• 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.

GA-965P DS3 Issues

CaptRingold

Limp Gawd
Joined
Feb 12, 2005
Messages
256
I got together this new system a couple days ago, and after much tinkering, I seem to have some bum RAM.

Relevant parts:
E6600
4x2 1GB SuperTalent 4-4-3-8 (though I've loosend it to 5-5-5-15 till stability is established) @DDR800
Rev 2.2 with F10 BIOS
Watercooled, Corsair Nautilus 500 (side note: loud pump!)

I first tried to jump right in with the OC I wanted, 9x400, and thought I'd try to creep up the voltage to make it stable. I eventually realized that was going to be impossible.

Based on how Orthos works, it actually seems like the CPU itself is stable here, so then I went to Memtest86+ v2.7.

I lowered it to 6x400 to simulate stock CPU frequency but the higher FSB, and had the RAM 1:1 so it shouldn't of been OC'd. Memtest bled errors before freezing.

I play with voltages, +.1, +.2, +.3, +.4, +.5, some of which make Memtest bleed less, but never close to stable. I take the FSB back to stock 266 with the proper divider to get back to DDR800, same result. I try to underclock it, same result. I try +.1 to +.3v on the FSB and MCH, all to no avail.

I've got the C1E and whatnot disabled, but thats the only other major change. I tried all sorts of vcores up to 1.6 before realizing the CPU was happy as a lark and the errors were getting thrown by RAM.

Is there a problem having all 4 DIMM's populated in this board, or am I missing something? Or is my next recourse to try each individual pair of sticks to see if one pair is stable and the other not?

Edit: I know there's a similar thread going, but with a different problem I didn't want to hijack it with my own issues.
 
try 401 FSB, it jumps to next FSB strap which is less stressfull on the ram & northbridge.

9x multiplier (ie: 3.6Ghz) demands a pretty good chip.
try the 8x multiplier, a lot more chips can do 3.2Ghz with out too much voltage / effort.
 
install them in pairs (only 2 sticks at a time) in the primary slots and if they still will not pass memtest at rated FSB and timings, send em back.

Swap thru them and see if you can find one good pair.

edit: I can see nothing you are doing incorrectly, lol tho you might want to take it easy, zapping them with +.5 was a little radical. If you do find a good pair, I suggest you continue your memtest experiment and determine their max clock at stock timings and on "auto" . You might drop down into the memory sub-forum and see if anyone knows anymore about those chips. Some memory responds well to extra voltage, some does not and some "non-standard" timings can be usefull as well.

edit 2; I agree with chris, establishing a 8x400 or 401 baseline OC is a good practice and then go from there.
 
Thanks guys, great advice.

After testing all the RAM individually, they all checked out -- even when I OC'd them.

So I used what you guys said about bumping it up just past 400mhz; I went with 403mhz, and I could tell that helped, but not 100%. +.1v vdimm helped more, and rock-solid RAM at +.2v. With all 4 dimms populated though I did have to loosen a bit from the 4-4-3-8; I went all the way back to 5-5-5-15; I think I can make it tighter but that's a battle for another day.

8x403 is rock-solid, and just tweaking vcore to get 9x403. But if I have to settle for 8x403, well, battle well fought anyway. :)

Again, thanks guys! Never would've guessed about the FSB bands; thats really what was killing it. The timings and vdimm I would've figured out eventually, but lol, raising FSB to stabilize an OC wasn't my first thought.
 
Hey Chris,

Good to see you're still rockin' the boards. You helped me out about a year ago with my Pentium D overclock adventures and mobo selection. Still cranking my little 920 at 4.0Ghz. Updated to a 640MB 8800GTS and it's still a very capable system. I see you updated your box to Shiznit 3. My brother is shopping for an E4300 mobo. Based on the conversations I see here and in other threads I'm going to recommend the DS3. He's not going for monster clocks, just wants a good solid board running 2.4-2.7ghz with stock heatsink.
 
but lol, raising FSB to stabilize an OC wasn't my first thought.

heh yeah, I think most mobos have holes like that. something to do with that chipset strap if i recall correctly.

if i wer you though i would have tested each stick individually, at stock speeds first before i began OCing. I made that mistake with some of my old sticks and could never find out what was wrong with my system.
 
Everyone seems to be able to get a FSB of 400 or above with this board, and I've seen reviews get over 500.

With my processor multiplier at 6x so it's clocked well below what I've had it at and RAM multiplier at the minimum so it's running below its rated speed I can't get the FSB on my DS3 rev3.3 F10 above 365. I've disabled pretty much everything and increased the FSB voltage but doing so did not help the FSB go above 365 :(
 
Everyone seems to be able to get a FSB of 400 or above with this board, and I've seen reviews get over 500.

With my processor multiplier at 6x so it's clocked well below what I've had it at and RAM multiplier at the minimum so it's running below its rated speed I can't get the FSB on my DS3 rev3.3 F10 above 365. I've disabled pretty much everything and increased the FSB voltage but doing so did not help the FSB go above 365 :(

Have you tried the F7 bios?? The f10 bios killed my overclock on the rev 1 board.
 
I have the rev 1 board that I bought in October of 2006 and the f7 bios seems to work the best for my mild overclock. It is also the first bios that supports quad-core processors, which is what I want to upgrade to eventually. I just know that I personally had problems with the f10. It wouldn't overclock with that bios at my current settings.
 
Hey Chris,

Good to see you're still rockin' the boards. You helped me out about a year ago with my Pentium D overclock adventures and mobo selection. Still cranking my little 920 at 4.0Ghz. Updated to a 640MB 8800GTS and it's still a very capable system. I see you updated your box to Shiznit 3. My brother is shopping for an E4300 mobo. Based on the conversations I see here and in other threads I'm going to recommend the DS3. He's not going for monster clocks, just wants a good solid board running 2.4-2.7ghz with stock heatsink.

Its been a solid board for me. I've had a Rev 1 board since like August.
The first few BIOS's were not great, but F7+ are all very solid.

After F7 I could drop to 6x multiplier and the board would go as high as 473 FSB. I've also run my ram at 3x multiplier at 333 = 1Ghz ram clock.

The best part of the DS3 is the huge user base. Look at the Conroe / Core 2 overclocking database. And look at how many people have the DS3 (78 currently - way more than any other). That means great information available from other users. (other people with similar configs, etc) And Gigabyte will continue to support it longer with BIOS updates.

summary image from OC'ing DB

Technically nearly any mobo should work for a 4300 OC, b/c of the lower FSB starting point (and 9x multiplier) so you could go really cheap. Max you would need MIGHT be 400 FSB for 3.6Ghz (100% OC!), if he was really lucky. But the DS3 is still already fairly inexpensive and its results have been as good or better than many of the $200++ mobos.

Anyway, good luck, post back some results.
 
Back
Top