• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

IP35 Pro, E8400, 8GB ram. Need advice on settings.

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
N

Nypyren

New Member
Joined
Feb 4, 2008
Mobo: IP35 Pro v1.1
CPU: E8400
RAM: 4 sticks of 2GB GSkill DDR2-1000.

I've almost got it stable at 3.6GHz (which should be cake), yet I still have random BSODs and other strange behavior under certain conditions.

I'm convinced it's a voltage issue of some sort since the system becomes much more stable as I raise MCH and DDR2 voltage. Yesterday I was able to run The Witcher all day without any issues, so I believe I found a stable or nearly-stable setup.

BIOS: 16 Beta 04.

VCore is 1.3625 (reads 1.22V after droop)

DDR2 is at 2.2V (the sticks are rated for 2.1V), and the timings are all Auto. I would have left this at 2.1V except that I get an error immediately upon booting windows. The sticks get pretty warm to the touch.

CPU VTT I left at stock settings because I have no idea what it's for.

MCH is currently 1.42v (I kept bumping it up until it was stable... how much is too much?).

I left both ICH voltages at stock, which might be the cause of my remaining issue:

I have four hard drives in the system, and two are RAID0'ed. Just today I got another BSOD in the middle of a 100+GB transfer from a storage drive to my RAID array.

Anyone have any advice on whether (or how far) I should increase/decrease voltages?
 
Nypyren said:
Mobo: IP35 Pro v1.1
CPU: E8400
RAM: 4 sticks of 2GB GSkill DDR2-1000.

I've almost got it stable at 3.6GHz (which should be cake), yet I still have random BSODs and other strange behavior under certain conditions.

I'm convinced it's a voltage issue of some sort since the system becomes much more stable as I raise MCH and DDR2 voltage. Yesterday I was able to run The Witcher all day without any issues, so I believe I found a stable or nearly-stable setup.

BIOS: 16 Beta 04.

VCore is 1.3625 (reads 1.22V after droop)
I would bump the cpu voltage and see what happens , btw
I just bought that board , can you explain to me how to set up raid 0 on it , the board should be in by 2 days

DDR2 is at 2.2V (the sticks are rated for 2.1V), and the timings are all Auto. I would have left this at 2.1V except that I get an error immediately upon booting windows. The sticks get pretty warm to the touch.

CPU VTT I left at stock settings because I have no idea what it's for.

MCH is currently 1.42v (I kept bumping it up until it was stable... how much is too much?).

I left both ICH voltages at stock, which might be the cause of my remaining issue:
bump the voltage on cpu should take care of it

I have four hard drives in the system, and two are RAID0'ed. Just today I got another BSOD in the middle of a 100+GB transfer from a storage drive to my RAID array.

Anyone have any advice on whether (or how far) I should increase/decrease voltages?
Click to expand...
 
Put 1.3V actual in that CPU to give you some room and start bumping up the clocks. Leave the other votages stock until your moving past 475FSB or so.
 
Still having issues. Orthos is happy, memtest is happy, but programs still crash. I think it actually might be that my RAID setup is horked and everything else is fine.

I updated to BIOS 16.b09, set almost everything to stock/auto settings and put the DDR2 at its manufacturer spec (2.1v :: 5-5-5-15). Windows loaded up happily. However, every few minutes a random program that uses the hard drives will crash. A large game I tried installing is the only thing that's consistent -- it gives me a CRC error in the exact same spot while installing (the DVD drive and disc are good - I can take the entire disc and drive to another machine and it'll work fine).

The thing that I noticed too late is that the RAIDed drives aren't perfectly identical. They're just the same size. One is SATA2, the other is SATA1 (but both are WD drives). I'm wondering if that could cause these kinds of issues. I'm pretty sure the drives themselves are stable since I was using them (not RAIDed) on my old KN8 Ultra.

Anyway I'm going to reinstall without any RAIDing this evening and report back.
 
Update: Not a hard drive or RAID issue; Issue was reproducable with stock settings, 2 sticks of RAM (4GB total), and one hard drive. I disconnected everything else that was drawing power. Both sticks passed 9 hours of memtest and the hard drives all passed SMART and full surface scan testing.

The only thing I have left to isolate is the power supply. I have an Enermax Liberty 500W modular PSU. I'm going to borrow my roommate's Corsair HX620 tonight to solve this once and for all.

