Please report all spam threads, posts and suspicious members. We receive spam notifications and will take immediate action!
Page 3 of 5 FirstFirst 12345 LastLast
Results 21 to 30 of 45

Thread: GA-EP45C-UD3R & E8500 OC help




  1. #21
    Chike is offline Senior Member
    Join Date
    Jun 2009
    Location
    Israel
    Posts
    6,531

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Quote Originally Posted by Cell247 View Post
    I ran Blend for passed 8k and no errors. Lowest temps 56c and highest 63c. =\ I guess thats pretty warm?
    I get the same with 3.6GHz, what is the ambiant temperature?

    EDIT: by lowest you mean Idle?

  2. #22
    Join Date
    Sep 2009
    Posts
    18

    Default Re: GA-EP45C-UD3R & E8500 OC help

    No, during Blend. Idle is 39c. Well I live in Texas and so its pretty hot. =) My office AC is ****ty so I would say its about 80f in my office. I moved from 444 to 445 and running at 4004.9 with no other settings changed. Im going to run a few passes of Blend and see how it heats up. Seems to be staying around 58-60c. What should I work on next? Oh and so far all I use is CPU-Z, Real Temp and Prime95. Should I be using other appz also?
    Last edited by Cell247; 09-03-2009 at 08:23 AM.

  3. #23
    Chike is offline Senior Member
    Join Date
    Jun 2009
    Location
    Israel
    Posts
    6,531

    Default Re: GA-EP45C-UD3R & E8500 OC help

    The Linpack based stress programs LinX/IBT(IntelBurnTest) may indicate if the overclock actually achive it's expected results. Aside from errors, low results may indicate need for more vcore.
    Note that these programs produce much higher temps, 10c+ over Prime95.
    When you satisfied with CPU overclock you can use memory benchmarks to set memory for best performance, and of cours any games/software you usually use. Even after all the stress tests other software may cause crashes.

  4. #24
    Join Date
    Sep 2009
    Posts
    18

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Is it safe for me to run 10c more? Im on the 10k test now in Prime95 and highest heat is 63c. If I go up to 73c with Intel Burn is that safe? What temp and voltage is safe for every day use? When does a C2D start to degrade from OCing? I thought for sure 4.2 was obtainable with my hsf and some low voltage. I must have read wrong.

  5. #25
    Chike is offline Senior Member
    Join Date
    Jun 2009
    Location
    Israel
    Posts
    6,531

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Core temps of 73c is not much to worry about, especially if it's only for a short period of time. No need to tun IBT/LinX for more than 10-20 loops with max mem.
    Some say up(down) to 20c distance from TJmax (80c for E8500) is safe, other will say even more. I myself rather keep them under 70c for everyday use.

  6. #26
    Join Date
    Sep 2009
    Posts
    18

    Default Re: GA-EP45C-UD3R & E8500 OC help

    I just ran Intel Burn on standard with high of 75c.



    Im now running it on High and still hasnt broke 75c. How should I go about figuring out the best memory timings and settings? Again thanks for the help.

    EDIT: Just hit 76c right before it finished 5 runs on High. But that was with 2gigs of ram setting. So run custom with 3292, because thats all it sees, and run 20 loops?
    Last edited by Cell247; 09-03-2009 at 09:24 AM.

  7. #27
    Chike is offline Senior Member
    Join Date
    Jun 2009
    Location
    Israel
    Posts
    6,531

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Nice GFlops.
    As I said I myself don't bother much with memory settings, just set the lowest tRD (Static tRead Value) and let BIOS set whatever is needed.
    You can run memory benchmarks, like the one in Everest or MaxxMEM2 preview.
    Post a screen capture of MemSet main and spd(click spd) windows we'll see what can be done.
    What memory multipliers you have for straps B & D?

  8. #28
    Join Date
    Apr 2009
    Location
    J'Habite En Angleterre
    Posts
    2,431

    Default Re: GA-EP45C-UD3R & E8500 OC help

    As Chike says, IBT/Linpak is a useful program.

