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

After suggested hardware upgrade, how to consolidate it all?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
I have done most of my testing lately with Cinebench R 11.5 since they have a large data base of scores for the CPU test and the Video test. I don't bench ram specifically so I have no preference there. I can tell if my ram is overclocked too far because of failures in stressing or other benches.
RGone...
 
pzykotik said:
Ugh. Just realised that ever since the last picture, the CPU isn't overclocked AT ALL. Will...be...back...later...
Click to expand...

I thought the image showed no overclock but took you at your word since there are times when certain situations will lower the clock speed and not show the overclock.
RGone...
 
Must've hit a default button and didn't realised. All that time I thought it was OC'd @ 4.3ghz. So I took it back to 4.3 @ 1.65v, failed prime95.
Then I've came up with this as 20min prime95 stable, which is not too bad.

The only thing that bothers me is the temps... wasn't it that amd was 15°C off, and true tolerance was up to 70°C? If so, am I in a dangerous zone here?

cpu-ram-overclocked.jpg

Will try some games with it, report later.
 
I think after all the cussing and discussing was over with it is pretty well understood that the cpu makers are not in the thermometer business and at lower temps the readings are off by maybe 10c but as the limit is reached, the temps are more accurate. I know AMD or Intel neither is concerned about lower temps. They are determining max temps to shutdown at not what is an actual lower temp.

Somewhere in the area of 62c core temp is when the internal cpu temp controls may step in and begin to clock back to keep higher temps from hurting the cpu. That is what we have come to understand now.
RGone...
 
Last edited:
Ever since the overclock, I've experienced wierd problems.

First of all, keyboard's ctrl and shift being stuck like if I pressed them, when I was not.
Changed the keyboard, same problem. Rebooted, was just fine.

Then, sound card crashed while playing a game, which crashed the comp. Once rebooted, it was fine.

What would be the cause of that?
 
If the issues do not repeat themselves time after time, then I would not worry. If you reboot and the issues return 'after' awhile, then you have a problem with too much overclock at high temps or instability of some kind.
RGone...
 
The issue about the keyboard was fixed as soon as I rebooted, but I will look into that. So far, it didn't repeat. As for the soundcard, I didn't think the drivers would be a problem, since it didn't cause a problem before the oc. Just updated them, and fixed.
 
The issue about the soundcard crashing the whole pc repeated itself. I took off the cpu oc for now, and until I find a solution (which might be buying a soundcard, maybe onboard soundcard doesnt support well the oc?)

After reading even more on the OC of the FX-4100, I have found a thread were you explained your own way of overclocking, as well as some references around the lines of "some cpu might be bad apples".

So far, none of the overclocks that I have done have been satifying. Either it was crashes even after 24h of prime95, devices failing to respond or loss of speed instead of a gain.

The -only- all-around stable oc that was possible was the auto-oc from the motherboard (CPU@ 18x216 3.9ghz 1.3v, ram @ 1439mhz)...

Am I doing something wrong, or is it possible that I just happened to have that "bad apple"?
 
Last edited:
Okay let us get as honest about some things as we/I best understand them from hours and hours of typing helps and also testing myself.

You write >> The -only- all-around stable oc that was possible was the auto-oc from the motherboard (CPU@ 18x216 3.9ghz 1.3v, ram @ 1439mhz)...

Answer >> You check that Vcore in HWMonitor and I can almost assure you it is not just 1.3V Vcore. That board is auto changing the Vcore to a higher value underload. Heck do an extended testing of the Auto Overclock. It might be the only way to use that board.

You write >> Am I doing something wrong, or is it possible that I just happened to have that "bad apple"?

Answer >> Something wrong? Probably not but IT HAS been a long time since you posted the full complement of four captures individually so we can see what was going on inside the case. Often things are not as they seem until we all see the real deal in the four captures.

Bad Apple? Yes and No. There is a thread from a few weeks or so ago in the AMD motherboard forum section and a user was in there with a M5A97 and FX processor. I think it was a 6 core though. He said he knew the M5A97 was going to die overclocking an FX processor and so it DID die and he wanted to know what to get to replace it.

So bad apple? Probably just a board that is not at all suited for any FX processor. They are just FINE for Denebs and Thubans but the FX processor just draws so much current thru the weak (for FX processors) VRM circuit. That might be the major issue with not enough power left over for the rest of the motherboard. OR it could be a weak power supply that is multi-railed on the +12V rail and the split rails with low amperage are not able to hold-up the +12V rails from which the VRMs draw their power. And there is one ore thing that just hit me after reading your list of parts. Kingston ram has proven over the last months to be just a pain on these AMD boards. You have to make sure the ram is set by the SPD of the ram and I would have no doubt to add 0.05V to the ram voltage. Like I said it has been sometime since we saw the four screen captures so most of what is going on now has not been pictured lately. We are just guessing from old data likely.

