• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

e6750 overclock

apresutt

n00b
Joined
Dec 26, 2007
Messages
47
At the moment i have my cpu overclocked to 3.0 GHz, Running Prime95 i hit a steady temperature of 45 C. Is this acceptable? how high can the temperature go? Think i can hit 3.2 without a problem?

Thanks for these answers,
Aaron
 
First of all, I assume you are loading BOTH cores fully, using 2 instances of Prime?

45c isn't anything to worry about. Get CoreTemp and then go into options and set the checkbox to show "remaining delta to tjmax" this will show you, accurately, how warm your cpu REALLY is. If you are 20c or more away from tjmax, man, you are golden, tjmax is the temp at which your cpu will start to throttle. For the record, my E6750 at 3200mhz loads fully with a DELTA to tjmax still, of ~45c, well within safe limits, in fact, that's good for air cooling.

3.2ghz with that cpu should be a GIVEN....just make sure when overclocking your ram is within limits, or use a lower ram divider to keep the speed down.

Many of those cpu's get far more than 3200mhz....
 
I'm only opening one instance of Prime, but in my task manager both cores are stressed to 100.

I'm gonna bump it up to 3.2 and see what i get. Thanks
 
First of all, I assume you are loading BOTH cores fully, using 2 instances of Prime?

45c isn't anything to worry about. Get CoreTemp and then go into options and set the checkbox to show "remaining delta to tjmax" this will show you, accurately, how warm your cpu REALLY is. If you are 20c or more away from tjmax, man, you are golden, tjmax is the temp at which your cpu will start to throttle. For the record, my E6750 at 3200mhz loads fully with a DELTA to tjmax still, of ~45c, well within safe limits, in fact, that's good for air cooling.

3.2ghz with that cpu should be a GIVEN....just make sure when overclocking your ram is within limits, or use a lower ram divider to keep the speed down.

Many of those cpu's get far more than 3200mhz....

I'm at 375x8 (3ghz). I'm pushing to 385x8 today for 3.06. I'm at stock cooling so i won't be able to push it too far.

You should be able to go far higher. Use ORTHOS, not prime, for stress testing for both cores. Otherwise ya, prime95x2. Run smallFFT to stress the o/c.
 
See my sig for my set-up. I have run my E6750 at 444x8=3.6GHz, 1.35v, 53C load for three months with absolutely no issues. I have not tried to push it higher, as this is a gaming rig and I want it stable.
 
My E6750 is also a 3.6 with a Thermalright Ultra 120 eXtreme. Saying what your temp readings are is pretty useless unless you indicate what program you are getting this information from as temp readings on current intel CPU's are very erratic and can be quite inaccurate from program to program.
 
silentcoercion: 3.5 stock? What mobo are you running on? That's an impressive overclock! I'm not sure i can get past 400 just based on the fact i'd have to also overclock my ocz plat rev 2's (ddr2-800)

Rob-J - are you on stock cooling? what mobo are you on?

CoreTemp is the authority for measuring temps on the conroes.
 
I helped a friend OC an e6750 on a P5K Premium. We had no problems hitting 3.8ghz, 12 hour Orthos stable @ 1.44v running 55c under H20.
 
Just got the cpu and board (Abit IP35-E) this week. I am at 3.8Ghz (475x8) so far stable. Running 2x1 GB Crucial Ballistix at 950Mhz 5-5-5-12 T2. Had to up the vcore to 1.3+ something. I'm thinking of picking up 4GB (2x2GB) of 1066mhz ram so I won't be RAM limited. This is on a Thermalright 120.
 
i'm at 3.4 with a thermalright ultima 90, and i haven't had any troubles with temps
 
most e6750s should be able to hit 3.2 without breaking a sweat. on stock cooling and voltage to boot.
 
My 6750 runs 450 x 9 1.41v on stock air, Provided your mobo (and the bios you select) Will take yours that high your chip could possibly run any where from 3.4 to 3.8ghz on decent air.
 
before anyone directs me to read the sticky how-to... i already have. multiple times. Please don't point me there.

