Phenom 2 1100T High Temps on Low Voltage

[Bioshock]

Hey guys, needing to get some input from other people who own this processor.....first off let me start by saying i think i got a really good chip in the batch because i am perfectly stable at 4.1Ghz using 1.4v. this chip pushed 3.8 before i needed to add any voltage to it all at to keep it stable........that includes a bit of FSB OC to push my ram a bit further, and a NB OC of 2665Mhz. Voltages being put into the CPU/NB are 1.25v.......

now for most reviews ive read and OC benches ive seen around the internets, it seems people are hitting a wall around 4.1 - 4.2 with these chips, and to reach those speeds they have to pump 1.55v into them to get there........the guru3d review specifically said, quote:

"Here's an overclock at 4.2 GHz 100% stable. We boosted Voltage towards 1.55 V in the BIOS and simply applied a multiplier of 20. Temperatures are now higher, but really acceptable for an AMD Phenom II at just under 50 Degrees C with the core peaking to 55 Degrees C. Again, and I can't stress this enough -- we are only using a 35 USD air based Vendetta cooler here, nothing fancy."

which leads me to my question.

at 1.4v under intel burn test and prime95, im maxing out my load temps at 55-57c, while idling between 28-29c......

i have a Cooler Master V8 sitting on the 1100t with an aftermarket fan replacing the stock fan that came with it....all of this is inside an Antec 902 case, so airflow is just about as good as it can possibly get. Cable management is well done and nothing is blocking any airflow inside the case, Ambient temps are at 72f.

Arctic Silver 5 TIM is under the V8, as well as the NB and SB heatsinks on the MB, which is an MSI 790fx-gd70

i think i should mention i just got this chip 3 days ago as well as the case. before i was running a 720be using the same motherboard and v8 cooler (without the aftermarket fan at the time)......i had the 720 OC'ed to 3.6Ghz running 1.52v to it and my load temps never reached 50c....and that was in an NZXT Guardian 921 case, which has substantially lower airflow compared to the Antec 902. Being as the 2 chips run at the same voltage (i know the 720 was a 95w chip and the 1100t is 125w so there would be a bit more heat) i figured temps should be relatively close between the 2, at least not as high as they are under stress @ 1.4v right now

i should also mention that i have reseated and re-applied the AS5 paste under the V8 twice just to make sure there were no air bubbles or bad contact between the cooler and processor

so given that information, having AS5 paste on all the components, best airflow possible, and the lowest volts ive seen pushing one of these chips at these speeds, what can be causing the load temps to shoot up that high under stress?


below i have attached a picture of CPU-Z showing the clock speeds and voltages for your reference

thank you guys for taking the time to read this and i welcome any information you all may be able to provide!

 

[DigitalUK]

as5 needs to cure and no way it could have done that in 3 days i prefer mx-2/3. try droping your cpu voltage (will lower temps) and turning up your nb to at least 2200 and nb voltage up as well.

 

[Law-II]

Hi

I do not wish to sound rude; but you have essentially answered all your own questions

Summary
-your new CPU is 125w
-your V8 was good for 95w
-as5

Things you could change
-Add a more effecient fan to your V8
-MX4 or another non curing TIM

nb: when you fitted your aftermarket fan did you check that the airflow is going from right to left on your v8 so that it vents hot air to the rear of the case.

-What aftermarket fan did you fit? dose it exceed the reference fans sepc's of 69.69CFM?

http://www.coolermaster.com/product.php?product_id=5279

atb

Law-II

 

[LordJummy]

Get some SHIN ETSU X23-7783D. It has 0 curing time and works wonders.

 

[repman244]

I wouldn't worry about 55C on prime95 (you will probably never get there with real world applications), second thing you should consider when comparing to others is that the ambient temperatures are definitely different (did guru3d state the ambient temperatures?).
And like DigitalUK said AS5 needs to cure.

 

[techtard]

AS5 is pretty old school. There are lots of pastes that are better, even if only a little. The pluses of the newer pastes : No cure time, non-capacitive and non-conductive.

