DS3 3.3 won't run at 1100+ on memory speed

[zbomb5610]

I am running the 2gb G.Skill HZ kit, just got it in this monday.

I can run 1050 4-4-4-4 @ 2.3v easy, and i can do it a little bit higher, usually with a 1:1 ratio, but my cpu isn't that stable.

This is with many passes in memtest.

I decided to see how high the memory would go, changed the timings to 5-5-5-18 and raised the voltage.

I set the memory multiplier to 4 and tried to get the computer to run at 1120 memory speed, which was only 280 fsb, and it wouldn't do it.

I even bumped the voltage up to 2.5 but it freezes in memtest, doesn't show any errors, just stops running memtest and the percentages stop, but the little (+) sign by memtest in the top left is still blinking. No errors are displayed, just stops.

I then tried to do 1100 with memory multiplier of 3, but that did the same thing.

I am pretty sure it is the mobo that is keeping me from doing that, but not sure what else i would check.

I wanted to see how high the memory would go, i would hope it isn't the memory that is holding me back. I was hoping these D9's would go higher, assuming they are D9's since i really can't take off the heat spreaders to check since I don't want to void the warranty.

Any suggestions on what i should try.

[Lestat]

this is normal the DS3 has never been able to do 1000mhz +

its a design flaw and there is no fix. and after 3 revisions gigabyte pretty much refuses to fix it.

[zbomb5610]

this is normal the DS3 has never been able to do 1000mhz +

its a design flaw and there is no fix. and after 3 revisions gigabyte pretty much refuses to fix it.

I can do 1050 no problem, good to know it is most likely the mobo and not the memory, still sucks though since i wanted to see how well these HZ's would OC.

[Stapler]

this is normal the DS3 has never been able to do 1000mhz +

its a design flaw and there is no fix. and after 3 revisions gigabyte pretty much refuses to fix it.

Eh, that was true with my rev. 1 board, but my rev. 3.3 does near DDR1300. I've had a lot of success with the F9 BIOS, so you could try that. Then again if your ram needs 2.3v to get DDR1050 4-4-4-x the ram might have some issues.

[Hutch]

You need to set the vFSB to the max, as well as the vMCH.

[zbomb5610]

You need to set the vFSB to the max, as well as the vMCH.

i was currently running +.2 on vFSB and vMCH, but i moved those to max at +.3, Putting +.5v on the memory, i got 2000 errors within a few seconds in memtest, bios will freeze if i try to do +.6v. Very unstable with 2.4v+.

I have a feeling i got a set of HZ's that don't oc so well. Maybe ill exchange them for the ballistix 1000's at newegg.

[kiwi]

This board also needs more volts in comparison to p5b dlx (all DMM) to even try running above 1000 or so

[Hutch]

i was currently running +.2 on vFSB and vMCH, but i moved those to max at +.3, Putting +.5v on the memory, i got 2000 errors within a few seconds in memtest, bios will freeze if i try to do +.6v. Very unstable with 2.4v+.

I have a feeling i got a set of HZ's that don't oc so well. Maybe ill exchange them for the ballistix 1000's at newegg.

I can't even put +.2v to my ram without clocks degrading on my S3 rev 1 c1. Have to use +.1V vdimm for the best clocks on my ram (terrible ram, low grade 1 1/2 year old elpida's.) I have managed to hit 1000mhz with only +.1V on them though... Stable enough to run spi1m ;)

[ahmad]

You tried option 2 for memory dll right?

[zbomb5610]

You tried option 2 for memory dll right?

yep, played with both options, played with pretty much everything under the oc section of the bios

[zbomb5610]

so i am leaning toward it just being the memory, tried the HZ's in my cousin's p5b deluxe, and no luck getting the memory over 1080, even with 2.5v.

Wouldn't even boot at 1000 w/ 2.15v on the p5b.

