AMD 8150

[vanhline]

so i missed the part about the prime95; i have that loaded up now and running, within about 5 mins my CPU hit 70*C made its seemed to of leveled out there but its making me a bit uneasy so im gonna shut down the stress test and google a bit of info regarding safe temps for my cpu to be at before i continue.

 

[vanhline]

alright so i let it run about 20mins and here is the result

 

[RGone]

Okay I am going to take every scrap of information shown in your post #22 including your signature and explain how I interpret what is seen.

1. When a P95 Blend result is attached for viewing it should have the CPUz CPU Tab uploaded as well "OR" the speed of the cpu at the time of the HWMonitor should be specifically stated.

2. And HWMonitor gives us a view of a temp that we call CPU temp when monitoring most later Asus mobo's instead of some off the wall TMPINx number that we have to 'guess' what those temps apply to and it varies from Giga to MSi etc. Then we have a cpu package temp in later versions of HWMonitor which is the readout of the Core Temp or temperature of the internal cores. It is this core/package temp we are somewhat more concerned with but not to the exclusion of that other temp called CPU Temp on your Asus mobo.

3. In general the temp HWMonitor calls CPU temp is best kept to about 70c or less. In general the temp that HWMonitor now calls out as "package" temp should be kept to about 60c or less. Lately I have seen some say you can fudge those temps a little higher even, but for my part I think that you can fudge anything you want to fudge, but in the relation to heat...why would you want more heat to deal with? Even more heat to have to remove from the cpu and eventually from the case itself. But I have outlined what temps we have been using in this forum secion for 18mos and without undue strain.

4. If you have 69c CPU temp and 54c Package temp, then you are real close to the limits shown above. If that is truly 20 mins of P95 Blend testing AT the 4.5Ghz speed in your signature and will now run that same setttings and do so without a failure in P95 Blend for 2 hours and temps remain within 2c of what you are showing in the HWMonitor in your post #22, then you should be very well off for daily use.
RGone...

 

[Johan45]

I think that you can fudge anything you want to fudge,
RGone...

Not sure why but man did I laugh when I read that line
But Rgone is right this is probably the end of the line and if you intend to put the rest of your ram back in we may have to drop a bit since I'm certain we'll have to increase the NB_CPU voltage to accomodate that. In turn that makes more heat which you don't have much room for ATM.

 

[RGone]

But Rgone is right this is probably the end of the line and if you intend to put the rest of your ram back in we may have to drop a bit since I'm certain we'll have to increase the NB_CPU voltage to accomodate that. In turn that makes more heat which you don't have much room for ATM.

We have also got to remember that on the Bulldozer processors there was almost no need to go beyond 4.3Ghz. In the CPU forum section "ssjwizard" was one of the first to see the trend that there was little real performance to be had from pushing BD beyond 4.3Ghz. He had charts of such behavior as did I and then I did that thread on does FX-8350 seem to flatline beyond 4.3Ghz?, thread. Lots of graphing in that thread.

So then you try and put 32gigs in a system that is approaching its' max and there can be problems.
1. The cpu is near all it can run at X temp and the IMC is living inside that cpu.

2. The cpu heat is effecting the IMC. This makes the load on the IMC even greater it would seem.

3. So stepping back away from the max speed and thus the max temps to a more managable situation, can make 32 gigs easier to use.

4. The ram speed and ram amount has to be balanced against the cpu speed and the CPU_NB speed and voltage. Again with the IMC = CPU_NB, the temps and extra volts to the CPU_NB are related. Fully related.

5. As regards the ram speed to cpu speed and needed CPU_NB speed, the BD and PD processors really seem to like the DDR1866 ram. But here too you have to do the homework. AMD only warrants or specifies DDR1866 ram in "one" channel of the IMC and motherboard. That amounts to only two sticks of fast ram.

6. I believe that the BD/PD likes DDR1866 ram very well. Especially when the cpu is overclocked. The CPU is not waiting on the ram to move on. If that is a truth, then the slower ram makes the cpu wait longer the faster the cpu is run. I see this with my own parts and pieces and very similar conditions with others I speak with on a regular basis. Stuff we never get to cover in the forum because of all the time spent with "I cannot get my stuff to overclock". Not a bad thing just a reality perhaps. So here I write a bit about it.

7. So 32gigs of ram is not a picnic of plug and play with a greatly overclocked cpu. Impossbile? I did not say that, but only that there are a number of factors that figure into a fully workable overclock. Factors that, if ignored, can bring unexpected results. Perhaps slowing a cpu overclock can be beneficial to an attempt at large amounts of ram.
RGone...