 

[scaminatrix]

Unless you're a competitive OC'er, the TIM you use makes pretty much no difference. That's why everyone's getting masses of the cheapest stuff rather than spending $xx on a couple of grams.

Also, the curing time myth has been pretty much debunked

We also debunked the "curing time myth": at least with the compounds we tested, the performance of a recently-applyied thermal compound was the same as after the "curing time". We cannot claim that no thermal compound needs this curing, but the "curing time" seems to be a negligible factor when testing CPU coolers or thermal compounds.

http://www.hardwaresecrets.com/article/Thermal-Compound-Roundup-March-2011/1207/7

I'm leaning towrds Law-II's answer.

 

[LordJummy]

Unless you're a competitive OC'er, the TIM you use makes pretty much no difference. That's why everyone's getting masses of the cheapest stuff rather than spending $xx on a couple of grams.

Also, the curing time myth has been pretty much debunked


http://www.hardwaresecrets.com/article/Thermal-Compound-Roundup-March-2011/1207/7

I'm leaning towrds Law-II's answer.

Actually they can make a huge difference, and a lot of the difference depends on the type and quality of the actual base on the heat sink. I tested 5 different TIMs with the H70 and saw a massive difference between AS5 and shin etsu. The AS5 applications saw over 8c increase on each core vs. shin etsu x23-7783d.

 

[DigitalUK]

admittedly we are talking 1-3c between some white gunk and something more expensive but it does make a difference, also the clean white stuff breaks down after about 1-2years specially when its been on something that runs hot. also if the curing time was a myth why would they put it in the instuctions for as5 as it could put some users off.

 

[micropage7]

cooling performance affect by many things, like thermal paste that you use and how you place it, how good the contact area between processor and hsf, the hsf performance itself, airflow and your room temperature, the fan you use, how tight between hsf base to processor etc
so its more complicated so check it item per item

 

[scaminatrix]

I'm just reporting what hardwaresecrets has reported, and they are a reputable site so I think I'll go with them on this one

 

[repman244]

BTW I also have a question about TIM and my CPU. Is it safe to put a blob in the center and press the cooler so it spreads by itself since I won't know if it will cover the CPU die bellow IHS? I usually used a card and spread it myself but I want to try the blob and see the difference.
I'm worried about the TIM not covering the die completely since its 6 cores and the die is not square but rectangular.

 

[Law-II]

BTW I also have a question about TIM and my CPU. Is it safe to put a blob in the center and press the cooler so it spreads by itself since I won't know if it will cover the CPU die bellow IHS? I usually used a card and spread it myself but I want to try the blob and see the difference.
I'm worried about the TIM not covering the die completely since its 6 cores and the die is not square but rectangular.

Hi

Look here http://www.youtube.com/watch?v=EyXLu1Ms-q4

nb: use a blob no bigger than a grain of rice

atb

Law-II

 

[repman244]

Hi

Look here http://www.youtube.com/watch?v=EyXLu1Ms-q4

atb

Law-II

Thank you for that, I'm using AS5, will that spread like the AS3 showed in the video or will it be like the first that's shown?
If I applied AS3 like that it would not cover all 6 cores IMO (or would it?).

 

[Law-II]

Thank you for that, I'm using AS5, will that spread like the AS3 showed in the video or will it be like the first that's shown?
If I applied AS3 like that it would not cover all 6 cores IMO (or would it?).

Hi

If you use AS5 in the shape of a grain of rice in parallel to the six cores it should work fine; although it needs a little more pressure to spread it will get this from your heatsink once it is secured

I cannot comment with authority on AS3 spreading in a similar way to AS5; however I have a feeling that the numbering has to do with the purity and not the consistency of this TIM, you may have to look in to this.

nb: I do not use AS5 anylonger, I have converted to MX4

atb

Law-II

 