[zbomb5610]

UPDATE:

So after not getting the results we wanted with the HZ's in my ds3 or my cousin's p5b deluxe, he started playing with his ballistix, he realized the 2.5 multiplier was holding him back. So he changed it to 3x multiplier on the memory and had his memory running 1200 with no problems.

We decided to try the hz's in his p5b and we were able to hit 1185 max stable with them @ 2.45v 5-5-5-18.

Tried playing with subtimings and all multipliers along with both options for the memory, and i can not hit 1100 on the memory without going into a horrible cycle of booting.

I am almost 100% sure it is the motherboard that will not allow me to run memory above 1100, his ballistix was also capped at about 1090 in my motherboard which further confirms it is most likely the mobo holding back the ocing of the memory.

This sucks and makes me want to swap out the mobo, but even then, i usually run a 1:1 multiplier so there really isn't much need to get rid of it. Just sucks that i can't test memory out on it.

[n91htmare]

I'm the cousin.

My Anniversary edition of ballistix has no problem doing DDR2-1200 at just 2.3V, but Same ram in Zbomb's DS3 does not even hit DDR2-1200 with 1.45V.

Maybe we have a wrong setting somewhere but where?

[zbomb5610]

Eh, that was true with my rev. 1 board, but my rev. 3.3 does near DDR1300. I've had a lot of success with the F9 BIOS, so you could try that. Then again if your ram needs 2.3v to get DDR1050 4-4-4-x the ram might have some issues.

just tried the f9 bios, still no luck breaking 1090. Either goes into a boot cycle or it just disables the oc.

[zbomb5610]

UPDATE: I tried to use setfsb, but i had better luck just using easy tune (program that is for the ds3). Again, 2.5 multiplier does not work whatsoever when trying to oc the memory, it won't even do 1000+ without having problems and freezing on me.

I changed to multiplier of 3, and i had some more luck.

Problem is that I can only boot at 1140 in bios sometimes, and sometimes I can’t. So I broke the 1080 barrier after tweaking with the subtimings in the bios. But I had a lot more luck with ocing the memory when I tweak with the subtimings in memset and use easy tune to bump up the fsb.

I changed the subtimings to what are shown in the first pic for the hz’s, and those subtimings were pretty much what my cousin’s p5b deluxe are set to. I was able to hit 1170 fsb and run super pi 1m, but I also hit 1190 fsb before the comp froze trying to run super pi. Seems 1170 on the hz’s was the most stable for the ds3 with 2.5v. But I really didn’t do much testing, pretty much trying to figure out if the ds3 is limiting the memory overclock or if it is the memory.

http://aycu09.webshots.com/image/12288/2000806478947689544_rs.jpg

I then ended up testing my cousin’s crucial, which also wouldn’t boot at over 1140 in bios with loose timings. Under easy tune though, with setting the subtimings pretty much to what the hz’s were at, I was able to hit 1260 and get off a round of 1m super pi. I hit 1272 but the comp froze with the crucial at 2.4v, I really didn’t put any more voltage into the crucial since they are my cousin’s and he would probably get butt hurt if he knew I did.

http://aycu39.webshots.com/image/11758/2002928387080807565_rs.jpg

So thanks hutch, I originally didn’t mess with multipliers and subtimings in windows much because I really wanted to get the memory to boot higher from the bios, but that just wasn’t happening. So you getting me to try setfsb made me play with a few other things in windows and I pretty much figured that I could hit higher memory overclocks if I played with them in windows, meaning my ds3 is capable of doing it, just not from bios.

THE DS3 WILL RUN THE MEMORY AT OVER 1250 EVEN, JUST NOT THROUGH BIOS. I AM PLEASED WITH MY DS3 SO FAR, JUST WISH I COULD DO THIS THROUGH BIOS.

Eh, that was true with my rev. 1 board, but my rev. 3.3 does near DDR1300. I've had a lot of success with the F9 BIOS, so you could try that. Then again if your ram needs 2.3v to get DDR1050 4-4-4-x the ram might have some issues.


are you doing near 1300 by booting up through bios? or is that once you are in windows.