Sign in with
Sign up | Sign in
Your question

PRIME 95 TEMPS

Last response: in Overclocking
Share
June 2, 2009 12:20:44 AM

CASE: NZXT TEMPEST
PSU: CORSAIR HX-1000
CPU: INTEL Q6600 GO 3.0ghz lapped
GPU: EVGA GTX 285 x2 SLI
MB: EVGA 750i SLI FTW
RAM: CORSAIR XMS2 8gb 800mhz
HS: MASSCOOL 8WA741 lapped
HDD: SEAGATE 500gb
OS: Windows VISTA ULTIMATE x64

CPU Temps: 40-42 idle / 60-68 load
GPU Temps: 40 idle / 60 load

When you use prime 95 what test do you run? my temps are different on each test. Whats the most important and are my temps going to fly? I dont use my PC for more than 4 hours a day. My temps are as follows at 3.0ghz with no voltage change what so ever.I can reach 3.2 with no voltage chage also but temps are too high. N E ways

Blend: 42 idle - 54 load
large FFT's: 42 idle - 58 load
small FFT's: 42 idle - 67 load

also how many number of torture test threads is everyone running? Im running 4 on everyone of these tests

More about : prime temps

a c 172 à CPUs
a c 197 K Overclocking
June 2, 2009 12:27:22 AM

If you want to fully load the CPU, run 4 instances of small fft's. That should be apparent from your results.

Your CPU temps seem pretty high for only running at 3.0 GHz. My Q6600, Oc'd to 3.6 GHz with a TRUE/Scythe S-Flex SFF21-F in an Antec 900 case, with P95 running small fft's gives Cpu core temps of 61 C - 65 C.
June 2, 2009 1:35:37 AM

jsc said:
If you want to fully load the CPU, run 4 instances of small fft's. That should be apparent from your results.

Your CPU temps seem pretty high for only running at 3.0 GHz. My Q6600, Oc'd to 3.6 GHz with a TRUE/Scythe S-Flex SFF21-F in an Antec 900 case, with P95 running small fft's gives Cpu core temps of 61 C - 65 C.



prime 95 maxes out the cpu though right? so if i play games like crysis,crysis warhead and far cry 2 I should be ok...??? maybe at least until i can get a different heatsink that is. please correct me if im wrong.
Related resources
June 2, 2009 4:34:22 AM

^ Yes it does, but I hear that Prime95 doesn't have support for multi core CPU's (Dual Core, Quad Core) so you need to run 4 instances (since you have a Q6600) And a program called OCCT does have support for multi core CPU's. You can try that.
June 2, 2009 6:41:08 AM

^ Thanks for clearing that, I use Prime95 the most, but occasionally use OCCT.
June 2, 2009 11:26:48 AM

but again, mostly anything I do like playing games will never stress my core like prime does, right? So even though I close to 68 degrees with primes small ftt's I should be able to play game with no worries.....I hope. the problem is that I cant get a new heatsink for 2 weeks maybe more and I just want to know I will be o.k. for a while. played crysis warehead maxed out 4xaa and core only went to 47 degrees. ok or not?
a b à CPUs
a b K Overclocking
June 2, 2009 2:47:41 PM

jdigital1983,

The information you want is covered in my Guide under Section 6, which I had hoped you would read for yourself. The Stickies at the top of these Forums are provided for everyone's benefit, so that we don't have to repeatedly post the same answers. Nevertheless, I'll put this into perspective for you.

Core 2 Quad's have a single CPU temperature sensor (Tcase), and four Core temperature sensors (Tjunction). For the Q6600 G0, Intel's Thermal Specification is 71c, which is shown in their Processor Spec Finder - http://processorfinder.intel.com/details.aspx?sSpec=SLA...

71c is maximum CPU temperature (Tcase Max), NOT Core temperature, which is a common misconception among many users. Also, there's a 5c Gradient between CPU temperature (lower) and Core temperature (higher), which is shown in the following Intel document - http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

Although maximum Core temperature (Tjunction Max) is 100c, which is for Throttle and Shutdown protection, the corresponding CPU temperature would be 95c; far too hot for sane operation. As such, CPU temperatures above 71c and Core temperatures above 76c should be considered as an "overtemp" condition.

SpeedFan shows CPU temperature AND Core temperature, while Real Temp shows Core temperatures ONLY. Both programs allow calibrations. Intel has stated that the Digital Thermal Sensors (DTS) used for monitoring Core temperatures are accurate at very high temperatures, become less accurate as temperatures decrease, and may be unreliable at idle temperatures.

