The AMD Phenom II OC'ers Club

[erocker]

Other than my cores (which seem to be somewhat accurate) temps are always different.

The Northbridge/Southbridge temps in Everest seem to be correct. I should have my CHIV waterblock in this week and I'm interested to see if things will change... they better!

 

[Corduroy_Jr]

Yeah motherboards have to do with it as well.

mine under full load

 

[AphexDreamer]

Other than my cores (which seem to be somewhat accurate) temps are always different.

http://i403.photobucket.com/albums/pp112/erocker414/amdtemps.jpg

The Northbridge/Southbridge temps in Everest seem to be correct. I should have my CHIV waterblock in this week and I'm interested to see if things will change... they better!

Hey whats your performance like with your 965 and Mem at those settings? I'm trying to find out if I should aim for lower timings instead of a higher frequency and your running pretty much the same speed as I (4095mhz) 1680 9-9-9 or should I aim for 1333 cas 6.

 

[mastrdrver]

Other than my cores (which seem to be somewhat accurate) temps are always different.

http://i403.photobucket.com/albums/pp112/erocker414/amdtemps.jpg

The Northbridge/Southbridge temps in Everest seem to be correct. I should have my CHIV waterblock in this week and I'm interested to see if things will change... they better!

My MSI 890GX board reports the cpu temp 5C higher than core but I only pay attention to core temps on my 955BE since it is a C3. I had a 555BE that was accurate on the cores with in 2C of the cpu temp on my Gigabyte 790ftx. 955BE show the same variance. The three C2s all report core temps off by 5C at minimum on the same board at idle. On those, cpu temp has always seemed to report closer to reality.

My 435 will report an ambient temperature when at idle. I know the D14 is awesome, but even I don't believe it is that awesome.....though it is just a X3 so who knows!

Btw E, how is that 1.8v Crucial stuff working out for you? Is that the same stuff that Darwin is selling in his thread?

 

[cadaveca]

The difference between "core" temps and "socket" temps has nothing to do with anything other than how effective your cooling is. The better the cooling, the closer these two numbers will be.

 

[mastrdrver]

Well then my cooling (being the same) is somehow less efficient with the 435 (as an Athlon X3) than with the 955BE as the separation is higher with the C2 than with the C3.

 

[cadaveca]

You cannot compare across generations of chip, as calibration points may have changed, though.


It's best to take the same cpu, and try different cooling methods, to see this. Because each chip's leakage is different, even comparing different chips doesn't show the real picture either. Thuban naturally reads low, for example...so we know the sensor doesn't work the same there, for sure.

 

[TheMailMan78]

I built a computer for a friend of mine this weekend. I installed a 955 with a 1090T cooler. Under OCCT it hit 56c after an hour at 100% load. Stock clocks and volts. Please keep in mind my house was at 79f (26c) but does this seem normal?


Also here are my temps for my 1090T with 1.47v

 

[Chicken Patty]

I built a computer for a friend of mine this weekend. I installed a 955 with a 1090T cooler. Under OCCT it hit 56c after an hour at 100% load. Stock clocks and volts. Please keep in mind my house was at 79f (26c) but does this seem normal?


Also here are my temps for my 1090T with 1.47v
http://img.techpowerup.org/100713/Untitled2.jpg

Core temp is about 10ºc lower. PC Probe and AOD display more accurate temps. Core temps are off on 1090T's, gotta go by socket temp.

 

[cdawall]

If it doesn't crash temmps are ok socket temps are not accurate nor are the on chip diodes if you want cpu temps get a thermal probe.

 

[Chicken Patty]

That's true, but I rather guide my self by the higher temp just in case. Speaking of probes, my board brings temp sensors. Wouldn't that help is placed properly?

 

[cadaveca]

you'd not get anything more accurate than the socket temps, unless you cut a slot in the IHS for the probe. I'm using the probe headers on my CH3 to regulate fan speeds.

 

[Chicken Patty]

Hmm, crap! maybe I should use them for that as well. Where do you have you're placed?

 

[cadaveca]

I have one on memory to control memory fan, one on vga backplate to control fan blowing over the cards, and the third is smushed into the NB heatsink, to give temps there, but is not controlling the fan.

 

[Chicken Patty]

Thanks bro. This gave me some ideas

 

[cadaveca]

No problem! The functionality has proven quite useful for me....but is kinda limited...yuo can set 100% speed, percent from 40-90 in 10% steps, or set a temp range from 25c to 65c(I have vga fan @ 35min, 65c max/memory @ 30/55).

 

[Chicken Patty]

Ill def check this out when I get home.

 

[erocker]

Hey whats your performance like with your 965 and Mem at those settings? I'm trying to find out if I should aim for lower timings instead of a higher frequency and your running pretty much the same speed as I (4095mhz) 1680 9-9-9 or should I aim for 1333 cas 6.

I much prefer lower latencies. I like running my CPU/NB at double the frequency of my ram, so I'm running my ram at 1400mhz and my CPU/NB (memory controller) at 2800mhz. Give it a try, I think you'll like it.

Btw E, how is that 1.8v Crucial stuff working out for you? Is that the same stuff that Darwin is selling in his thread?

I bought this stuff new back when I bought my M4A79T Deluxe. This stuff is great, I've tried lots of other ram and always go back to this stuff. Lol, I paid $64.99 for a 4gb set. Prices sure have gone up!

This is what I get with them (max) at 1.85v (all 1T)

1450mhz-ish 6 6-6-18
1680mhz 7 7-7-21
1860mhz 8 8-8-24

Since I need quite a lot of voltage for my memory controller, I prefer keeping them around 1400mhz and the CPU/NB at 2800mhz.

 

[TheMailMan78]

Ill repeat. Does this seem normal?

I built a computer for a friend of mine this weekend. I installed a 955 with a 1090T cooler. Under OCCT it hit 56c after an hour at 100% load. Stock clocks and volts. Please keep in mind my house was at 79f (26c) but does this seem normal?

 

[Corduroy_Jr]

with stock cooler or corsair H50?

 

[cadaveca]

Ill repeat. Does this seem normal?

Yes.

 

[TheMailMan78]

with stock cooler or corsair H50?

A 1090T stock cooler on a 955.

 

[cadaveca]

Cooler design doesn't provide enough airflow to be very good. drop 10-12c or so with a TRUE, but more with water...

I tihnk this is why Thuban temps read that much lower, to prevent throttle...they just simply can handle higher temps. The larger core surface contact with IHS may account for this alone, though, and has thrown off the sensor they didn't both to re-design...

 

[TheMailMan78]

Cooler design doesn't provide enough airflow to be very good. drop 10-12c or so with a TRUE, but more with water...

Hes not OC. I just wanted to make sure he was safe. I personally didn't remember it running that high under load. I gave him a Xiggy "Red Scorpion" to put on later just in case.

 

[Corduroy_Jr]

temps are reasonable yes, try undervolting processor?