Gigabyte P965-P35 X38 overclocking and BIOS tweaking Guide

Hiya guys, just finished my first build this week and have been playing around with trying to overclock it. Getting couple of errors.

First of all my hardware:
Intel e6600
Bufalo Firestix 4 x 1gb
Gigabyte DS3 rev3.3, F11 bios
PNY 8800GTS 320mb
Corsair 620HX
Barracuda 7200.10 320gb HD
Arctic Freezer 7 pro
Thermalright HR05 SLI
Antec p182 Case

I've been following this guide to OC my DS3

From the many hours lurking these forums I felt that the following overclock was very achievalbe:

CPU @ 3150mhz
350 FSB x 9
Vcore 1.375

RAM @ 437.5mhz (4:5 ratio)
875mhz 4-4-4-12 timings @ 2.2v

So, the above is the OC I was intending to keep. However, something is apparently not stable. I've had a few problems at this speed.

1) sometimes when I boot the computer for the first time that day, the computer would load but the monitor would act like no signal is recieved. This happened a couple of times, I've reseated my vid card. Its has since happened once again but been behaving for the last couple of days.:rolleyes:

2) After setting the ram timings and FSB speed, I boot into memtest86+ to check the stability of the ram. I let it run till it gets 1 pass without errors (approx 45 mins). The ram is pretty hot to touch (I can put my finger on the heatsink for 5 secs or so) but looks ok according to memtest86+. I quit the program and reboot. The computer would not post, it attempts to boot a couple of times then reverts to lower FSB/RAM settings.:confused:

3) Being stubborn, :D I reinput same settings to see what temps look like. I start up CoreTemp, SpeedFan, SuperPi, CPUZ, and Orthos. Orthos ran fine the first time at priority1 for about half and hr with no errors. Idle temp is 39o and never goes above 55o under load. Later, I run all the above again and orthos stops about 5 mins in :mad: . The following is from the splog file in orthos directory:

Type: Blend - stress CPU and RAM Min: 8 Max: 4096 InPlace: No Mem: 3837 Time: 15
CPU: 3150MHz FSB: 350MHz [350MHz x 9.0 est.]
5/1/2007 9:37 AM
Launching 2 threads...
1:Using CPU #0
2:Using CPU #1
1:Beginning a continuous self-test to check your computer.
1:press Stop to end this test.
2:Beginning a continuous self-test to check your computer.
2:press Stop to end this test.
1:Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
2:Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
2:Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
1:Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
2:Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
1:Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
1:FATAL ERROR: Rounding was 0.5, expected less than 0.4
1:Hardware failure detected, consult stress.txt file.
1:Torture Test ran 4 minutes 50 seconds - 1 errors, 0 warnings.
1:Execution halted.

2:Torture Test ran 4 minutes 50 seconds - 0 errors, 0 warnings.
2:Execution halted.

I cannot find the 'stress.txt' file it refers to.

Any thoughts as to what you think is unstable in this system? Any thoughts are greatly appreciated. (I've also posted this in the OC forums)

KD

Edit: here are the messages from inside orthos (its greek to me)
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 4 minutes 50 seconds - 1 errors, 0 warnings.
Execution halted.

Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
Torture Test ran 4 minutes 50 seconds - 0 errors, 0 warnings.
Execution halted.
 
Haven't read the whole thread yet but just wanted to say the thread is well written and the colours and sections help readability a lot.

Not a lot of new info for anyone who's OC'd before but definitely a great guide for beginners.
 
KD28 said:
Hiya guys, just finished my first build this week and have been playing around with trying to overclock it. Getting couple of errors.

First of all my hardware:
Intel e6600
Bufalo Firestix 4 x 1gb
Gigabyte DS3 rev3.3, F11 bios
PNY 8800GTS 320mb
Corsair 620HX
Barracuda 7200.10 320gb HD
Arctic Freezer 7 pro
Thermalright HR05 SLI
Antec p182 Case

I've been following this guide to OC my DS3

From the many hours lurking these forums I felt that the following overclock was very achievalbe:

CPU @ 3150mhz
350 FSB x 9
Vcore 1.375

RAM @ 437.5mhz (4:5 ratio)
875mhz 4-4-4-12 timings @ 2.2v

So, the above is the OC I was intending to keep. However, something is apparently not stable. I've had a few problems at this speed.

