Help Overclock E7200

[geokilla]

Hi everyone. I have a E7200 paired with a EP35-DS3L using the F5 BIOS. I stressed 25 min of OCCT last night at 333X9.5 and it ran fine for that time. At the moment, I'm pushing 350FSB with 1.24Vcore. However, no matter what voltages I give it, the system can't pass over a min of OCCT. CPU-Z will read the vcore as 1.184V under idle, and 1.168V under load whether I give it 1.22Vcore or 1.25Vcore.

I already have the RAM running at 1:1, so it's running at DDR700 5-5-5-15 @ 1.9V. The MCH voltage is at +0.1V.

Can anyone help me? My temps are at 30C and 36C under idle. They are measured using the latest Core Temp.

Edit: May have found the problem and solution. link

 

[geokilla]

New question.

How much vcore should you give a E7200 from the BIOS and CPU-Z? At the moment, I'm sitting at 1.28Vcore (CPU-Z idle) running the CPU at 3.6Ghz.

 

[mancunian]

I went for just a little bit more than you, mine's at 1.296v in CPU-Z, rig in sig.

But your voltage seems about right to me. You can probably get it to 3.8 with a slight increase, 4Ghz might require too much juice.

But then again, it might not. I'd give it a shot if I were you.


Edit: You might wanna change the topic summary, otherwise some people might just not bother reading the thread, thinking that you don't need any more answers.

 

[geokilla]

Originally posted by: mancunian
I went for just a little bit more than you, mine's at 1.296v in CPU-Z, rig in sig.

But your voltage seems about right to me. You can probably get it to 3.8 with a slight increase, 4Ghz might require too much juice.

But then again, it might not. I'd give it a shot if I were you.


Edit: You might wanna change the topic summary, otherwise some people might just not bother reading the thread, thinking that you don't need any more answers.

Ya, thanks for the reminder and everything.

Currently, I'm sitting at 3.8Ghz with 1.36Vcore at idle, and 1.328V under OCCT load (CPU-Z) with +0.1MCH and FSB voltages. Is it safe to change the MCH/FSB to +0.2 and try to lower my vcore? What's your MCH and FSB voltages?

 

[mancunian]

Originally posted by: geokilla
Ya, thanks for the reminder and everything.

Currently, I'm sitting at 3.8Ghz with 1.36Vcore at idle, and 1.328V under OCCT load (CPU-Z) with +0.1MCH and FSB voltages. Is it safe to change the MCH/FSB to +0.2 and try to lower my vcore? What's your MCH and FSB voltages?

No probs.

I'd say you're at the sensible max, Intel's safe spec is up to 1.36v I think. So it'd be up to you if you wanted to go any further.

It won't matter if you raise MCH/FSB, it wouldn't enable you to lower the vcore.

My MCH/FSB voltages are at default, I can't see any reason why you would increase yours as that board supports 1600Mhz FSB, I would only increase CPU voltage when a stress test fails. You might wanna try setting those back to default and seeing if your CPU voltage on its own has made things stable. Also, if you are using DDR2 667, I can understand you bumping the RAM voltage a bit, but not if you're using DDR2 800 or above.


Good luck with it, it's a great chip.

 

[geokilla]

Here's an update to my overclocking. I'm trying to reach 3.8Ghz, but currently sitting at 3.7Ghz. 3.8Ghz requires a bit too much voltage as it can't even pass 1 minute of OCCT. Any advice is greatly appreciated.

Here are my BIOS settings:

C1E, EIST, TM2: All disabled
FSB: 390Mhz
PCI-E: 100Mhz
C.I.A.2: Disabled
RAM Multiplier: 2.00 (1:1)
RAM Timings: 5-5-5-15
FSB Voltage: +0.1V
MCH Voltage: Standard
CPU Voltage: 1.3875V in BIOS. CPU-Z Idle: 1.344V. CPU-Z Load:1.312

@Man, how'd you get it to 3.8Ghz stable using just 1.3V? I'm trying to get 3.7Ghz stable with 1.31V.

 

[mancunian]

I suppose I just got lucky with this chip. Some got even luckier, there's a guy on here has his at 4Ghz using about 1.28v. But 1.3v is not bad at all either, and I think it might have a little more in it.

Your settings look ok to me, aside from the FSB voltage. Does that need to be at +1? Because you aren't overclocking the FSB beyond what the board supports, you are actually 10Mhz under the highest supported FSB.

But the rest look ok. You might want to flash to the latest BIOS if you haven't already. But to me, flashing the BIOS for an extra 200 or so Mhz just isn't worth it.


If it was me, I'd keep it at 3.7 if the rig is stable and your temps are acceptable. These chips tend to top out at 4Ghz, so you're not far off the max as it is. Anything over 3Ghz on a Core 2 is plenty. And a 1.2Ghz overclock is not at all shabby. Your old X2 would have to be running at about 4.1Ghz to match what you have now.

 

[big4x4]

I just got a e7200 at frys for $99 to replace the cpu in the sig. Never booting up at stock, I immediately set it at 3.8ghz w/ 1.25v. Well, that was 100% stable so now I am down to 1.21v 24 hours orthos and 2 hours OCCT stable. I would push for 4.0 but it is over 100 degrees outside and my temps would be too high. Give it sometime and you will eventually find the right setting to get it stable!

Forgot to mention my settings:
Abit Ip35-e
e7200 @ 3.8 (9.5*400)
2gb ddr2 800 @ 4-4-3-10 (1.9v)
vcore- 1.21v bios (load = 1.75)
Everything else stock

 

[mancunian]

Originally posted by: big4x4
I just got a e7200 at frys for $99 to replace the cpu in the sig. Never booting up at stock, I immediately set it at 3.8ghz w/ 1.25v. Well, that was 100% stable so now I am down to 1.21v 24 hours orthos and 2 hours OCCT stable. I would push for 4.0 but it is over 100 degrees outside and my temps would be too high. Give it sometime and you will eventually find the right setting to get it stable!

