• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Questions concerning OC'ing FX-4100

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Svnryn

Registered
Joined
Jun 27, 2012
Hi all,

Looking to OC my 4100 to get some more performance out of a few of the games I play (primarily Planetside 2). I'd like to hit somewhere around 4.4 and from what I've been reading, that should be possible given the cooling I have.

I've been reading this guide here but find myself a little overwhelmed by it. Should I simply change the multiplier and see how far I can go on stock voltages followed by tuning the voltage and THEN the NB/HT stuff?

Also, will my memory at 1066Mhz be a problem? Should I overclock it or look into getting faster memory?

Any and all help is much appreciated. Thanks in advance.
 
Last edited:
No sense in RE-inventing the wheel everytime somebody wants to overclock a cpu. You can see what voltage he used and the speed of his overclock in about the first or second post.

OC'ing FX4100
 
OK so I went ahead and pushed the multi to x20 and turned off everything I need to (Cool n Quiet, C6, C1e, APM, Core Booster). Ran two hours of Prime 95 blend on stock voltages with no errors and will now move to x21.

But before I do that a few questions:

When Prime 95 isn't running, the voltage is at 1.392V. However, as you can see in the image it drops to 1.376. Is there a setting I forgot to turn off in BIOS?

Also, my 1066 memory timings are 7-7-7-18 at 1.5 volts. Everything is set like that automatically in BIOS expect for the last timing is set to 20 instead of 18. Should I change this or leave it?

Finally, in BIOS at the bottom where it says CPU Voltage, I left it at the Normal preset and below it, it says VCore is 1.4125. Is that what the vcore is set at currently even though HWmonitor is saying its 1.392?

Oh and when you say give the voltage a "bump" is that in .01 increments? Gonna try out x21 here shortly and will report back what I find.
 

Attachments

  • x20FX4100.jpg
    x20FX4100.jpg
    337.7 KB · Views: 89
Crap, just noticed this is in the Motherboards section. Must've been more exhausted than I thought last night and didn't pay attention what AMD section I was in.

Could a Mod please move this to the CPU section? Much appreciated.
 
a bump is one increment of what ever your bios allows, loke one tap of the "+" key would be a "bump".
 
and you need to expand the hardware monitor window down to where we can see the package temp.
 
The vcore fluctuation you mentioned:

This is Normal (vdroop) and can usually be combatted by adjusting the level of LLC (load line calibration)


Leave you're memory as it is for now as adjusting the last timing is not overly important right now but you can adjust it when you get everything sorted out.


As for the vcore discrepancy, I have had boards that have said a lot worse than that, but usually that doesn't happen until nearing the end of the vcore limit for the board, plus sometimes CPUz doesn't work very well on some boards/bios


Sorry for the bad layout, but I can't multi quote as I'm using my mobile :(
 
Caddi and Keny - thanks for the replies. Will be sure to expand HWMonitor on the next set of SS's.

Put it on x21 and got 2 hours of no errors on stock voltage with pretty much same temps.

Thanks for clearing the vcore flux up Keny. I figured it had to do with Load Line Calibration since that is the lone setting I didn't mess with. Which setting should I use in LLC? I believe there's only two aside from auto (Extreme and I forgot the other).

About to try x22 for four hours later today and will report back my findings. Thanks for the help so far guys - much appreciated.
 
Last edited:
Ok, so bumped it up to x22 and did a P95 blend test for four and a half hours and no errors. (pic) I've also included an expanded view of my HWmonitor window. Do the temps look good?

Also, I'm still at (what I assume) to be stock or close-to-stock voltages. When I set the multi to x22 earlier, vcore was at 1.375 in bios. So my question is, where do I go from here?
 

Attachments

  • x22fx4100.jpg
    x22fx4100.jpg
    362 KB · Views: 74
  • x22fx41004hr.jpg
    x22fx41004hr.jpg
    365.3 KB · Views: 104
That all looks good..... Where to go from here is, keep bumping up the multiplier until it fails p95 and when it fails, that is when you need to start adding vcore to stabilise the overclock, but once you start adding vcore you're core temp will start to rise so make sure you monitor it and try to keep it under ~55c under load
 
OK. Which part does TMPIN0, 1 and 2 designate? Is TMPIN2 the CPU or is the reading under FX-4100 package my CPU temp?

Just want to be sure what belongs to what before I go any higher, as to better follow your recommendation, Keny.
 