1) sometimes when I boot the computer for the first time that day, the computer would load but the monitor would act like no signal is recieved. This happened a couple of times, I've reseated my vid card. Its has since happened once again but been behaving for the last couple of days.:rolleyes:

2) After setting the ram timings and FSB speed, I boot into memtest86+ to check the stability of the ram. I let it run till it gets 1 pass without errors (approx 45 mins). The ram is pretty hot to touch (I can put my finger on the heatsink for 5 secs or so) but looks ok according to memtest86+. I quit the program and reboot. The computer would not post, it attempts to boot a couple of times then reverts to lower FSB/RAM settings.:confused:

3) Being stubborn, :D I reinput same settings to see what temps look like. I start up CoreTemp, SpeedFan, SuperPi, CPUZ, and Orthos. Orthos ran fine the first time at priority1 for about half and hr with no errors. Idle temp is 39o and never goes above 55o under load. Later, I run all the above again and orthos stops about 5 mins in :mad: . The following is from the splog file in orthos directory:

Type: Blend - stress CPU and RAM Min: 8 Max: 4096 InPlace: No Mem: 3837 Time: 15
CPU: 3150MHz FSB: 350MHz [350MHz x 9.0 est.]
5/1/2007 9:37 AM
Launching 2 threads...
1:Using CPU #0
2:Using CPU #1
1:Beginning a continuous self-test to check your computer.
1:press Stop to end this test.
2:Beginning a continuous self-test to check your computer.
2:press Stop to end this test.
1:Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
2:Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
2:Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
1:Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
2:Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
1:Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
1:FATAL ERROR: Rounding was 0.5, expected less than 0.4
1:Hardware failure detected, consult stress.txt file.
1:Torture Test ran 4 minutes 50 seconds - 1 errors, 0 warnings.
1:Execution halted.

2:Torture Test ran 4 minutes 50 seconds - 0 errors, 0 warnings.
2:Execution halted.

I cannot find the 'stress.txt' file it refers to.

Any thoughts as to what you think is unstable in this system? Any thoughts are greatly appreciated. (I've also posted this in the OC forums)

KD

Edit: here are the messages from inside orthos (its greek to me)
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 4 minutes 50 seconds - 1 errors, 0 warnings.
Execution halted.

Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
Torture Test ran 4 minutes 50 seconds - 0 errors, 0 warnings.
Execution halted.
Click to expand...

Maybe try more core voltage, or 2 sticks of RAM to see if either of those are stable.
 
I finished putting my new system together last nite and have been tweaking the bios ever since.

So far I've been able to get
e6600 @ 400x8 3.2 1.400 v
Corsair DDR2 800 2.1 v 4-4-4-12
Temps are 41°c idle and 65°c load. (would like the temps to be lower) Zalman 9500 at full speed

Failing Orthos though, any suggestions?

I'm a novice at this over clocking thing and this guide has helped me get the specs mentioned above.
 
1) sometimes when I boot the computer for the first time that day, the computer would load but the monitor would act like no signal is recieved. This happened a couple of times, I've reseated my vid card. Its has since happened once again but been behaving for the last couple of days.
Click to expand...

I've noticed this on my 8800GTS as well while using DVI. I'm guessing that DVI handshake is being slow - if you let it finish booting, you do eventually get the Windows login screen with no issues. This doesn't happen at all with D-Sub, which is why I assume it's just a slow DVI handshake.
 
What is the best way to get 1066 Mhz on my PC2-8500 without raising the FSB? I tried just raise the ram multiplier to x4 and it shows 1066. It also shows voltage not optimized. Is that all I need to do?
 
Thank you soooooooooo much for this thread. This is so amazing for noobs like me who have never overclocked before! Will you be adding more to this thread is is this all that I will have to know?
 
Here goes. 1. What is the purpose of lowering your multiplier from 9 to 8? With a higher FSB you will get higher speed? With a lower multiplier, you have to raise FSB that much higher to get to same speed. What are the advantages of changing from 9 to 8? What am I missing? I am currently at 3ghz with a 9 x 333, 1.375v. 31c on Hyper XT cooling on a e-6600, Giga rev. 3.3. Corsair XMS 4-4-4-12 memory. Next, I am not clear on the settings in Integrated Periphs. What is the setting for a single Sata II drive? My onboard Sata/Ide control mode is set at IDE. For a sata drive should it be set at AHCI? My SATA AHCI mode is diabled. Should it be set to enabled for a SATA drive? Last question. Gigabyte includes a tool Easy Tune 5. Is this program any good for O/C? Does it run in the background or just make changes to your bios. Sorry for all the questions, but I am a green NUBEE:D . Thank for this wonderful guide. It sure is fun. Larry
PS What is the advantage of AHCI over IDE? Thanks, Larry
 
llbts1 said:
Here goes. 1. What is the purpose of lowering your multiplier from 9 to 8? With a higher FSB you will get higher speed? With a lower multiplier, you have to raise FSB that much higher to get to same speed. What are the advantages of changing from 9 to 8? What am I missing? I am currently at 3ghz with a 9 x 333, 1.375v. 31c on Hyper XT cooling on a e-6600, Giga rev. 3.3. Corsair XMS 4-4-4-12 memory.
Click to expand...