If his PSU doesn't fix it, I'm going to RMA the motherboard.
 
Yes, check out the psu angle. I've read over at the jonnyGURU forums of problems with the Enermax liberty psu's dying or giving problems.
 
This solution is easy in my opinion.

Start by lowering the memory voltage. Those modules shouldn't need more than 2.0v, I use the same memory and have ran them as high as 1066MHz at the stock 2.0v. Excessive voltage just creates more heat and possibly instability, not to mention those modules simply don't respond to more than 2.1v and even that isn't necessary from my own experimentation.

Your instability is either in the MCH voltage or the CPU voltage as both are required for stability. If your CPU voltage is high enough which it looks about right for your speed, increase the MCH voltage. I have run as high as 1.52v MCH without any issues at all and in fact I found it necessary in order to maintain stability at higher FSB speeds. The P35 and X38 chipsets need more voltage than the default voltage in order to unlock their potential, 1.44v is fine for two 2GB modules, but with four you will probably need more voltage. Because you are running four high density smodules, it's creating more stress on the P35 chipset and the CPU.

And this is all of course assuming your power supply isn't low quality and/or underpowered since it appears you're running alot of power hungry devices in your system. It would also be helpful if we knew what multiplier and FSB you're trying to use to achieve 3.6GHz.
 
I currently have my roommate's Corsair HX620 in my machine. I'm having about the same luck.

The system seems slightly more stable than with my other PSU so I think I may make a run to Fry's tomorrow.


I think you're right on the money as far as "lower the DDR2 voltage" goes; I've been setting the RAM to 2.1v thinking "more is better" all this time. I just set them to 2.075 and noticed a bit of improvement so I went back to the BIOS and let it use AUTO (which is decided was 1.93v according to the in-bios voltage viewer).

So far, Orthos has gone longer at 1.93 than it did with 2.1v, which is in accordance with your advice.


As far as CPU overclocking goes, I was using 400x9.0. Right now I'm trying to eliminate all possible variables of instability so I have it at 340x9 (near stock) and 1:1 memory clocking ratio (DDR2-680).

Orthos just bit it, so I'm going to go set DDR2 to 2.0 and mess with the MCH and CPU voltages ONLY for a while to see how that goes.

(Edit)
Is adjusting voltages using uGuru in Vista64 trustworthy? The very newest version runs with 16.b09 but the CPU voltages begin at 0.00 and work up to 0.03 (which seem to be offset from the 1.20 or 1.21 minimum). I'm never sure if I can REALLY trust the thing so I go back to the BIOS to tweak volts, but if it works it would drastically speed up my tweaking...
 
Last edited:
Ok, problem confirmed. I found a nice "stable" (40 minutes in Orthos - longer than I've usually achieved this week with all four sticks in place) setting for all 8GB sticks with the processor at stock speed. This is using my roommate's Corsaid HX620 still.

So I decided to connect my other power supply to see if it was part of the problem. Orthos died after only 2 minutes. I reconnected the Corsair and it's still running Orthos as I type.

So tomorrow I'm heading up to Fry's and getting a new power supply.

I'm torn between a few PSUs that I've heard good things about: The Corsair HX620 (obviously since it works, that's a good sign), the PC Power & Cooling 610W (single rail) or a larger, single rail Corsair.

Now, I have to ask those with experience: For THIS setup, should I focus on getting a massive, single rail PSU, or should I just get an HX620? I'll never get a second video card, but I will put in up to 6 SATA hard drives.
 
Single rail PSU's are considered the most reliable under extreme applications. There is a reason why the major manufacturers are still producing large single rails PSU's.
 
Nypyren said:
I currently have my roommate's Corsair HX620 in my machine. I'm having about the same luck.

The system seems slightly more stable than with my other PSU so I think I may make a run to Fry's tomorrow.


I think you're right on the money as far as "lower the DDR2 voltage" goes; I've been setting the RAM to 2.1v thinking "more is better" all this time. I just set them to 2.075 and noticed a bit of improvement so I went back to the BIOS and let it use AUTO (which is decided was 1.93v according to the in-bios voltage viewer).

So far, Orthos has gone longer at 1.93 than it did with 2.1v, which is in accordance with your advice.


As far as CPU overclocking goes, I was using 400x9.0. Right now I'm trying to eliminate all possible variables of instability so I have it at 340x9 (near stock) and 1:1 memory clocking ratio (DDR2-680).

