SOLVED FX 6100 is acting on me. trying to OC.

[Mandrake4565]

Cpu v is on Ultra High cpu/nb v on high.

But how in the heck are you to 'see' what the real differnce is when the cpu shows 100% load in either situation? Does a 100% load have a difference?

your guess is better then mine

 

[RGone]

Cpu v is on Ultra High cpu/nb v on high.

your guess is better then mine

Yeah that is what I generally use for those two LLC's also. Well crap, now I got to go try what just entered my feeble brain. I will take looksee with OHM since it monitors each core load. Maybe I can look at it and windows monitor and catch a minor difference.
RGone...

 

[drosera01]

Oh well,

Did try different stressing methods to check how it is acting this morning.
P95: CPU usage 100%-all cores, Clock throttles to 3GHz (x15) and back to 3.8GHz every other second voltage drops to 1.21v from 1.3v
OCCT Lin-pak: CPU Usage-100%-all cores; Clock throttles to 3GHz (x15) and back to 3.8GHz every other second voltage drops to 1.21v from 1.3v
IBT: Same as other two's above.
AIDA64 Built in stability test: CPU Usage 100%-all cores; Clock stays at 3.8 GHz and voltage stays 1.3v
F@H: CPU usage about 90%- Apparently default F@H client uses 5 out of 6 cores at 100% and other core stays at around 50-60%. This core could be anyone and keep changing every second or so. No throttling, clock speed always @3.8GHz. No drop on voltage as well.

Very carefully observed temperatures in all these testing: On any of the stress testing
- Package temperature never exceeded 31C (cores)
-TMPIN0 is 35-36 C (ambient MB temperature)
-TMPIN1 did not exceed 35 C (Socket temp aka CPU temp)
-TMPIN2 reached up to 41 C (NorthBridge temp)

Admittedly, I did not run either of thee tests for long enough like hours but about 10 minutes on each which showed the picture.

Definitely, its not a thermal throttle.

This picture of OCCT stressing shows CPU clock throttle.

 

[Mandrake4565]

interesting Drosera, I would think the reason for the throttling would be temperature, but that doesn't seem the case. Rgone would know better then me but is it possible that non of the temps that are monitored are the VRM area and that is what is causing the throttling. I would have to believe that, that section with no heatsinks on it would be getting really hot.

 

[drosera01]

interesting Drosera, I would think the reason for the throttling would be temperature, but that doesn't seem the case. Rgone would know better then me but is it possible that non of the temps that are monitored are the VRM area and that is what is causing the throttling. I would have to believe that, that section with no heatsinks on it would be getting really hot.

I doubt VRMs are getting that hot. LAst time I had pushed my hand through the ally to the MOSFETS to touch, if they are running hot, but were not.
MY case is still open, so let me run OCCT one more time and poke my hand to the MOSFETS to feel it. I will update in few minutes............

UPDATE: NO, MOSFETS are not hot at all, I just touched some of them and kept my figure touching it literally 2 minutes, no pain no burning feeling, just warm, lukewarm. I guess sure it is not more than 45-50 C or even less.

 

[Mandrake4565]

ok, I mentioned it because it's the only thing I could think of that could be causing your issue. Maybe a long shot but have you tested the rails on the psu to see if your getting good power?

 

[drosera01]

ok, I mentioned it because it's the only thing I could think of that could be causing your issue. Maybe a long shot but have you tested the rails on the psu to see if your getting good power?

I am definitely keeping eyes on the rails as well. Last BIOS picture on Post #11 shows 12v, 3.3v and 5v reading on BIOS and they are all pretty good. PSU is the thing I never compromise and I always get scared with people overclocking like hell whose 12v reads less than 11.75. This is Seasonic build Corsair branded 650W PSU.

I did check 12, 5 and 3.3v reading using different software as well and they all look good, I do not have multimeter to check though.
This is HWMonitor screenshot showing all voltages while running Prime95 on the back ground.

 

[Mandrake4565]

You said you had cool and quiet off as well as the other power saving features? I ask because I noticed your CPU V core dropped to .896.

 

[drosera01]

You said you had cool and quiet off as well as the other power saving features? I ask because I noticed your CPU V core dropped to .896.

Yes, All the green setting on my BIOS is off, I guess its the problem on BIOS not on CPU and definitely temperature related. It just does not have an option to keep from throttling on my BIOS.
Today, I set it up on default 3.3GHz and still throttles back to 3.0 on Prime 95 but not on F@H though.

thanks

 

[Mandrake4565]

Do you have windows in performance mode?

 

[drosera01]

Do you have windows in performance mode?

Yes, in high performance mode on power setting.

 

[drosera01]

ok,
As FX 6100 did not work well on my board, instead of buying another board, I changed FX 6100 to Phenom II x4 965.
by some reason stock voltage on 965 is whooping 1.475. That's a lot for stock 3.4 GHz, so I tried to reduce voltage and Was able to run Prime95 2+ hours at 1.35........... great.

Now time to start overclocking... immediately I noticed that 965 is not good overclocker either. I needed 1.425v to run it at 3.7 GHz Prime stable. I tried 3.8 GHz.... no go until I give it 1.5v then again one of the core appears to be very weak. on Prime95 while 3 other cores finish 150 tests the weak one does not even finish 70 tests. Its strange......... I had seen some cores running a bit slow like may be 5-7 or 10 tests but almost half is no good.

And temps at 1.5 v is already pushing temps to the max I want to see, so I pulled 965 out and installed Phenom II x2 555 BE, which easily unlocks to x 4 B55 and I can push to 3.8GHz with only 1.375v and very good temps.

What a bummer Fx6100 and 965 I got. I am back to original B55 and 965 is going back to Microcenter.

 

[Mandrake4565]

That's a bummer!