I'm new to overclocking and this is my first desktop (had a dell laptop before.) I got the E6750 cause i read all these reviews about getting to 3.3GHz with no problems at all (and even higher up to 3.6!!!). My motherboard is a Gigabyte P35-DS3L and i have G-skill DDR2-800. My issue is that i can't OC over 3.0GHz. completely frustrating. I can boot up at 3.3 but prime 95 fails immediately. I went through the newbie's OC-ing guide multiple times and i have all the power-saving settings disabled. My multiplier is 8x413 with a 1:1 FSB:RAM ratio so the ram is at 826MHz. I have my core voltage at the "normal" 1.35. I've tried increasing the DDR2 voltage +0.1 to 1.9v and that didn't help. (g-skill specs are 1.8-1.9). I have also tried all 3 performance enhance settings (standard, turbo, and extreme(what do these even do? the board manual isn't helpful at all)) all without any luck. My RAM timings are stock/loose at 5-5-5-15. worst of all is that if i set my core voltage to "Auto", CPU-Z reads it as 1.408!! adjusting back to what BIOS would say (assuming its linear and using the "normal" vcore(1.35) at 2.66GHz speed (reads 1.312 in CPU-Z and 1.35 in BIOS) it translates to a 1.4488V! Its completely disheartening that some people are running 3.5GHz on LESS than "normal" core voltage and i can't get mine past 3.0GHz without a huge bump! Is there anyone that can point me in the right direction as to what is going on? The only voltage changes I have made are to the core voltage. I did up my RAM to 1.9 with no luck either. I've also tried changing the MCH(?) latency (in the advanced memory timings) to 8 and higher for more stability but to no avail. I had heard that this would help for FSB's over 400MHz. I'm a little nervous to +.1 my NB voltage as that HS is pretty darn toasty as is at stock. I'm thinking about a little fan for it soon.

anyway....
Again its an Intel E6750(2.66GHz, VID=1.35) in a Gigabyte P35-DS3L (BIOS version F7) along with 2x 2GB G-Skill DDR2-800. PSU is a OCZ Game Xstream 700W and an EVGA 8800GTS(320MB). Also i'm running XP Pro (32-Bit) with SP2. ( I had XP Pro x64 and it wasn't OC-ing right so I though maybe it had something to do with windows... guess not.)

what the heck is going on?

I appreciate any help you guys can give me. Seems like a lot of you know what you're doin.
Thanks a lot!

-Mike
 
Mike:
it doesn't sound like you're doing anything wrong.
Just wanted to confirm that you've increased mch, fsb, pci-e etc all to +0.1? If you're worried about heat, add another fan to the NB.

what are your temps? It might be getting too hot, but i don't know - run coretemp v0.97.

what was your last stable overclock? Do a memtest on that, and also a memtest for each increase you make (running 10 rounds of them takes maybe 10 minutes.).

You've checked for the latest bios?
 
crunch,

i have NOT increased mch,fsb,pci-e and etc all to +0.1. I'm afraid of heat. I also had someone on another forum tell me his board fried after +0.2... makes me nervous.

My temps are fine. using core temp 0.97 my idles are 35/37 and never been above 60 running Prime 95.

My last stable OC was 3.00GHz (8x375) with normal (1.35) core voltage. It ran Prime 95 for 7.5 hours with no errors. Memory was at DDR2-900 speed (2.4 ratio, 100MHz over rated 800MHz) 5-5-5-15 with the "performance" setting on normal. All voltages at normal/stock. I can run a memtest and let you know how that goes.

As far as the BIOS version, i have F7. I believe there are beta versions available for F8, but i figured i should stay away from a beta when trying to OC so i know it not a BIOS issue. Is that good to assume or should i update? (as gigabyte states, apparently F3 is supposed to offer better OC stability. should i downgrade to that one?)


Another question... I was just playing Crysis last night (settings all set to high, 1680x1050, no AA) and there would be times in the game that I'd see artifacts(?). Like, one particular situation i was looking at a parked Hummer. Everything looked fine. But when i used the sight to zoom in on the truck though, certain parts (other peripheral sight places still looked fine) just turned kinda boxy and green/pink or purple. I got nervous and closed the game. Another incident was when i turned on my flashlight attachment, i saw pink lines extending from the light... not an actual white light beam. After that I left the computer for a bit, came back, and my display (entire LCD) was pink with blue/green static. I just pressed the reset button and its been fine since. I read that you should lock the PCI-E freq. to 100MHz. How important is this? Cause there was ONE time I reloaded stock BIOS settings after it failed to POST, and i didn't get it set to 100 again, it was on auto when i ran the OC. did i ruin my video card?!?! its been less than a week since i ordered it!