It's the package temp which you want to focus on, and I think tmpin2 is you're socket temp (usually around 10-12c higher than the core temp on aftermarket air coolers),as long as the other temps don't go stupidly high you have no need to worry about them
 
On a Gigabyte board the TMPIN2 is the CPU (socket) temp and with the FX CPUs what we used to see as "core" temp shows up as Package. With the FX CPU both are important to keep track of because of the complex nature of the FX's power management scheme.
 
So tonight I tried to bump things up to x23 and left voltage on normal in BIOS (1.375 so it claims). I made it into windows and opened P95 and it failed instantly on me, locking up the system in the process. I pressed the reset button and went back into the BIOS.

I gave the voltage a bump (+.025) to 1.400, according to the BIOS. Upon getting back into Windows, CPUID and HWMonitor are telling me that my vcore was set to 1.360 and would occasionally jump to 1.370ish.

Any idea what this is about? I've included pictures of this and of the BIOS. Currently back at x22 (4.4) at what I presume is stock voltage, but CPUID indicates its 1.392 (BIOS says 1.375 with "Normal" set in the CPU Voltage Control option at the bottom of the second BIOS image).
 

Attachments

  • x23troubles.jpg
    x23troubles.jpg
    291.4 KB · Views: 66
  • BIOS1.JPG
    BIOS1.JPG
    1 MB · Views: 70
  • BIOS2.JPG
    BIOS2.JPG
    1 MB · Views: 97
  • BIOSOptions.JPG
    BIOSOptions.JPG
    1,000.9 KB · Views: 67
On the last picture at the top of the menu (load line calibration) if you adjust the setting in there to a different setting like high it should stabilise the voltage reading as LLC increases/reduces the vcore under load and it takes a bit of messing around and adjusting before you can find the setting that suits you the most, also see if a different version of CPUz gives the correct max vcore reading as it should be reading ~1.4v
 
LLC only gives 3 options which are regular, extreme or auto. Depending on which you choose, Vcore either droops like a mofo or goes well beyond where you want it. That is pretty well known about that board across the net and why I don't like them. Other than that they sell quite a few since they are cheap to buy.
 
Well I currently have it on Auto and it's done quite well to reach my initial goal of 4.4. However, I feel like I can squeeze a little more out of it.

Since it droops as is on Auto and Extreme will likely yield beyond what I need, I'll give regular a try with a voltage bump to 1.4 and see if that yields the correct reading. I'm betting it will, but won't know until I try.

But I paid $50 in combo with the processor back in July at Microcenter and all things considered along with its features, it's a damn good board.
 
For cheap, they are okay boards. I just don't like dealing with quirks in bios at any price, if I don't have to.

Link below is a pretty good read for what happens when trying to overclock Bulldozer. Seems a pretty good number are using the Gigabyte mobo including the Thread Starter. I did not read but about 4 pages but it was good reading for understanding without having to do many posts and REinvent the wheel as it were.

OC FX-8120 BIOS settings for a stable 4500 MHz
 
Well I set the multi to x23, changed LLC to regular and bumped the voltage to 1.4. As I thought, setting it to Regular gave me a somewhat accurate reading in CPUID. However, running P95 resulted in worker #4 getting an immediate summout error and the voltage I believe drooped to 1.37ish-1.39ish. So I stopped the test and booted back into BIOS.

Bumping the voltage to 1.425 gave me a reading of 1.440 idle in CPUID and 1.425 during P95. Worker #3 stopped after 2 mins with a fatal error.

Upon raising the voltage to 1.45, CPUID gave me a reading of 1.472 idle and jumped between 1.440-1.456 during P95. I was able to get about 28 minutes of P95 before receiving a blue screen of death.

After my failure at 1.45 in the BIOS, I've set everything to what it was at x22. Seems 4.4 is the best I'll be able to do with the board. So I'll just take what I can get and say thanks to all you that helped. Much appreciated.

*NOTE: Extreme LLC setting gave a 1.360 reading in CPUID with voltage set to 1.4 in BIOS.
 

Attachments

  • x23worker3fail.jpg
    x23worker3fail.jpg
    357 KB · Views: 53
  • x23p95FifteenMinsPassed.jpg
    x23p95FifteenMinsPassed.jpg
    360.4 KB · Views: 48
It's not that setting the LLC to regular gave a more accurate reading in CPUID it's just that it changed the amount of voltage supplementation over baseline. That's what LLC does and how it works. Gives more voltage to the CPU when it's under load to keep it stable.
 
Back