    You will see temps as much as 10-15c higher than with Prime95 if Vcore is set correctly. The calculation speed is shown in GFlops and for 4GHz should be around 28-30 ish. If it's much lower or the CPU heats up to only the same as Prime95 levels, then increase Vcore by one notch at a time until you see an increase to the right speed or a temperature more than Prime testing gives. this change will be a sudden one, eg at 1.3V you may have cool temps and low GFlop values and increasing by one notch will see temps of 10c more and/or a significantly higher calculation rate.

    As for degradation, some say they've experienced it at relatively modest speeds and voltages, others at extreme levels and some (usually the majority) have no issues at all. Because degradation is hard to measure and everyone reporting degradation has had it possibly happen with different variables (Vcore, temps components used etc) it's hard to talk about in definite terms.

    70c with IBT is definitely nothing to worry about at all. It's such an extreme test that it's not a representation of the temps that will happen in real world usage. I find Prime95 reflects better the kind of temps I have during the main use of my rig which is gaming. Also with gaming, as with most real apps, the CPU won't be loaded to 100% constantly, meaning in my gaming trials, I hit around 55-57c only for very brief periods. The exception to the 100% load thing is in programs for video encoding/transcoding, compression, encryption/decryption and if 3D modelling then while rendering complex scenes that take a while.

    it would be a good idea to leave Prime running for a minimum of 8 hours, preferably (my own personal tastes) 12 hours +. Errors can occur at any time. The chances of stability increase with each passing minute of priming and to be honest anything that passes 8 hours is pretty much solid for every day use. I have had errors happen at about the 7 1/2 hour mark before and also around the 22 hour mark. In both cases I tweaked my settings to improve stability, but if you go 23 hours you can probably leave it alone really.

    Note that if you take part in distributed computing projects such as Folding@Home (protein folding to search for disease cures) then I would want a minimum (again personal preference) of 48 hours stable. I couldn't forgive myself if I supplied faulty data to such a project. Most of these projects have error checking and redundancy in the form of giving multiple PC's the same set to work and comparing, but even so, it's important in this case to be as stable as possible. Stability is relative though and a 48 hour stable PC could have failed at 48 hours and 1 second, you never know.

    For tuning RAM, two of my favorite programs are SuperPi Mod and Everest Ultimate's RAM benchmark. Super Pi calculates the mathematical constant of Pi to however many decimal places you specify. Lower times mean a faster system. You can expect a time of 11-12 seconds for a calculation to 1 million decimal places, maybe slightly less with super fast and super tight RAM settings. I can hit just over 13 seconds with my 3.9GHz and DDR-1093 RAM at 5-5-5-15-3-3-6-54-6-2T, which is a slight improvement over the auto and XMP settings. The manual settings give me about 200MB/s more bandwidth in Everest and a gain of over 0.1 seconds for a 1M super Pi. Using both these programs gives a synthetic result (Everest) and a much more real world application (SuperPi)

    Although temps may be good, when you've verified stability and tweaked your RAM to where you're happy, it's then time to see if you can reduce Vcore and MCH Core voltages. You may also be able to use a standard PLL of 1.5 and even reduce RAM voltage a little bit. All this will make the CPU run cooler and also hopefully lower case temperatures a little while using a bit less electricity. I like having a finely tuned rig with everything at the best it can possibly be. This may not be to your tastes though, and if so it seems that you don't really need to tweak much if you don't want.

    After you verify relative stability with 8 hours+ of Prime, post up your SuperPi 1M time and a screenshot of Everest Ultimate's RAM benchmark. there's a free trial of Everest availible, or if you're lucky you can catch a copy as it "falls off the back of a truck" so to speak.