Orthos just bit it, so I'm going to go set DDR2 to 2.0 and mess with the MCH and CPU voltages ONLY for a while to see how that goes.

(Edit)
Is adjusting voltages using uGuru in Vista64 trustworthy? The very newest version runs with 16.b09 but the CPU voltages begin at 0.00 and work up to 0.03 (which seem to be offset from the 1.20 or 1.21 minimum). I'm never sure if I can REALLY trust the thing so I go back to the BIOS to tweak volts, but if it works it would drastically speed up my tweaking...
Click to expand...


It sounds like you're heading in the right direction now. I'll give you my settings and see if it works just as well for you. Make sure you set this in the BIOS, and not using uGuru.

FSB: 500MHz
Multiplier: 8
CPU Voltage: 1.425v
MCH voltage: 1.44v
DDR voltage: 2.0v

Get a decent PSU and I see no reason why you can't run these settings. I've run Orthos for over 12 hours with these settings.
 
It's official: I suck at overclocking.

:bang head

OK. New PSU (Corsair TX750) makes the system a lot less prone to BSODs and freak lock-ups. Unfortunately I still completely suck at finding settings that are stable.

I haven't tried your settings because I honestly don't think my case's cooling configuration is up to the task of dealing with 1.42v. I don't entirely trust the temperature readings and I'm scared of melting components.

I have front and rear 120mm fans and an Arctic Cooling Freezer 7 on the CPU. The new PSU (Corsair TX750) has one massive fan but it only ramps up if its own sensor gets hot.

The RAM is hot to the touch at any settings and there's no air being blown directly at them, just some ambient airflow from the front intake headed toward the back.



The most I feel safe trying is 3.6GHz...

Linearly interpolating between [email protected] and [email protected] *should* land me at 1.34v but I don't know if it even works like that. I've heard people clocking to 3.6GHz at stock voltage (1.22v). What's up with that? My system is totally unstable at that voltage.


The most stable I've got is the following which lasted 2 hours and 52 minutes.

400x9.0
1:1 (DDR2-800)
CPU @ 1.35 (reads as 1.34)
DDR @ 2.0
VTT @ 1.25
MCH @ 1.52
ICH @ 1.09
ICHIO @ 1.50
5-5-5-15-etc(Auto)

I've tried 23 other slight variations which all fail in Orthos at some point. I had the CPU fan on 100% for all variants. I'm not sure if bumping up the CPU voltage any more is going to help or not.

Without multiple trials at the same settings, Orthos's time-to-failure isn't a linear indicator of how stable the system is. It's true/false - your system is stable or it's not, and the failure can happen at any point.

My MCH is maxed at 1.52v - I'm not sure if that's too high and contributing to an overheat instability or if it's too low to handle 8GB. Unfortunately using 1.55 or higher halts my box before POST and requires a CCMOS.

I haven't attempted any higher CPU with lower MCH (1.38+ / 1.44-) combinations yet.

Maybe I should change the DDR timings? How should I go about doing that? I tried simply incrementing all values (higher = slower, right?) but the system hated me even MORE when I did that.
 
Nypyren said:
:bang head

OK. New PSU (Corsair TX750) makes the system a lot less prone to BSODs and freak lock-ups. Unfortunately I still completely suck at finding settings that are stable.

I haven't tried your settings because I honestly don't think my case's cooling configuration is up to the task of dealing with 1.42v. I don't entirely trust the temperature readings and I'm scared of melting components.

I have front and rear 120mm fans and an Arctic Cooling Freezer 7 on the CPU. The new PSU (Corsair TX750) has one massive fan but it only ramps up if its own sensor gets hot.

The RAM is hot to the touch at any settings and there's no air being blown directly at them, just some ambient airflow from the front intake headed toward the back.



The most I feel safe trying is 3.6GHz...

Linearly interpolating between [email protected] and [email protected] *should* land me at 1.34v but I don't know if it even works like that. I've heard people clocking to 3.6GHz at stock voltage (1.22v). What's up with that? My system is totally unstable at that voltage.


The most stable I've got is the following which lasted 2 hours and 52 minutes.

400x9.0
1:1 (DDR2-800)
CPU @ 1.35 (reads as 1.34)
DDR @ 2.0
VTT @ 1.25
MCH @ 1.52
ICH @ 1.09
ICHIO @ 1.50
5-5-5-15-etc(Auto)

I've tried 23 other slight variations which all fail in Orthos at some point. I had the CPU fan on 100% for all variants. I'm not sure if bumping up the CPU voltage any more is going to help or not.