So let me just be straight-up about my suspicions based on hours and hours of my own testing and no telling how many help posts...I would not have an FX processor in an M5A97 motherboard. I prefer single rail power supplies for the power hungry FX processor so there is no issue with load balancing the +12V rails. Otherwise you have to buy a very hefty power supply to know you have enough power on the CPU +12V input. And I cannot invision Kingston ram over my G Skill ram on the later AMD motherboards. That is about it overall.
RGone...
 
Last edited:
Alright, basically, just a bad match overall. Good news it is.

As for the screenshots, I posted some by the 16th, which were the -almost- stable settings, but those which gave me some trouble with the soundcard.

As for running tests with the auto-oc, I ran prime95 for an hour, no failures. And if by extended you mean 24h... maybe at some point when I'm not home :)

I can post some ss of the current settings(auto-oc), if thats what you're asking.

/e: autooc.jpg
 
Last edited:
That screen capture is "exactly" what I wanted to look at. You say that is the Auto-Overclock setting with HWMonitor running in the background logging Min/Max temps and voltages while P95 Blend was running???

IF the answer to my question above is yes, then the problem is likely one of the suspicions I voiced above. The ram is a problem. You see that they downclocked the ram base setting and used the FSB/HT Ref setting to upclock the cpu and the ram at the same time. The Auto-overclock mechanism has the ram running way below its' rated speed and thus the system appears stable.

Not 24 hours of P95 Blend mode but realistically at least 2 hours or so to give a good pointer towards >> yes or no for stability.

Now based on what I see above, I know what I would do if I thought to overclock anymore. But there is a caveat to further overclocking...the VRM circuit of the board maybe unable to supply more cpu voltage accurately and without the board perhaps beginning to die. To me that possiblity is so relevant that I might just leave it where it is if the P95 is not failing for 2 hours in Auto-overclock mode. The way the cpu is running now it is doing so with ALL four cores on ALL the time and at 3.9ghz. That is a good bit more powerful than that cpu running as it was designed.
RGone...

EDIT:
Just in case you did not catch it when I said it above...it looks like the suspicion of the Kingston ram as creating a problem, is about 95% accurate.
END EDIT.
 
It will be either this effort-less auto-OC or (if any problem shows up) no overclock at all. It was just a bad match of component overall, and my next build will take MUCH more time to buy (I didn't search for compatibility issue, I bought what my money could buy at the time...).

This is the end of my overclock journey on this rig. I hope that this thread will help some people when they will search for the FX-4100.

Thanks to everyone involved.

For searches purposes, since my sig will most likely change over time:

MOBO >>>>>>>>>> ASUS M5A97
CPU & Cooler >>>>> AMD FX-4100 rated 3.6Ghz , OC @ 3.9Ghz & Kühler H2O 620
MEMORY >>>>>>>>> Kingston HyperX Genesis 4x 2gb rated 1333MHz 9-9-9-25-34, 1.5v , OC @ 1439Mhz 9-9-9-24-33
POWER SUPPLY >>>> ThermalTake TR-600, 600W /w 120mm fan
VIDEO CARD >>>>>> 1x Gigabyte Radeon HD 7750 1GB GDDR5
HDDs / Optical drives > 1x 3Tb Seagate Barracuda 7200rpm + 1x 350Gb SATA
SOUND DEVICE >>>> Onboard
USB >>>>>>>>>>>> 1x 1Tb External Drive
OS >>>>>>>>>>>> Win7 Home Prem x64 SP1
Case >>>>>>>>>>> Antec Two Hundred Three /w 4x Blue LED 120mm + 1x 120mm LOW/HIGH + 1x 140mm LOW/HIGH
 
Last edited:
Reserved for PCMark7 Results, HWMonitor after some Prime95 and CPU-Z screenshots:
1. Failed OC
2. Auto-OC
3. RAM OC only
4. Default
 
You must log in or register to reply here.

Similar threads

M
How difficult is it to run 24 GB of fast RAM on X58? (X58 / Socket 1366 / Westmere)
Replies
9
Views
234
freeagent
freeagent
U
Upgrading to AM5, G.Skill vs Kingston RAM?
Replies
2
Views
581
Woomack
W
R
Bad CPU, Bad Mobo, or 3 Bad Ram Kits?
Replies
14
Views
475
Rewire
R
mackerel
X299 and 7980XE in 2024
2 3 4 5
Replies
90
Views
3K
TechStuff
T
N
Intel 13th vs 14th gen and QVL
Replies
2
Views
878
nstgc
N
Back