Thanks!
-Mike
 
I ran a memtest86 using a cd iso and booting from it. I let it run all day while i was at class. when i came back, it was still running. i dunno if it was still going or had started the test again. regardless, i see it has some errors. I don't really know what the results mean but here's a picture. (sorry its so terrible)

the specs of the run are as follows:
3.0GHz (8x375)
2.4 ratio (900MHz, 100 over stock 800MHz)
Auto timings (5-5-5-15)
core voltage = 1.35
all other voltages auto/stock
pci-e at 100MHz
and the performance enhance set to default

picture (http://web.ics.purdue.edu/~mjlamb/Pictures/Image3.jpg)

I guess this means i should back down to 800MHz or below?
 
I ran a memtest86 using a cd iso and booting from it. I let it run all day while i was at class. when i came back, it was still running. i dunno if it was still going or had started the test again. regardless, i see it has some errors. I don't really know what the results mean but here's a picture. (sorry its so terrible)

the specs of the run are as follows:
3.0GHz (8x375)
2.4 ratio (900MHz, 100 over stock 800MHz)
Auto timings (5-5-5-15)
core voltage = 1.35
all other voltages auto/stock
pci-e at 100MHz
and the performance enhance set to default

picture (http://web.ics.purdue.edu/~mjlamb/Pictures/Image3.jpg)

I guess this means i should back down to 800MHz or below?

-Keep the ratio at 2.0 - i did that it things were unstable.
-you should not overclock your memory at this time. As you can see from memtest, you're getting tons of errors. Run @ 2.0 and see if you still get errors.

-pci-e @ 100 would be good
 
I re-ran memtest86 using the 2.0 ratio, and no errors this time. so the specs are:

3.0GHz (8x375)
2.0 ratio (750MHz, 50 under stock 800MHz)
Auto timings (5-5-5-14)
core voltage = 1.35
all other voltages auto/stock
pci-e at 100MHz
and the performance enhance set to default

picture


so now that I've effectively ruled out the RAM with my OC settings, whats next? I'm still at a loss and don't know why i can't overclock to 3.3 at normal vcore or slighty above.

another note, crunch, you said to do about 10 passes, it'd take about 10 mins... 10 passes of what? a single pass when i run it takes about 2 hours from what i can tell....
thanks for the help with the RAM too! i'll assume i'll be fine to run it up to 800MHz or lower then.
-Mike
 
Mike,
My bad - i meant ten tests or so, not 10 passes. You can do longer tests later. We just want to find some basic stability at lower overclocks.

Just to understand, you tried to overclock past 3.3 @ spd = 2? I think you should be able to get there. You have good cooling. Try 400x8 from where you are (5mhz at a time!). Run memtest to rule out any memory issues. From there, if you get problems, you'll have to start increasing your voltages for your CPU. You have a good cooler so I think that can be done.

if u didn't read my older posts, i'm 400x8 all stock volts and cooling. Very happy with it.
 
Highest I can get currently is 3.1Ghz. Currently, I think my motherboard is holding me back. It is a P6N Diamond, but I think it is having issues and does have excessive coil-whine... along with some other problems . I've got a new one coming my way and I'll see if I can push it further. I do have TRUE on it and at 3.1Ghz running Orthos, it stabilizes at 48C.
 
well, I just spent a while doing tests.....

basically, im still left scratching my head. There's a lot of notes i took. take a look....
Results

At 3.04GHz on normal vcore, i got a 33% PRIME95 fail rate.
Any speed higher resulted in an almost immediate fail. At 3.2GHz(normal vcore) i couldn't even boot.
I ran memtest86 for a Wall Time of 7 mins each. (i don't know how you figure i can finish 10 tests in ~10 mins. I only got through ~67% of the 4th test after 7 minutes). For that 900MHz RAM test, after 7 mins i had 0 errors, but i let it go to 11 mins just to see cause you said ~10 mins... but it was still had 0 errors. Looking at my first picture of the memtest results, it doesn't look like i got errors until late into the 7th test.

