz lower but cooler and quieter. You got to make your own up on that sort of thing.

Well right now I'm at 4.2 ghz and everything appears to be stable.. Is there any additional fine tuning I should do at this point or should I keep going?

I'm trying to understand what you meant by temperatures -- I really was sitting in my computer chair watching TV and glancing over every couple of minutes, and temperatures were stable around 49-52 degrees on each core.. as long as I don't go near 60 right?

 

[Psykoikonov]

1. When you say temps, we normally 'see' a capture of HWMonitor with CPU CORE TEMP visible and HWMonitor was open on the desktop logging min/max temps during the P95 blend testing. That is what we like to 'see' when talking about temps to decide on going to more Mhz.

2. Going up by 0.5 multiplier is okay, but 'how far' do you really think you want to go?

3. If you plan on going to 4.2/4.4Ghz, then I would jump multipliers by 1.0 and test.

4. Currently even if you do not see it since you are not going to run an Intel rig right beside your AMD rig, the AMD is giving away power to do work to Intel, so most want to clock the AMD to gain back what they feel they have lost. The other side of that coin is that with my knowledge and yet I have an AMD system, then I must not care. So I will push upward to the point I don't feel like temps are just too much and cannot throttle my fans down and call the job done before I run some high MHz number that I will never know is better than somethinig 300Mhz lower but cooler and quieter. You got to make your own up on that sort of thing.

+1, on top of that the FX chips reach a point where they need a hefty increase in voltage = a hefty decrease in efficiency. Mine will do 4.1Ghz on stock voltage, and +0.05V for 4.2Ghz beyond that it really wants Vcore, again every processor is different. Yours may do 4.2Ghz on stock voltage. The only way to tell for sure is to try. Aiming for 5.0Ghz is more a "it can be done" than anything that will benefit from day to day, if that is the goal all the power to you and I hope luck is on your side.

 

[BeepBeep2]

By the way, if Aero is crashing then your system is unstable.
For some reason with FX, Aero is always the first thing to crash...especially when the CPU is under a large load.

Some FX CPUs crash Aero when at stock settings (ie. Load Line Calibration/Control off) and must be RMA'ed.

I had an FX-8150 that did 4.6 GHz on 1.42v on water cooling but I had Aero crashes at stock speeds, stock voltage, and no LLC. That one went off to RMA.