On the other hand, the Analog Thermal Diode used for monitoring CPU temperature is linear from idle thru high temperatures, which is why, as per agreement with Intel, motherboard manufacturers do not include Core temperature in their monitoring utilities found on the installation disk, such as Asus Probe.

Never assume that default temperatures are accurate. The accuracy of CPU temperature is determined by BIOS calibrations, but can be closely calibrated in SpeedFan. The accuracy of Core temperatures are determined by Intel factory calibrations, but can be closely calibrated in SpeedFan and Real Temp.

Prime95 Blend, or OCCT (Linpack), or CPU Burn Test (LinX) are cyclic workloads, which produce fluctuating temperatures. While these are useful for stability testing, they're inappropriate for thermal testing. Prime95 Small FFT's is the standard for thermal testing, because it's a steady-state 100% workload. Since thermal saturation is reached within 7 to 8 minutes, a 10 minute test is adequate. Keep in mind that even the most processor intensive games or applications will rarely exceed 70% to 85% sustained workload.

From the Guide:


Scale 3: Quad
Q9x50: Tcase Max 71c, Stepping E0, TDP 95W, Idle 16W
Q9x50: Tcase Max 71c, Stepping C1, TDP 95W, Idle 16W
Q9400: Tcase Max 71c, Stepping R0, TDP 95W, Idle 16W
Q9300: Tcase Max 71c, Stepping M1, TDP 95W, Idle 16W
Q8x00: Tcase Max 71c, Stepping R0, TDP 95W, Idle 16W
Q8200: Tcase Max 71c, Stepping M1, TDP 95W, Idle 16W
Q6x00: Tcase Max 71c, Stepping G0, TDP 95W, Idle 16W<--Q6600 GO

-Tcase/Tjunction-
--70--/--75--75--75--75-- Hot
--65--/--70--70--70--70-- Warm
--60--/--65--65--65--65--Safe <--
--25--/--30--30--30--30-- Cool


CPU temperature = Tcase
Core temperature = Tjunction

Clear enough?

Comp :sol: 
June 2, 2009 3:49:08 PM

CompuTronix said:
jdigital1983,

The information you want is covered in my Guide under Section 6, which I had hoped you would read for yourself. The Stickies at the top of these Forums are provided for everyone's benefit, so that we don't have to repeatedly post the same answers. Nevertheless, I'll put this into perspective for you.

Core 2 Quad's have a single CPU temperature sensor (Tcase), and four Core temperature sensors (Tjunction). For the Q6600 G0, Intel's Thermal Specification is 71c, which is shown in their Processor Spec Finder - http://processorfinder.intel.com/details.aspx?sSpec=SLA...

71c is maximum CPU temperature (Tcase Max), NOT Core temperature, which is a common misconception among many users. Also, there's a 5c Gradient between CPU temperature (lower) and Core temperature (higher), which is shown in the following Intel document - http://arxiv.org/ftp/arxiv/papers/0709/0709.1861.pdf

Although maximum Core temperature (Tjunction Max) is 100c, which is for Throttle and Shutdown protection, the corresponding CPU temperature would be 95c; far too hot for sane operation. As such, CPU temperatures above 71c and Core temperatures above 76c should be considered as an "overtemp" condition.

SpeedFan shows CPU temperature AND Core temperature, while Real Temp shows Core temperatures ONLY. Both programs allow calibrations. Intel has stated that the Digital Thermal Sensors (DTS) used for monitoring Core temperatures are accurate at very high temperatures, become less accurate as temperatures decrease, and may be unreliable at idle temperatures.

On the other hand, the Analog Thermal Diode used for monitoring CPU temperature is linear from idle thru high temperatures, which is why, as per agreement with Intel, motherboard manufacturers do not include Core temperature in their monitoring utilities found on the installation disk, such as Asus Probe.

Never assume that default temperatures are accurate. The accuracy of CPU temperature is determined by BIOS calibrations, but can be closely calibrated in SpeedFan. The accuracy of Core temperatures are determined by Intel factory calibrations, but can be closely calibrated in SpeedFan and Real Temp.

Prime95 Blend, or OCCT (Linpack), or CPU Burn Test (LinX) are cyclic workloads, which produce fluctuating temperatures. While these are useful for stability testing, they're inappropriate for thermal testing. Prime95 Small FFT's is the standard for thermal testing, because it's a steady-state 100% workload. Since thermal saturation is reached within 7 to 8 minutes, a 10 minute test is adequate. Keep in mind that even the most processor intensive games or applications will rarely exceed 70% to 85% sustained workload.