Now THAT'S a good chip.

:thumbsup:

 

[geokilla]

Originally posted by: big4x4
I just got a e7200 at frys for $99 to replace the cpu in the sig. Never booting up at stock, I immediately set it at 3.8ghz w/ 1.25v. Well, that was 100% stable so now I am down to 1.21v 24 hours orthos and 2 hours OCCT stable. I would push for 4.0 but it is over 100 degrees outside and my temps would be too high. Give it sometime and you will eventually find the right setting to get it stable!

Forgot to mention my settings:
Abit Ip35-e
e7200 @ 3.8 (9.5*400)
2gb ddr2 800 @ 4-4-3-10 (1.9v)
vcore- 1.21v bios (load = 1.75)
Everything else stock

Wow that's very nice. I might be able to put it at 3.8Ghz too, but I'd probably need around 1.36V under load. I'm scared going above 1.4V since it's 45nm, and too much vcore will significantly shorten the lifespan of the CPU. One example is AnandTech's QX9650, which I believe died from too much voltages.


I'm sitting at 1.312V under load right now. I found something weird with Linpack though. I already raised the voltages up a bit notches, and Linpack constantly fails at the 5556 second mark, saying that it's passed only 1 out of 50 tests. Anyone got an idea what the problem could be?

 

[error8]

Originally posted by: geokilla


I'm sitting at 1.312V under load right now. I found something weird with Linpack though. I already raised the voltages up a bit notches, and Linpack constantly fails at the 5556 second mark, saying that it's passed only 1 out of 50 tests. Anyone got an idea what the problem could be?

The problem is your voltage, which is too low. Check my sig.
I found it to be stable at 1.41V ( 1.38 V load) for 3,9 ghz , 1,375 V ( 1,35 V load) for 3,8 ghz and 1,45 V (!) for 4 ghz ( 1,42 V load).

At 4 ghz I also fail in Linpack , but I can pass hours and hours of OCCT and any other program out there. If I'll get any instability issues, then I'll back it down at 3,9 ghz where I know is even Linpack stable ( 10 tests ), but I don't think that this will ever happen. Linpack is the best tool for measuring stability, but it has nothing to do with real life situations.

 

[geokilla]

Originally posted by: error8

Originally posted by: geokilla
I'm sitting at 1.312V under load right now. I found something weird with Linpack though. I already raised the voltages up a bit notches, and Linpack constantly fails at the 5556 second mark, saying that it's passed only 1 out of 50 tests. Anyone got an idea what the problem could be?

The problem is your voltage, which is too low. Check my sig.
I found it to be stable at 1.41V ( 1.38 V load) for 3,9 ghz , 1,375 V ( 1,35 V load) for 3,8 ghz and 1,45 V (!) for 4 ghz ( 1,42 V load).

At 4 ghz I also fail in Linpack , but I can pass hours and hours of OCCT and any other program out there. If I'll get any instability issues, then I'll back it down at 3,9 ghz where I know is even Linpack stable ( 10 tests ), but I don't think that this will ever happen. Linpack is the best tool for measuring stability, but it has nothing to do with real life situations.

That's true. I believe Linpack is the current stress-testing program that stresses the most out of your system.

This time, the test finished in 5455 seconsd, passing 7 out of 50 tests. I'll now try raising the voltages a bit more and see what happens. However, I probably won't raise the FSB and MCH yet till I find out whether it'll kill my system components or not since I don't have active cooling on the MCH.

 

[error8]

Originally posted by: geokilla

Originally posted by: error8

Originally posted by: geokilla
I'm sitting at 1.312V under load right now. I found something weird with Linpack though. I already raised the voltages up a bit notches, and Linpack constantly fails at the 5556 second mark, saying that it's passed only 1 out of 50 tests. Anyone got an idea what the problem could be?

The problem is your voltage, which is too low. Check my sig.
I found it to be stable at 1.41V ( 1.38 V load) for 3,9 ghz , 1,375 V ( 1,35 V load) for 3,8 ghz and 1,45 V (!) for 4 ghz ( 1,42 V load).

At 4 ghz I also fail in Linpack , but I can pass hours and hours of OCCT and any other program out there. If I'll get any instability issues, then I'll back it down at 3,9 ghz where I know is even Linpack stable ( 10 tests ), but I don't think that this will ever happen. Linpack is the best tool for measuring stability, but it has nothing to do with real life situations.

That's true. I believe Linpack is the current stress-testing program that stresses the most out of your system.

This time, the test finished in 5455 seconsd, passing 7 out of 50 tests. I'll now try raising the voltages a bit more and see what happens. However, I probably won't raise the FSB and MCH yet till I find out whether it'll kill my system components or not since I don't have active cooling on the MCH.

Really, there is no need for extreme voltages for your MCH. If you want to do something like 8X500 then you should up that voltage to something over 1,4 V. At 421 mhz I'm using 1,375 V on MCH. We have different mobos, but you should be around the same voltage as I am for 421 FSB.

 

[geokilla]

Update: Got my rig to 2hrs+ OCCT stable last night. Then I went to sleep, dropped the core voltage by one notch, and it was 1 hour OCCT stable. So I raised the voltages, and it was on 30 min OCCT stable. Raised it again, and it was only 7 min OCCT stable?! I'm running OCCT again without touching the voltages, and it was 1 hour stable. Now, it's the same voltages, but +0.1V on the FSB

Is it just me, or is OCCT not that reliable anymore?

This is with the CPU at 3.75Ghz. Temps top out at around 51C.