How much does DDR voltage affect OC`ing...? (take a look)

[Bartman39]

I have this awesome "THUGSROOK" chip (1.6a @2655mhz /1.68 volts on an Epox 4G4A+)... But I`m running some OCZ PC3000 (yea I know but bought this before I found out about their BS) Well before I could only run at 160fsb (2.56ghz) and with slightly higher core voltage (1.71) but I increased the DDR memory voltage to 2.9 and now I can boot and even load Win9X @3Ghz...? Or run perfectly stable @2655mhz (166fsb)

The real question...? Will the 2.9 volts on the memory cause a big problem or will it be fine...? (yes it has the copper heat spreader and I have a nice quiet 120mm fan blowing on the HS/fan & memory)

BTW the cooler is an AVC Sunflower (damn nice cooler)

 

[THUGSROOK]

hiya Bartman39!

vdimm voltage is BS! (depending on the board)
run a program called memtest86 to find your true max speed/max cas settings.

boot to it and run "ALL" 11 tests for 1 pass. itll only take about 30mins.
post your results when done.

here's my results on a stick of XMS3000...
2.6v = 436mhz max
2.7v = not stable at any speed
2.9v = not stable at any speed

g'luck

 

[nemo160]

ddr voltage can occasionally give you that little bit extra out of your mem
i have two sticks of old (july '01) crucial pc2100 that would boot at 166 but not stable, boosted vdimm to 2.7 and rock solid on an 8k3a+
didn't try to push them past that, i was running 166/166 on my tbird 1.4 at 1.6-1.7
2.9 seems pretty high though and will probably shorten the lifespan on the memory...but its ocz so thats probablya kindness to the earth

 

[Duvie]

I agree...My epox is 2.6v default and I ran into a wall with the memory and when I bumped it to 2.7v it ran up quite a bit further. It isn't BS on my board...Anything after 2.7v is a waste though...


Check out my rig!!!

I booted into win98at 3ghz but I needed 1.79v and I didn't even test stability...I know it would have needed like 1.9v or higher to be stable at that speed and likley would have crapped out in any program....

 

[Bartman39]

Thanks Thugsrook... I checked it out and found my memory works fine up to about 400mhz after that voltage does not matter... But it runs rock solid with the increase of .4 volts...

Another question is this also the chipset voltage...?

 

[THUGSROOK]

....is this also the chipset voltage...?

actually im not really sure :Q
in your case i would suggest running @ 166fsb 1:1 ratio ~ see if it can run as low as cas2226 for ya.
that would be better/faster than 150fsb 400ddr.

HTH

 

[Bartman39]

Well I tried but it wont run at those settings so looks like I`ll just get me a stick of that good Corsair XMS 3200 Cas2 stuff... Can you recommend a good place for a 512meg stick...?

Thanks & this chip is just plain awesome...

 

[gururu]

2.9 volts seemes high to me, as far as cooking your mem chips is concerned.
i personally haven't seen any difference in playing with ddr voltage. my chip o/c to
406 MHz at 2.5 (2.46) volts. increasing the voltage hasn't helped get any faster.
it may go higher with voltage, if my cpu is actually the bottleneck...
i wouldn't get near 2.9v though :Q

 

[Bartman39]

Well strangely enough I could not get stability @2.5 volts on the memory but with 2.9 its rock solid and the mem-test program also reflects this... So possibly with this board its something that is needed...? BTW I checked the temp on the memory and its not even warm...

 

[THUGSROOK]

it could just be that board undervolts mem and mine overvolts....

 

[Duvie]

very much the possibilty...My epox undervolts everything so it makes since to me!!!

 

[Matt2]

How do u get results from that program? I ran it, went to sleep and woke up this morning and it was still going, like on pass 9 or something. I couldnt find how to get the results in the menu so I just hit esc and the comp rebooted.

 

[SupermanCK]

memtest86 will tell your if there are any error ... if no error it would say passed... so you should look fine with 9 pass...but i think thugsrook said you have to pass 11 tests... i don't remember exactly tho

 

[THUGSROOK]

hehe ~ yea that program has stupid terminology.....

by default that program runs the short 7 test "loop". (or pass)
you have to hit (C), (2), (3), (0) for it to loop all 11 tests.
completing all 11 tests = 1 pass. (the programmers shouldnt have used the term "pass" this way!)
you only really need to run it for 1 "pass", takes about 30mins.
if there are errors they will be listed on the screen.

HTH

 

[Mikki]

Voltage didn't help my XMS3200 either...

 

[Stallion]

So how exactly does this test work.. You just load it and let it run? And then it tells you if you Mem is being utilized to the maximum?

 

[Vehementi]

Higher voltage = less or NO stability
Higher voltage = Higher heat
Higher voltage = less life

 

[Bartman39]

Just cant agree with it...

Higher voltage = less or NO stability

I`m running even less core voltage (1.63-1.65) and still rock solid... BUT...! If I lower the memory voltage below 2.8 volts I get lockups and all kinds of crazy errors... And memtest86 shows this as well...


I do agree higher voltages mean higher temps but good-great cooling works wonders...

Also higher voltages are an OC`rs way of life... (we do what have to do to achive what we should`nt)


A 1.6a (1600mhz) that will run up to 3Ghz (3000mhz) or a rock stable 2.66Ghz is where its at...



BTW a last note I just ordered a stick of Corsair XMS 3200 C2...

 

[Bartman39]

Well just an update and some interesting findings...

I just did some testing with my new Corsair XMS 3200 C2 memory and find that with it ANY increase in the voltage results in errors even lockups in Win9X...? So of course I run it at default voltage and right now I`m at 415 on the DDR (4:5 ratio) Cas2 and the memory scores are sweet...


So what it comes down to is the memory NOT the motherboard... The old PC3000 OCZ memory that I was using (red PCB...?) needs alot more voltage to run at higher speeds but this XMS is cool and I mean cool...

 

[THUGSROOK]

kewl ~ nice info

 

[bigboxes]

Good to know! I am running Prime95 right now to burn my cpu in. I will run that memtest86 tomorrow morning on my XMS3200C2... can't wait!

~box