I have an answer to your first question. The point of running with a lower multiplier is the higher FSB. 8x375 is faster than 9x333 because the motherboard/RAM are running faster.

Right now I am running my e4300 at 9x358 but I have been trying to run at 8x400 because it would be faster. However, my S3 motherboard won't get to 400MHz :(
 
Thanks Ken for that answer. I keep getting different readings on temps from different programs. How do you know which one is correct? Speed of chip is different than what I have set up in bios also. This can be very confusing. By the way, from my bios settings I am running at 3200. From core temp, I am running at 3600???:confused: Thanks, Larry
 
Thanks for this amazing thread but does anyone have a guide on how to set up RAID. I'm a total noob and have no clue how to do it.
 
Why does Orthos give an error sometimes, and then without any adjustment in bios or anything it will run without error seconds later? This causes me to question its reliability. Tried to switch over to ACHI. No go here. The computer just gets caught up in a loop trying to restart. I am running at 3.2 right now and stable. (except with the Orthos, then its Wheel of Fortune. Thank You for your time, Larry
 
KD28 said:
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 4 minutes 50 seconds - 1 errors, 0 warnings.
Execution halted.

Test 1, 4000 Lucas-Lehmer iterations of M18074369 using 1024K FFT length.
Test 2, 4500 Lucas-Lehmer iterations of M17432577 using 1024K FFT length.
Test 3, 4500 Lucas-Lehmer iterations of M17432575 using 1024K FFT length.
Torture Test ran 4 minutes 50 seconds - 0 errors, 0 warnings.
Execution halted.
Click to expand...

Specifically "FATAL ERROR: Rounding was 0.5, expected less than 0.4" is a memory error. Need to adjust some setting for your ram.
 
petern said:
Thanks for this amazing thread but does anyone have a guide on how to set up RAID. I'm a total noob and have no clue how to do it.
Click to expand...


Ok I will bite, but please tell me what kind of raid you wish to use first please so I can go right to the point.

If you did not install the Raid/AHCI drivers (called the F6 preinstall drivers on Gigabytes site) during windows install we might be in for a bumpy ride but beleive we can still do it.


Also be aware that to properly do raid you should use drives designed for raid although its not absolutlely necessary i.e WD RE series of drives instead of the SE series but I digress. Be aware that a 2 drive raid 0 with nothing else doubles your chance of loosing ALL YOUR DATA due to a drive failure.

Tell me your OS, and the desired raid configuration,
 
Thank you soooooooooo much for this thread. This is so amazing for noobs like me who have never overclocked before! Will you be adding more to this thread is is this all that I will have to know?
Click to expand...

I hope to start working on it again, but a bit slower, finish section on taking the memory OC info and using it to overclock the CPU in a rational method. ..Raid setup. Then basic updates and repost all in one bunch in proper order. LOL I dont know if this is all you have to know cause I dont know if its is all I have to know. You know ?
 
BillParrish said:
Ok I will bite, but please tell me what kind of raid you wish to use first please so I can go right to the point.

If you did not install the Raid/AHCI drivers (called the F6 preinstall drivers on Gigabytes site) during windows install we might be in for a bumpy ride but beleive we can still do it.


Also be aware that to properly do raid you should use drives designed for raid although its not absolutlely necessary i.e WD RE series of drives instead of the SE series but I digress. Be aware that a 2 drive raid 0 with nothing else doubles your chance of loosing ALL YOUR DATA due to a drive failure.

Tell me your OS, and the desired raid configuration,
Click to expand...

Yes! Yes! Yes! :D
Thanks a lot. I'm thinking raid 0 and in case you needed to know, I'm planning to run vista. Thanks again!
 
BillParrish said:
I hope to start working on it again, but a bit slower, finish section on taking the memory OC info and using it to overclock the CPU in a rational method. ..Raid setup. Then basic updates and repost all in one bunch in proper order. LOL I dont know if this is all you have to know cause I dont know if its is all I have to know. You know ?
Click to expand...

