Stress Testing: Post Your REAL Temperature Challenge

[simoko]

---

thanks master, good to know gai mi pointers how conquer 222 bclk bclk? unsa limiting factor ani..ang board?procie or ram?

glide, dagkoti nanang pugon sa bibingkahan diha kay magsugod napud ko ug tuno sa lubi hehe

@ centax: onsai mobo nimo? kadtong hinampak nako nga Bloodrage 216 ra gyud ma kaya....(hehehehe wala btaw tu kasabot ang among nibor nga nakapalit unsa ni ato storyahan)...bai, w/ that MSR tool diay i sent u malingaw gyud ka...u can even set different multipliers on the different cores & verify it w/ CPUz...

@ DerFuhrer/FU/vashtinone: nangawas na gyud ang mga Ozawa slaves...form unya ta atong fans club...si Der ato president, ikaw vash ang bise, si FU ang Sec...ako lai tig-research/post stories,videos & pics

by the way...DILI LAGI KO MASTER....slave lagi ko ^_^

[glide 1]

Finally, I was able to bring down max temp below 80 deg but had to strap the Delta screamers. DK and Thunderblade just doesn't cut it below 80 mark. This is how far I can push it at the current cooling set up. Next would be to lap the cooler and get a very good thermal compound and prolly strap in 2 Deltas and give 4.3-4.4 a shot if possible tempwise.

i7 D0 @ 4.2 ghz
Asus RE 2
GTX 295
Corsair Dom DDR3 1600
Silverstone ST 1000
Prolimatech Megahalem
Vista 32

after 8 hrs of P95


idle temps

nel - very nice results, lupig pa akoang h2o cooling ani 4.3 - 4.4 next nel!!!

simoko - very informative ni inyong discussion diri and is a big help namong tanan. Most of it i cannot grasp at the moment pero with a bit of reading and experimenting maka sabot ta eventually.

[vashtinone]

@sir simoko..
hahaha.. dli ko oi, ako lng muse.. hahaha..lol..

OT na kaayu ta dah, masuko na si master poldo ani iya raba ni nga thread.. heheh.. PEACE!!

[poldopunk]

@ simoko:

i too, have been basing my temp analogies from the datasheet pero sa E4000, E6000 and Q6600 series lang (65nm), the Q9X50 is 45nm so the measurements may differ (wa lang ko motan-aw kay gikan pa ko work, giuna pa ang mga pr0n updates ). from how i understand these assertions, my processor which is an E4500 has a tjunction max of 85c although these are not official since Intel has contradicting releases on the REAL tjunction max for some reason and the 85c are taken from calibrations made by CoreTemp, RealTemp and other known authors of these utilties.

As for #PROCHOT and TCC, TCC is also activated once the processor is on #PROCHOT status hence why my system generated a stop error (hardware failure - 0x21 and restarted because it started lowering down clockspeeds in an attempt to cool down the temps which resulted in instability and poof, restart. RealTemp actually reports the status from the bottom ('Thermal Status') and I was actually aiming for a screenshot for #PROCHOT but as per Intel's datasheet, it only lasts for a short period if the thermal solution is adequate and well-designed. For some reason the temperature may have spiked up higher than RealTemp and CoreTemp is reporting since as I've noticed because as far as I can remember, I was still at 75c at that time.

Truth be told, I'm not really as scared with temperature as before because these features are very well implemented despite the fact they were never designed for accuracy but protection. My goal right now is to get a screenshot of a #PROCHOT status followed by a #THERMTRIP. Of course no one can take a screenshot of the #THERMTRIP assertion since the computer would be off but I still wanted the experience. My original statement was that the processor started throttling 5-10c from the assumed tjunction max at that time which was expected since the processor is on #PROCHOT and that the TCC has been activated.

[poldopunk]

Crystal CPUID's MSR Editor to determine the accuracy for temperature is very interesting. Case in point:

IDLE



2C = 44
Assumed Tjunction Max = 85c

x = Tjunction Max - Hexadecimal Value
x = 85c - 44
x = 41c

Exactly as how CoreTemp and RealTemp reports it. They must be using this formula/method as well(?). Let's check out load with Prime95 Small FFTs.

LOAD



19 = 25
Assumed Tjunction Max = 85c

x = Tjunction Max - Converted Hexadecimal Value
x = 85c - 25
x = 60

Both RealTemp and CoreTemp are reporting the same value...dead accurate BUT the question is, are we correct to assume the Tjunction Max at 85c for E4500(?), only Intel can answer.

[COMCAM]

OT: mu take na jud ko ug OC class ani kang Simoko-San,

Simoko-San, when will OC class start?

[vashtinone]

OT: Simoko-sensie teach us Ozawa-no-jutsu.. hahaha...

now back to the topic.. PEACE!!

[isyot_daut]

(lain man gud nga klasing OC ako lingaw karon...hehehe OC=Ozawa's Compulsion)

ahaka maau pa kaau ko og basa nya ni kalit lng og liko...pwerte nako katawaa da...hehe

bogoya oi....heheh

nice one boss...heheh

[simoko]

@ poldo: the MSR value extraction is usually the method use by these temp monitoring software to arrived at what they display...the TJunctionMax & other values are normally found in the processor datasheet from Intel...

@ COMCAM/vashtinone: dili lang ko motudlo anang OverClocking oi...noob pa ko ana...dili lang ko pa-tuga2x...pero kung parti pa nuon na ni Maria O...pwede kaayo bisag usa pa ka semester

[Kenshiro]

Wow very impressive overclocks. Too bad for me I'm on a tight budget and I can't risk breaking my CPU. Since i really don't have a anything to buy a new CPU if it burns out. It would be good if I had a spare CPU.