[Bioshock]

wow. thanks for all the replies everybody. i had this same post on guru3d for 2 days and not one reply. i had a feeling i shouldve came over here to begin with.....so thanks alot for taking your time to discuss this with me

with that being said, i have also tried MX2 and 3 paste. honestly, from my results, the AS5 performs better and doesnt turn to cement after a few months of really high temps......i know, i was confused by that too....i gotta say that the paste isnt gonna help my issue.......a difference between 1-2c is beyond what im dealing with under load. i simply dont see why its going that high at 1.4v while every single review or OC'er ive seen has control of their temps even at close to the max volts these chips take.....just as a reference, when running prime or IBT, withtin 3 seconds of clicking start on the test, the temps jump from idle around 28c to 45c. withtin another 5 or so seconds im at 50c, on up to 55-56c where it holds steady.............but i do welcome the suggestions on the TIM. so thanks for that guys

in response to Digital UK's post, my NB is sitting at 2665 as per the picture i posted.

Law-II, haha, yea the fan is blowing the right direction......its an OCZ (not sure what model etc) but it moves a ton more air than the stock V8 used to. when cranked up full speed by the fan controller in front of my case it really moves.........would you care to link me to a spreadsheet or a site that says the V8 is rated only for 95w cpu's? i have never heard that before, considering the guru3d review of using a basic el-cheapo cooler and keeping temps under load below 55c @ 1.55v i dont think the V8 should be falling that far behind

again, thank all you guys for taking time to go through this with me. it very well could be the v8 if it is only rated for 95w, but that has to be something i looked over and never heard of that before. for such a big cooler i dont see how it couldnt keep a 125w cool at the volts im running.

ill look into the heat rating of this thing and see if i can find any more info about that.

any more suggestions are welcomed. thanks again



edit: doing a quick search, the v8 is rated to cool up to 180w according to the CM website and a review on oc club


sources:

http://www.coolermaster.com/product.php?product_id=5279

http://www.overclockersclub.com/reviews/cm_v8/

that being said, the guru3d review said under stress, the 1100t can draw up to 225w of power at max voltage. again, 1.4v is far from that. i need a way of testing the draw under load at these volts but thats something ive never attempted. will a simple volt meter do the trick or will i need a more sophisticated approach to do that?


edit 2:

heres a post of a guy getting over 4400mhz using a v8 @ 1.56v...........not that i would dare do that.....but just a reference that the V8 can cool this chip under stress adequately

http://www.eteknix.com/forum/cooling-overclocking/1100t-challenge-3768/

 

[Law-II]

wow. thanks for all the replies everybody. i had this same post on guru3d for 2 days and not one reply. i had a feeling i shouldve came over here to begin with.....so thanks alot for taking your time to discuss this with me

with that being said, i have also tried MX2 and 3 paste. honestly, from my results, the AS5 performs better and doesnt turn to cement after a few months of really high temps......i know, i was confused by that too....i gotta say that the paste isnt gonna help my issue.......a difference between 1-2c is beyond what im dealing with under load. i simply dont see why its going that high at 1.4v while every single review or OC'er ive seen has control of their temps even at close to the max volts these chips take.....just as a reference, when running prime or IBT, withtin 3 seconds of clicking start on the test, the temps jump from idle around 28c to 45c. withtin another 5 or so seconds im at 50c, on up to 55-56c where it holds steady.............but i do welcome the suggestions on the TIM. so thanks for that guys

in response to Digital UK's post, my NB is sitting at 2665 as per the picture i posted.

Law-II, haha, yea the fan is blowing the right direction......its an OCZ (not sure what model etc) but it moves a ton more air than the stock V8 used to. when cranked up full speed by the fan controller in front of my case it really moves.........would you care to link me to a spreadsheet or a site that says the V8 is rated only for 95w cpu's? i have never heard that before, considering the guru3d review of using a basic el-cheapo cooler and keeping temps under load below 55c @ 1.55v i dont think the V8 should be falling that far behind

again, thank all you guys for taking time to go through this with me. it very well could be the v8 if it is only rated for 95w, but that has to be something i looked over and never heard of that before. for such a big cooler i dont see how it couldnt keep a 125w cool at the volts im running.

ill look into the heat rating of this thing and see if i can find any more info about that.