Without multiple trials at the same settings, Orthos's time-to-failure isn't a linear indicator of how stable the system is. It's true/false - your system is stable or it's not, and the failure can happen at any point.

My MCH is maxed at 1.52v - I'm not sure if that's too high and contributing to an overheat instability or if it's too low to handle 8GB. Unfortunately using 1.55 or higher halts my box before POST and requires a CCMOS.

I haven't attempted any higher CPU with lower MCH (1.38+ / 1.44-) combinations yet.

Maybe I should change the DDR timings? How should I go about doing that? I tried simply incrementing all values (higher = slower, right?) but the system hated me even MORE when I did that.
Click to expand...

The memory could be the problem. Do not go past 2.2v on the memory, mine died at 2.3v. And yes, higher timings means slower. But 5-5-5-15 @ 2.2v should be just fine. My memory is running at 4-4-4-12 @ 2.2v @ 458Mhz 1:1.
 
For anyone still wondering whether I ever fixed it:

I gave up and got a completely different motherboard and all of my issues are gone (and I didn't have to change a single voltage setting from Auto). So at least one of the following is true:

- Abit IP35 Pro hates 8GB of G.Skill PC2-8000.

or

- I had a defective motherboard and the whole experience was a fluke.


Thanks for all of the help, anyway - it was quite a learning experience!
 
I had similar issues w/ my Wolfdale. Read some info on XS about 45nm CPUGTL REF voltages. When this was set to default on the 9 BIOS equals 67% - lowered it to 63% and I am rock solid stable. Check it out, I can prolly push this a bit more, but I'm happy here...till...Monday. :D

IP35.PRO.440x8.3960.Wolfdale.jpg
 
Nypyren said:
Mobo: IP35 Pro v1.1
CPU: E8400
RAM: 4 sticks of 2GB GSkill DDR2-1000.

I've almost got it stable at 3.6GHz (which should be cake), yet I still have random BSODs and other strange behavior under certain conditions.

I'm convinced it's a voltage issue of some sort since the system becomes much more stable as I raise MCH and DDR2 voltage. Yesterday I was able to run The Witcher all day without any issues, so I believe I found a stable or nearly-stable setup.

BIOS: 16 Beta 04.

VCore is 1.3625 (reads 1.22V after droop)

DDR2 is at 2.2V (the sticks are rated for 2.1V), and the timings are all Auto. I would have left this at 2.1V except that I get an error immediately upon booting windows. The sticks get pretty warm to the touch.

CPU VTT I left at stock settings because I have no idea what it's for.

MCH is currently 1.42v (I kept bumping it up until it was stable... how much is too much?).

I left both ICH voltages at stock, which might be the cause of my remaining issue:

I have four hard drives in the system, and two are RAID0'ed. Just today I got another BSOD in the middle of a 100+GB transfer from a storage drive to my RAID array.

Anyone have any advice on whether (or how far) I should increase/decrease voltages?
Click to expand...

exactly which ram?
planning q6600 oc to 3.2-3.6 on 8gb of ram and ip35 pro mobo 6 sata hdd's etc etc - 3 that i'm considering amongst others:

http://www.newegg.com/Product/Product.aspx?Item=N82E16820231122
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231148
http://www.newegg.com/Product/Product.aspx?Item=N82E16820231145
 
You must log in or register to reply here.

Similar threads

S
4x4GB DDR3 Overclock on AM3+
Replies
2
Views
338
Sologuy
S
S
CMT32GX5M2X6200C36 *4 Sticks Z790-E GAMING WIFI i9-13900K XMP Unstable
Replies
4
Views
664
Woomack
W
K
Overclocking RAM to 6000mhz on 7800x3d instable
2
Replies
27
Views
3K
Woomack
W
O
Need advice on Single CCD Overclock for 5900X CPU: Lower DPC Latency without Sacrificing Stability?
Replies
14
Views
2K
freeagent
freeagent
D
  • Question
Need timings to finish "overclocking" RAM to run at actual speed (1600Mhz) when motherboard only accepts 1333Mhz unless "overclocked"
Replies
7
Views
1K
WhitehawkEQ
WhitehawkEQ
Back