    Here's an example of the two benchmarks on my current config. IBT isn't shown but it calculates at 25GFlops. Sorry for the long post by the way, but there's a lot of detail to consider, even after you get the acceptibly stable clock speed you want.
    Coolermaster CM 690 II advance Case
    Corsair HX750 (CWT, 91%(80+ Gold rated @230V) single 62A 12V rail
    P55A-UD4 v2.0 @ F14
    Core i5 760 @ 20 x 201, 4.02GHz
    TRUE Black with a single Noctua NF-P12 pumping out 55 CFM @ 19db .
    2 x 2GB Mushkin Ridgeback (996902), @ 7-10-8-27, 2010-DDR, 1.66v
    2 x Gigabyte GTX 460 1024MB in SLI (Pre OC'd to 715MHz core and 1800MHz VRAM) @ 850 Core / 4100 Mem.
    Intel X25-M Boot Drive (OS and Programs) 200MB/s Read & 90MB/s Write
    Corsair X32 200MB/s Read & 100MB/s Write
    WD Caviar Blue 640GB C (Steam, Games, Storage, Temp Files & Folders, etc)
    Samsung F3 500GB Backup/Images
    Noctua 1300RPM 19dB case fan (rear extraction)
    3 x 140 MM Coolermaster LED fans (one front intake, one top extraction, one side intake)
    Dell Ultra Sharp 2209WAf E-IPS @ 1680x1050

  9. #29
    Join Date
    Sep 2009
    Posts
    18

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Chike:

    Here are my B & D multipliers:

    2.40B
    3.20B
    4.00B
    2.00D
    2.66D
    3.33D
    4.00D

    Here is the memset shot you asked for:



    Psycho101:

    It may take a few mins to absorb your post.

    The calculation speed is shown in GFlops and for 4GHz should be around 28-30 ish.

    Mine was a little less on 10 runs at max memory. Its pretty close to 28 though. Is this fine?



    I like having a finely tuned rig with everything at the best it can possibly be.


    I would love to get this new box fine tuned. Its my first build in a few years and I want to put some love into it. Ill get those other apps and post back.

  10. #30
    Join Date
    Apr 2009
    Location
    J'Habite En Angleterre
    Posts
    2,431

    Default Re: GA-EP45C-UD3R & E8500 OC help

    Your IBT figures are bang on the money. You may be able to improve them with som RAM tuning. Trying to reduce Static tRead one level at a time (lower is better for tRead) is the most effective single value after CAS latency with regards to speeding things up.

    After having a look at SuperPi and Everest and posting your figures, the very next thing I would do is give Prime95 it's 8-12 hours on Blend. It may well be that some slight tuning of CPU Termination, MCH Core and MCh Reference are needed, but hopefully not. You could use Orthos Prime in place of Prime95 if you wish, as I do. The only real benefit of orthos is that when an error on one core is found, Prime carries on runing with the other core (pointless) where as Orthos stops both threads. this is great for unattended testing. I test overnight while I sleep, so I like knowing that the whole process stops when it's pointless running any more. it saves electricity if nothing else.

    Edit: If you're stable at 8-8-8-24 on your RAm that's great, but remember any errors could be because the RAM may need to be run at 9-9-9-26. Better to have looser timings and a stable system than tight timings on an unstable one. I'd say that you should be fine at 8-8-8-24 though seeing as you're not running all that much above 1066 and the RAM can do 7-7-7-20 at that speed. tweaking later may let you run something like 8-8-7-20 or maybe even 8-7-7-20.
    Last edited by Psycho101; 09-03-2009 at 10:42 AM.
    Coolermaster CM 690 II advance Case
    Corsair HX750 (CWT, 91%(80+ Gold rated @230V) single 62A 12V rail
    P55A-UD4 v2.0 @ F14
    Core i5 760 @ 20 x 201, 4.02GHz
    TRUE Black with a single Noctua NF-P12 pumping out 55 CFM @ 19db .
    2 x 2GB Mushkin Ridgeback (996902), @ 7-10-8-27, 2010-DDR, 1.66v
    2 x Gigabyte GTX 460 1024MB in SLI (Pre OC'd to 715MHz core and 1800MHz VRAM) @ 850 Core / 4100 Mem.
    Intel X25-M Boot Drive (OS and Programs) 200MB/s Read & 90MB/s Write
    Corsair X32 200MB/s Read & 100MB/s Write
    WD Caviar Blue 640GB C (Steam, Games, Storage, Temp Files & Folders, etc)
    Samsung F3 500GB Backup/Images
    Noctua 1300RPM 19dB case fan (rear extraction)
    3 x 140 MM Coolermaster LED fans (one front intake, one top extraction, one side intake)
    Dell Ultra Sharp 2209WAf E-IPS @ 1680x1050

Page 3 of 5 FirstFirst 12345 LastLast

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •