OCing ram

[Mindflux]

I have a pair of cmx1024-3200 Corsairs and I'm trying to run them at 225 right now. It occured to me however that there aren't any temp sensors or anything on the ram, how can I tell if I'm going to burn them out? I s'pose I'll get errors first...Basically, is it a bad idea to OC the **** out of my ram or is it reasonably safe so long as I don't do something stupid?

[Gipse]

What kind of voltage are you using for your Ram? Anything below 2.9v is OK without active cooling (in my opinion!). You can check the temperature by touching the modules (be carefull) while running memtest for some time.

Gipse

[Mindflux]

I'm pretty sure it's at 2.6, and I have case-fan that's set next to them (slightly forward of them)

[Gipse]

You should be safe with that setup

Gipse

[smolt]

you can hit DDR500 with that ram at 3-4-4-8

[Kaxx]

What does your smartguardian say your DRAM voltage is??

[Mindflux]

I don't use SG it doesn't seem to give me accurate readings and likes to turn my cpu fan off.

in the bios it's set to 2.6 and reads at 2.69

 

I passed 6 passes of memtest @ 3-3-3-7 in 9/10 (225mhz) but when I get into window it resets within a minute of running...I take this to mean I should loosen the timings? or is it just failing completely at that speed?

 

edit: I also just noticed that Trc is set to 7 rather than auto (like everything else)...is this my problem? I don't remember changing that but it says "7 is usually much too tight for most average ram" in the OC guide...I'll set that to auto and see if it runs I guess

[osen]

Never heard of anyone killing their RAM due to heat actually. It will most likely generate errors long before that. And at those voltages you don't have to worry. The MHz doesn't really add to the heat output that much.

 

Actually never heard of anybody killing their RAM with to high voltage either :S My oldschool BH5 for example doesn't like voltages over 3.6V but it hasn't killed them either.

Same thing with some TCCD-sticks I experimented with (different voltages though ). If somebody knows anything about this, please share it.

 

Right now I'm at 3.4V without active cooling and it has worked like a charm for over a year. I attached a temp probe to it and it's about 50C at load.

[michaelzhao]

I run TCCD Ram at DDR 560@ a 1T command rate (3-3-3-8) with voltages of 2.8 and it absolutely fine.

 

The temperatures you'll need to kill your RAM are far higher than the voltages you need to get there. Even if I run 2.9volts (max through TCCD) its still much lower than the volts needed to get extremely high temps. And by that time, my voltages will kill my RAM before my temps do.

[Mindflux]

my settings are as follows:

CPC=auto=1t

Tcl=auto=3

Trcd=auto=3

Tras=auto=7

Trp=auto=3

Trc=7

Trfc=auto=?

Trrd=2

Twr=2

Twtr=2

Trwt=3

Tref=3120

Bank interleave=enabled

DQS skew control = auto = ?

DQS skew value = 0

Dram Drive Strength = auto = ?

data drive strength = auto = ?

max async latency = auto = ?

Dram response time = normal

read preamble time = auto = ?

IdleCycle limit=256

r/w queue bypass = 16x

bypass max = 7x

32bit granularity = disabled

 

is my ram failing at this speed or are my timings messed up?

[lowboy]

Try trc of 10....... trc=tras+trp

[Mindflux]

failed at 3-3-7-3-10

failed at 3-3-8-3-11

 

I'll try 3-4-8-4-11 later...