What??? Just do you thing I guess since I won't really know what it is anyways.
 
petern said:
Yes! Yes! Yes! :D
Thanks a lot. I'm thinking raid 0 and in case you needed to know, I'm planning to run vista. Thanks again!
Click to expand...

okie dokie, give me a day or two to write it up but do be sure to load the F6 preinstall driver found on gigabytes web site. Actually if you do that you will be in great shape its pretty easy as there is a menu when you boot that allows you to add drives to make the arrary etc. its all in the manual but I will do a quick quide.
 
Awesome guide bill, haven't OC'd since 4 years ago when I built my last rig w/a p4 2.4c. This guide helped me get right back into things. So far I've got my e6600 stable @ 3.267 (363x9), DS3 rev1.3(stupid mistake), zalman cnps9500, OCZ 8500 5-5-5-15. Average temp. under load is 60, has jumped as high as 62. I think I can do even better but I'm gonna back it down a bit until I get a fan on NB. Thanks again bill.
 
Remember people, cool down that northbridge with a 40mm fan or replace the sotck NB HS with a thermalright HR-05 like I did. Especially if you're trying to overclock to 400+
 
BillParrish said:
okie dokie, give me a day or two to write it up but do be sure to load the F6 preinstall driver found on gigabytes web site. Actually if you do that you will be in great shape its pretty easy as there is a menu when you boot that allows you to add drives to make the arrary etc. its all in the manual but I will do a quick quide.
Click to expand...

For latest Intel Inf, Intel Matrix Raid, and JMicron drivers, check:
http://www.station-drivers.com
The drivers offered on Gigabyte's site are usually outdated, in comparison.
Setting up raid:
Preparation: Use WinRar to extract the contents of the JMB36X_WinDrv_WHQL_R1.17.17.exe file
downloaded from:
http://www.station-drivers.com/page/jmicron.htm
Once extracted, you find a "Floppy32" folder.
Copy the contents of that folder to a floppy disk.
F6 floppy should look similar to this:
[Copy and paste in browser window:
http://pics.bbzzdd.com/users/vailr/2007-04-12_161410.png ]
Use when installing WinXP SP2 (F6 floppy driver mode).
Install WinXP SP2 on raided drives, using F6 floppy install for the JMicron Raid driver.
 
Really nice guide. Got my QX6700 upto 3.4G (425@8) right now. Million thx. :)

The board I got is DQ6-Rev3.3.

I already put a 40mm fan on the MCH heatsink. Quite a bit of pliers work.

I noticed one thing. According to EasyTune 5 and other monitoring software like CPU-Z and Everest Ultimate. The CPU VCore dropped from 1.3875 (this is the voltage I set in the BIOS) to 1.33 ~ 1.34 during 4 instance of prime95 stress test. Is it normal to have this kind of voltage drop or it's just a bug needed to be fixed?

In addition, what MCH & FSB voltage do people with this board usually set to to reach 500+ FSB? What's the actual default voltage of the MCH & FSB on this board?

WC Setup on CPU only.

CPU Temp: 23 @ Idle, 40 @ load
Motherboard Temp: 40 @ both Idle and load

CPU VCore: 1.3875
MCH Voltage: +0.15
FSB Voltage: +0.15
Ram Voltage: +0.4
PCI-E Voltage: +0.1 (not overclocking my video card ATM)

Thx in advance for any advices. :)
 
BillParrish said:
I suggest replacing the push-pins with 6-32 nylon screws washers and nuts found at Lowe's in the specialty hardware drawers.
Click to expand...
Just a note here: I got some 6-32 nylon screws from Home Depot for my DS3 (2.0) yesterday. They did not fit through my holes. I got 4-40's instead and they work like a charm. If you cannot find #4 nylon washers, you can use the #6 ones on the #4 screws.

Thanks for this excellent guide, Bill.
 
drizzt81 said:
Just a note here: I got some 6-32 nylon screws from Home Depot for my DS3 (2.0) yesterday. They did not fit through my holes. I got 4-40's instead and they work like a charm. If you cannot find #4 nylon washers, you can use the #6 ones on the #4 screws.

Thanks for this excellent guide, Bill.
Click to expand...


gah I will fix that. thanks !!!!!
 
I dont Post. i hear one beep, but i get nothing on the screen, i clear the CMOS, still nothing, I detach all usb, and use 1 stick of ram, still nothing. I just get 1 beep. any ideas? I change vid cards, still nothing on monitor, remove everything, detached power, reseated CPU, clear CMOS, keep just 1 stick of ram, try again.. still nothing. just one beep, then after a minute a single beep, fan spins faster hd cycles, but still nothing on screen- help guys? demmit, i should stayed put at 400x8, i was trying to go 400x9. when this happened.
 