From the Guide:


Scale 3: Quad
Q9x50: Tcase Max 71c, Stepping E0, TDP 95W, Idle 16W
Q9x50: Tcase Max 71c, Stepping C1, TDP 95W, Idle 16W
Q9400: Tcase Max 71c, Stepping R0, TDP 95W, Idle 16W
Q9300: Tcase Max 71c, Stepping M1, TDP 95W, Idle 16W
Q8x00: Tcase Max 71c, Stepping R0, TDP 95W, Idle 16W
Q8200: Tcase Max 71c, Stepping M1, TDP 95W, Idle 16W
Q6x00: Tcase Max 71c, Stepping G0, TDP 95W, Idle 16W<--Q6600 GO

-Tcase/Tjunction-
--70--/--75--75--75--75-- Hot
--65--/--70--70--70--70-- Warm
--60--/--65--65--65--65--Safe <--
--25--/--30--30--30--30-- Cool


CPU temperature = Tcase
Core temperature = Tjunction

Clear enough?

Comp :sol: 





YES clean enough. you could of just said yes I will be o.k. thank you for your help ( great guide by the way)
a b à CPUs
a b K Overclocking
June 2, 2009 3:55:23 PM

I don't like to just throw quick and dirty one-liner answers at users, because it leaves a great deal open to interpretation. As an engineer, my work requires me to be very specific, which is why I take the time and care to write thorough explanations.

Overclocking and temperatures are all about understanding Intel's voltage and thermal specifications, and not exceeding them. Since I've made it my mission here at Tom's to help users understand temperatures, and as other users read my posts, I have no desire to write in an ambiguous fashion.
June 2, 2009 4:18:59 PM

CompuTronix said:
I don't like to just throw quick and dirty one-liner answers at users, because it leaves a great deal open to interpretation. As an engineer, my work requires me to be very specific, which is why I take the time and care to write thorough explanations.

Overclocking and temperatures are all about understanding Intel's voltage and thermal specifications, and not exceeding them. Since I've made it my mission here at Tom's to help users understand temperatures, and as other users read my posts, I have no desire to write with an ambiguous style.





well thank you i appreciate it. with that said then what will lower my temps and make my system run better. all I have done is increase fsb to 333x9 voltage for the cpu atomaticly changed to 1.4 and ram is unliked still at 800mhz. would i be better of with a lower multiplier with the same speed maybe 400x8 so my ram is in sync & most of all lowering my multiplier will it reduce my temps at all?
a b à CPUs
a b K Overclocking
June 2, 2009 4:39:41 PM

Temperatures are influenced primarily by Vcore, CPU cooler, ambient temperature and clock speed. 3.2 Ghz at either 356 x 9 or 400 x 8 will have no effect on temperatures, however, 400 x 8 is more desirable for the 1:1 ratio. Also, when overclocking, never run Vcore in "Auto" as it will typically apply more voltage than necessary to maintain stability, which in turn increases temperatures.
June 2, 2009 8:40:44 PM

thank you for all your help computronix i will try it
June 2, 2009 8:44:11 PM

CompuTronix said:
Temperatures are influenced primarily by Vcore, CPU cooler, ambient temperature and clock speed. 3.2 Ghz at either 333 x 9 or 400 x 8 will have no effect on temperatures, however, 400 x 8 is more desirable for the 1:1 ratio. Also, when overclocking, never run Vcore in "Auto" as it will typically apply more voltage than necessary to maintain stability, which in turn increases temperatures.


Don't you get tired of saying the same ole thing...... ;) 
June 2, 2009 9:30:36 PM

what about cpu fsb voltage?
leave it on auto or change it also its currently in the red
a b à CPUs
a b K Overclocking
June 3, 2009 12:25:47 AM

ac3144 ,

Thanks for asking. Yes, I do get damned tired of repeating myself. I'm always willing to help those who are willing to help themselves, however, I won't answer endless questions for those who won't read, or take responsibility for conducting their own research.

jdigital1983,

FSB at 400 Mhz should require no more than 1.25 volts. Have you tried reading the Overclocking Guide at the top of this Forum?

Comp :sol: 
June 3, 2009 2:04:31 AM

we all dont know what you know and rely on our own research and information from ppl like you for help. its a forum. thats why im here, to learn and gather information.
a b à CPUs
a b K Overclocking
June 3, 2009 6:24:29 AM

jdigital1983,

In all fairness, I've answered your questions, but you haven't answered mine;

Have you tried reading the Overclocking Guide at the top of this Forum? said:
Have you tried reading the Overclocking Guide at the top of this Forum?
June 3, 2009 3:54:39 PM

yes i have thank you. so thing are questionable though. what one person precieves is not what the other does (we all have questions and sometime don't get things) but thank you for being patient and helping the best you can. (sometime a yes or no is better than a long explaination.
a b à CPUs
a b K Overclocking
June 3, 2009 9:00:40 PM

jdigital1983,

Understood, however, my "long explanations" contain addition valuable information from which you, as well as other readers can learn, while a simple yes or no does not. I believe that more information is better than less, because the essence of the topic is in the details. The individual reader can then decide what to retain, and what to discard.