Another note: while performing all of these tests... not during ANY test did my core temperatures exceed 61*C and when Idle was around 34/35*C (using core temp v0.97.1, Tj. Max = 100*C)

Anyway. im still baffled about this whole OC experience. i have YET to read about someone with my same issue using my same CPU and board. Memory is different, yes, but i feel like its not an issue since i don't get any errors in it (at least at 750MHz i didn't. I have yet to test 750-800MHz ranges for more than 7 minutes i used as a standard for the latest results)

could this all be because i have not adjusted my NB voltage and some of those other adjustments? If so, im nervous to do so for fear of frying my board! It just doesn't make sense... Crunch is running ALL stock voltages (all? even NB and FSB voltages and the like?) at 3.2GHz perfectly fine!

Anymore help or ideas would be greatly appreciated. If it comes down to it, maybe i can RMA my RAM for some better stuff?
 
well, I just spent a while doing tests.....

could this all be because i have not adjusted my NB voltage and some of those other adjustments? If so, im nervous to do so for fear of frying my board! It just doesn't make sense... Crunch is running ALL stock voltages (all? even NB and FSB voltages and the like?) at 3.2GHz perfectly fine!

Anymore help or ideas would be greatly appreciated. If it comes down to it, maybe i can RMA my RAM for some better stuff?

NB/FSB et all are all +0.1, including the PCIe. Mind you, I've got the ds3R, not the L, so it runs a little cooler than your board.
 
thanks crunch. i'll keep that in mind. I'm probably going to get a NB fan for that hot HS to help with temps. I guess my only options are to:
1. start increasing NB/FSB and others voltages up along with small amounts of vcore until im stable.
2. be happy with 3.0GHz

Would most of you agree that a +0.1v on the NB (and other components for that matter) is the max if I don't want to ruin anything? i'm not sure of any specified operating voltages on a NB like you can find on CPUs.

And if my G-Skill is rated 1.8-1.9V is it safe to go to 2.0V?

Just another thought... theoretically speaking, a goal in overclocking is to keep vcore low. this 1.keeps temps lower and 2. improves the life of the CPU at higher speeds. So are you more or less negating the effects of a lower vcore by increasing NB/FSB and other voltages in an attempt to keep the vcore lower? if thats the case, maybe its simply better to throw the 1.4v at the CPU and leave the other board components at normal. Just a thought. Ideas on this?

I still think those results from my tests are interesting (provided i did them fairly and systematically, which i believe i did..)
especially how in that 3.04GHz run, some Prime tests failed so quickly, then worked, failed again, and then seemed to be alright...

Thanks a lot for the help so far guys. excellent forum! i like!
-Mike
 
hello all!

i just did some playing around with not a lot of luck.
I haven't fine tuned it, but i was able to do Prime95 for a half our with these specs:

3.2GHz (8x400)
Vcore at 1.45
+.01 on DDR2,MCH, and FSB
and ram running 5-5-5-15 at 800MHz
Max Temp = 71 (kinda made me nervous, seems a little hot no?)

so i know so far that that seems to be a stable point, however i don't like any of my voltages that high and it runs hot.

Later in the afternoon i finished up these test, check it out.

I ran Prime95 for 30 minutes with no errors at:
3.2GHz (4x800)
Vcore = 1.43125
all other voltages normal
1:1 ram running 800MHz at 5-5-5-15
Temps were around upper 60's, maybe reached 70 at a few points.

After this test, i went to a Vcore of 1.425 (just .00625 volts lessthan last stable) with all other settings left the same as the stable run. Prime95 gave an error after just 7 mins!

I then ran again at the 1.425, but upped vMCH to +0.1 (all else left the same) and that time Prime95 gave an error after 20 mins.

I then ran another test at 1.425, left vMCH at +0.1 and proceeded to up FSBv to +0.1 (all other settings left the same) and after that run, again Prime gave an error after 20 mins.

Finally i ran a last test at the 1.425 Vcore, vMCH, FSBv at +0.1 AND upped DDR2v to +0.1 (again leaving all other settings the same). Prime reported errors after about 20 minutes...