Try to test your RAM in a different PC if you can. Also, reassemble the PC outside of the case to rule out any shortage problem. Also, test the power supply, video card and hard drive in another PC if you can.

BTW, what were your temps like at 400 x 8? What HSF were you using?
 
Danny Bui said:
BTW, what were your temps like at 400 x 8? What HSF were you using?
Click to expand...

thanks I'll try that

oh my temps were like 32-34 idle, 42-45 loaded, in a Thermalright SI-120 + 1700 rpm 70 CFM fan, if i used Panaflo, 2500RPM, 104 CFM fan - i drop to 40 loaded.

..... now if i can only get this thing to post again...9x400 couldnt have fried it you think?:eek:
 
Possibly. You may have gotten an E6600 that could only do 3.2Ghz.
 
I am curious is this guide good to go by for a DFI P965-S MB? I am thinking of trying some Overclocking with my new E6600 once I break it in.
 
FragMan said:
I dont Post. i hear one beep, but i get nothing on the screen, i clear the CMOS, still nothing, I detach all usb, and use 1 stick of ram, still nothing. I just get 1 beep. any ideas? I change vid cards, still nothing on monitor, remove everything, detached power, reseated CPU, clear CMOS, keep just 1 stick of ram, try again.. still nothing. just one beep, then after a minute a single beep, fan spins faster hd cycles, but still nothing on screen- help guys? demmit, i should stayed put at 400x8, i was trying to go 400x9. when this happened.
Click to expand...

hmm the 1 Beep is good actually, my best guess, and thats all it is, a guess, is that the cpu cant do 9x400, it is posting and then rebooting trying to reset itself to defaults but not being able to for some reason. Clearing the cmos should have got you out of the problem. But it didnt !!! :eek: The only thing I can suggest is try clearing the cmos again, even take the battery out and when done, let it cycle at least 5 times. Pray, if it does not come back you might have to RMA it. If it has a floppy you might see if the bios is scrambled and feed it a floppy with a bios file on it in hopes the boot block loader will sense the floppy and reload a bios.

Darn things are susposed to recover from failed OC attempts however it seems a small percentage dont and go ga-ga (highy technical computer science term), best of luck.


General note.

The procedure for doing a Raid 0 array is outlined step by step in the owners manual and the procedure is the same for Vista, absolutely a waste of my time to recreate it here. Please refer to that little book that came with the motherboard.
 
ServerKing said:
I am curious is this guide good to go by for a DFI P965-S MB? I am thinking of trying some Overclocking with my new E6600 once I break it in.
Click to expand...


In general yes, however there could be major differences in the layout of the bios in particular how the memory dividers are labeled. I dont have one of those boards. The general advice etc. should still apply.
 
Thanks for the info, I am farely new to OCing and this guide looked extremely good to follow.
 
BillParrish said:
hmm the 1 Beep is good actually, my best guess, and thats all it is, a guess, is that the cpu cant do 9x400, it is posting and then rebooting trying to reset itself to defaults but not being able to for some reason. Clearing the cmos should have got you out of the problem. But it didnt !!! :eek: The only thing I can suggest is try clearing the cmos again, even take the battery out and when done, let it cycle at least 5 times. Pray, if it does not come back you might have to RMA it. If it has a floppy you might see if the bios is scrambled and feed it a floppy with a bios file on it in hopes the boot block loader will sense the floppy and reload a bios.
Click to expand...
Thanks Bill and Danny and everyone for the input, overclocking is a journey and a fun experiment. :rolleyes: I am wondering though if my no POST is either, mobo, ram, or cpu, I'll find parts to check..

my instincts tell me it may not be cpu- I'm sure its takes more than what Ive done to fry it. ( 400x8 to 400x9 min volts)

Ram? hmm i have f10 bios,patriots 2.1s, 5-5-5-15, nah.. it worked first time out .

I'm leaning toward the mobo ..rats!:mad: I'll try the remove battery and floppy disk trick.

thanks everyone.

PS. I love this thread - love to hear the victories. nice to know youre all out there for the defeats ( flubs) too. cheers ;)
 
My E6600 running at stock speed idles at 53 C. I'm using a Zalman 9500 with normal fan speed. Why so hot? Installing the heatsink seemed straightforward. If I overclock to 3.0 Ghz it idles at 63 C.
 
You must log in or register to reply here.
Back
Top