any more suggestions are welcomed. thanks again



edit: doing a quick search, the v8 is rated to cool up to 180w according to the CM website and a review on oc club


sources:

http://www.coolermaster.com/product.php?product_id=5279

http://www.overclockersclub.com/reviews/cm_v8/

that being said, the guru3d review said under stress, the 1100t can draw up to 225w of power at max voltage. again, 1.4v is far from that. i need a way of testing the draw under load at these volts but thats something ive never attempted. will a simple volt meter do the trick or will i need a more sophisticated approach to do that?


edit 2:

heres a post of a guy getting over 4400mhz using a v8 @ 1.56v...........not that i would dare do that.....but just a reference that the V8 can cool this chip under stress adequately

http://www.eteknix.com/forum/cooling-overclocking/1100t-challenge-3768/

Hi

"would you care to link me to a spreadsheet or a site that says the V8 is rated only for 95w cpu's?" I cannot do that as I linked you to a site that said you can; all I was trying to suggest is that the V8 coped better with 95w under load as mentioned in your original post; as you are now stressing 125w you should expect to see a rise in heat. I apologise if you thought in any way I was suggesting anything different.

nb: by the way your temps are good under load with Prime 95

atb

Law-II

 

[repman244]

Bioshock if you hit ~55C with prime95 that is perfectly fine. I would worry if I saw 59-60C.
I think V8 is good enough for your chip, don't forget when comparing to the guy that posted the chip at 1.56V can have different ambient temperatures than you have (while I'm writing this my CPU is at ~36C idle because of my ambient and I only have it at 3.4GHz and 1.224V).

@Law-II: Thank you for the reply, I'll try pressing it with a piece of glass like on that video and experiment a bit to see how much to use so I cover the die correctly. But I don't know how the die is rotated tho, is it vertical (vertical to the letters and numbers on the CPU) or horizontal [-] ? (my guess is vertical but you never know).

EDIT: never mind I found some pictures of the CPU without IHS

 

[Bioshock]

Hi

"would you care to link me to a spreadsheet or a site that says the V8 is rated only for 95w cpu's?" I cannot do that as I linked you to a site that said you can; all I was trying to suggest is that the V8 coped better with 95w under load as mentioned in your original post; as you are now stressing 125w you should expect to see a rise in heat. I apologise if you thought in any way I was suggesting anything different.

nb: by the way your temps are good under load with Prime 95

atb

Law-II

hey law, my mistake.....i should have clarified in my first post that the reason im concerned about my temps is because of the headroom i have left. i could possibly push more volts into it and get more speed before i hit a wall on mhz.

that is my main concern with my temps. i know in real world gaming or cpu intensive apps that i wont come near the temps a stress test would produce, so theoretically i should be safe pushing a bit more into it and oc'ing a bit more and be fine while gaming but i wouldnt be 100% sure if it was stable without being able to stress it because of the added heat...........but id like to see how far i could push this chip if i were able to add a bit moar juice to it

it isnt a big deal really, id be happy with 4.1ghz at 1.4v if thats all ill be able to reach given my current cooling, just curious to see where i could get at 1.5v or so

for now ill stick to what im running and be happy i got that far at 1.4v

thanks for the input guys. ill be sure to post here first before posting any where else in the future......its always good to get fast responses and move on

thanks again

 

[Law-II]

Hi

thanks

good luck

atb

Law-II

 

[Pestilence]

I'm just reporting what hardwaresecrets has reported, and they are a reputable site so I think I'll go with them on this one

Was that a Joke? Hardwaresecrets? LAWL

 

[scaminatrix]

You've pretty much proved my point. IMO 6 deg.C difference isn't worth double the price, unless, like I said, you're a competitive OC'er.
We're getting off topic here anyway. Sorry all.

 

[techtard]

6 degrees is gigantic, especially when it's just the TIM making the difference. Pair that up with a better cooler, or better fans and you'll get a more stable oc, maybe even a little higher max stable.
AMD cpus love the cold. The cooler the better.
Spend a few extra bucks on better thermal paste.