I also noted that at a Vcore of 1.425, CPU-Z reads my voltage as 1.392 at idle and under Prime95 load, it drops to 1.3600 which comes out to a Vdroop of 0.032 between the Idle and Prime95 load CPU-Z readings. Doesn't seem like a large amount, not enough to be ruining my overclock right?

another very interesting thing i've notice lately...
i live in an apartment, and i've got a mini-fridge in my room. I have noticed lately that when the fridge compressor turns on to keep cool, the Num/Caps/Scroll lock lights on my keyboard turn off. And other time's i've had my keyboard become unresponsive. Simply restarting fixes it and i can't prove that the fridge turning on causes that issue. But i know for a fact that when the fridge turns on, the lights on my keyboard go out. does this mean that my computer is barely getting the power it needs?!?!?!?! could this be the cause of my Overclocking issues?!?!

So... Does anyone have any further insight, suggestions, comments, or ideas? I'd love to get to 3.2GHz on a stock or very close to normal Vcore without any other voltage increases. I'd finally be happy and call it good then.

Thanks a lot guys and gals.
-Mike
 
Mike, I have the ga-p35c-ds3r and I'm at 3.6.

It's great to see you're trying hard to solve the problems, way to join [H].
However, I noticed you set your vcore pretty high (1.4 @ 3.2 seems high to me).

I currently run 1.40 vcore (in bios) @ 3.6 GHz.
I've also ran 3.2 GHz with only a slight increase in vcore, I think around 1.3625 or 1.375v. Maybe you can try that?
Given we don't have exact same hardware and no two cpu OC's the same, I hope that this helped.

Also, I have +0.1 MCH, and all the others at default with RAM at +0.3v (my ram needs the 2.1v).
 
I'll go ahead and give my Vcore a small bump up from normal, but i'm 100% sure it won't work cause i've tried it many many times, haha. i'll post up a screen shot as well.

Thanks for the support!
I'm really committed to solving my OC issues, whether i got a bum chip or there are issues elsewhere.

-Mike
 
No problem, I know how you feel...I started not too long ago too.

Screenies really help too, as everyone can see what you see :D
Looking forward to seeing a good oc!
 
Your temps are really high, but that is expected because of your higher voltages. I guess maybe you have a bad cpu in terms of overclocking. Then again maybe your power isn't steady. It's weird what's going on with your fridge and your pc. It's quite a bummer - you've good cooling, good power (750w).
 
well, i ran a test at 3.2GHz (8x400) at 1.35Vcore, all other voltages normal and the computer rebooted at the windows loading GUI. Then I bumped it up to 1.375 and it booted. here are the results....

Booted up, before test. Notice the Voltage in CPU-Z (its 1.375 in BIOS) as well as the temps at idle.
boot1375fz1.jpg


Here is a screen shot while testing... Notice the Voltage in CPU-Z (is that a bad Vdroop?) and the core temps.
fail1375of7.jpg



obviously it didn't work... failed immediately.
Also, i unplugged my fridge for this test too, so i guess i can rule that out as being the issue... bummer.

I don't know what my next step in the testing is.... do I give up and settle for 3.0GHz?
It makes me want to cry. What about replacement through Intel? my newegg 30-day is gone. Would i need to prove that even at its stock 2.66GHz it fails Prime95?

frustrated - mike
 
Wow, what are the chances...I really hope you don't have a half-dead e6750.

Are you saying it fails prime @ stock too?
 
no no, at its stock 2.66 it'll run Prime95 fine. I kinda wish it didn't, then i could get another one and say this one's defective. I'm wondering more and more if its my RAM. When i bought all my stuff, i think i went for quantity over quality. Will unstable RAM effect Prime95's small FFt test? I'm gonna try running some tests using lower than stock speeds with my ram at 800MHz to see if its really stable or not.
 
What is your memory configuration? Is it 4x1 or 2x2? 4 sticks will strain the overclock much more. Try running a memtest with the stock and with the overclock to see if it passes. Also, what are the volt requirements and max for the memory? remember the ds3's default is 1.8V. If memtest is passing but your o/c fails at 3.2, then I think you should just settle for 3.0 for the time being and do more reading/searching.

This is unfortunate: I'm okay at up to 400x8 at stock volts though I've run orthos and OCCT for only an hour 4 mths ago when overclocking.

My only anomaly was a desktop freeze (solved by pci-e +0.1, bugs in vmware so that's software). I also had a bootup with no network signal once. Still haven't figured out that one. Why do i tell you all this? Because if 3.0 is working, you'll be running at a stable o/c and can work from there. That's still better than trying to squeeze out more mhz but not finding the problem.
 
Back
Top