Help!!! Venice 300+ Failing OCCT

[umper22]

Ok--I hope I have obeyed all rules...this is my first build and 2nd post to this forum. I started building this system two weeks ago and have lived on these forums to get the most out of overclocking. I am not sure what bios I started with (ordered board from newegg a month ago) but at one point I was able to overclock my venice 3000+ at 2.6 ghz @ 1.47 volts (ldt=1.4/chipset=1.7). My PWMIC and Chipset temps never got above 55 degrees and my cpu temp never above 45. I was having problems with memory and posted on ocz board. They recommended bigtoe bios and new settings. I downloaded bios using iso image and installed. Ran memory @ 200Mhz with 2-2-2-8 timings and passed memtest. I proceeded with overclocking venice but I am now failing occt within a minute. I went back into bios and loaded optimized settings but still fail @ 1.8 ghz. Any ideas of how to isolate this problem? I'm guessing its processor but I want to be sure its not the board.

[lowboy]

My vote is memory just cause you can pass memtest does not mean your stable.

[CPDMF]

Put the memory on a divider to isolate it from the CPU at stock speeds. Make sure you are giving the Ram enough voltage as well.

[Guest Blooz1]

OCCT is rough on both memory and CPU. Are you getting the "Processor Not Stable" error?

 

If so, you'll probably need more voltage, as CPDMF said. I usually run my rig at 1.52/1.55V, but to pass OCCT I have to crank it up to at least 1.6V!

[KingYubbo]

OCCT is very CPU based.. memory isn't usually a factor unless you change the memory usage in the settings. I run my Venice @ 1.525 24/7 but it needs something like 1.62V to pass OCCT. Fine for Games. Try StressCPU... i used that and it's game stable. If it's not StressCPU stable, it's not Game Stable for me!

[umper22]

Thanks for all the posts. I am still confused, one says OCCT uses memory and antoher says it doesn't. I have ram on max divider as described in overclocking section of this board. For some reason I think cranking up voltage is the answer, because I was able to pass OCCT before when voltage was cranked. I'll give it a try tonight. My concern was that I can't even run at stock speeds, but my system is stable for simple things. I can also run 3DMark05.

[KingYubbo]

OCCT is CPU.. and hardly any RAM. Where ever you heard the RAM thing from, it's wrong. SuperPI is more memory than CPU.

[Guest Blooz1]

That's interesting, KingYubbo - did you read some info on OCCT, or is that from observation?

 

I usually use both OCCT and SuperPi 32M before running Prime or a variant "long distance".

 

Umper - by "max divider" do you mean 1:1?

[umper22]

Blooz--I am seeting Ram ratio to 100(Mhz)(1/02) setting.

[umper22]

This is just a follow up for anyone who runs into same problem as I did. I was running OCCT at stock speed and it kept failing around 6 percent thru occt. My problem was that I had disabled "Data Execution Prevention" in windows. To enable this, right click on my computer, goto advance and click on settings under performance. Make sure you enable "Turn on DEP for essential Windows programs and services only". The only reason i found this was b/c of a printer problem