Additionally, you might note that while graysky's Overclocking Guide is an excellent piece of work, and his topic and mine overlap considerably, he tends to be a bit conservative regarding overclocking. Also, he advises using Large FFT's for thermal testing, which isn't quite right.

  • Prime95's definition of Large FFT's: (maximum heat, power consumption, some RAM tested). While this sounds like the logical choice for processor thermal testing, it's actually not. Maximum heat and power consumption is generally true, however, it refers to "overall" heat and power spread across the processor and memory, which does NOT yield a steady-state 100% workload across all segments of the test. This can be visualized on the "Charts" in SpeedFan.

  • Prime95's definition of Small FFT's: (maximum FPU stress, data fits in L2 cache, RAM not tested much). The key statement here is "maximum FPU stress" which concentrates on the processor, and does yield a steady-state 100% workload. This can also be visualized on the "Charts" in SpeedFan.

    I hope that you find this additional explanation helpful.

    Comp :sol: 
    June 4, 2009 12:08:46 AM

    i do very much thank you and i understand we must help others to, so indepth explainastion are propaly the best.

    I have a question 4 you though relating memory and i will post this in a memory thread to.

    with all said and done with my overclock.
    my memory is corsair twin2x4096-6400c5 (6400) 800mhz 1.9 volts on website but 1.8 on chip its self with 5-5-5-18 timings.

    question is will i see more performance as is or lowing the timings?

    overclock is 1333 fsb x 9 multiplier 1.325v and small bump on other volts.
    a b à CPUs
    a b K Overclocking
    June 4, 2009 1:43:06 AM

    If your memory is currently at 1:1, then it's running underclocked at DDR2 667, in which case you can tighten the timings to 4-4-4-15.

    If your memory is currently at 4:5, then it's running overclocked at DDR2 833, in which case you have no room to tighten timings.

    This is why 5-5-5-18 memory is undesirable for overclocking, since 4-4-4-12 memory leaves headroom options for increasing frequency or tightening timing. Regardless, understand that the difference in memory performance between 4-4-4-15 and 5-5-5-18 is less than 2%, which is insignificant in terms of overall system performance.

    Many enthusiasts continue to get stuck on small differences in memory performance, believing that it makes a huge difference in overall system performance, when in fact it has very little effect, and is imperceptible except to benchmarking utilities.
    June 4, 2009 2:05:41 AM

    well that takes care of that. you really are a big help thanks
    June 4, 2009 2:09:48 AM

    one more question (i know you must hate me by now) full load will prime max temps 60 cpu and 55 cores. in cpu-z core voltage will rarely go to 1.28. does this mean I could possibly lower my vcore volts to 1.28ish insted of 1.325 to reduce heat and maybe achieve a 3.2 overclock. (my HS really sucks)
    a b à CPUs
    a b K Overclocking
    June 4, 2009 3:13:22 AM

    jdigital1983,

    Yes, there's always "one more question". Considering the many threads you've started, you're rapidly approaching my 20 question limit, and I see that our Moderator, randomizer has already locked one of your threads, however well intended.

    If I may respectfully suggest, consider focusing more on research through reading. This will help you to become better informed, and ask fewer, but better questions. Most overclocking enthuasts appreciate the learning curve, and respect this approach, which allows for discussions of more interesting problems.

    With this in mind, I offer the following:

    (1) Relative to your Core temperature, your CPU temperature is wrong. CPU temperature should be 5c lower than Core temperature, not 5c higher.

    (2) Your EVGA motherboard shows excessive Vdroop (Google it). If BIOS has "Line Load Calibration" then enable it.

    (3) You can not typically decrease Vcore AND increase clock frequency AND maintain stability. As you increase clock, you must increase Vcore to maintain stability, which in turn, increases temperatures.

    (4) And yes, by all means, push your overclock to 3.2 @ 400 x 8, or better yet, get a high-end cooler such as the Xigimatek HDT-S1283, then go for 3.6 @ 400 x 9. Either setting will allow your memory to run 1:1 at it's intended stock frequency of DDR2 800.

    Comp :sol: 
    !