Tetedeiench

I can't use anything else than OCCT, being who i am I have to format Vista to go back to XP (who said "hasta la vista" ?) tonight, i'll rerun it again for the screenshot

Why would it be a suicide ? Rock stable

So : C2D E4300 @ 3Ghz FSB 335 G.Skill 2x1GB PC6400HK @ 3-3-3-5-3-1T synced (@335Mhz) (factory timings : 4-4-3-5-24-2T@400Mhz) Asus P5N-E SLI 2x7900GTX (Club3d & Leadtek) Seasonic S12-500W Better late than never OCCT 8hours stable + 3hours memtest stable :)

I'll see to that later, but i don't think it'll make it live I'm really not convinced on how useful this would be... Especially Uber, tell me how you would measure this offset ? How would you know to use 5

It could be possible, yep. I had one request about that from a Croatian guy (i told you OCCT is realy wide spread). Problem is it is not OCCT's job to do so. OCCT just reports readings and graph them. It does not get values by itself, it's just reading them. This option should be in the monitoring softwares (Speedfan, Everest, etc), and not OCCT. Moreover, how would you specify such values ? it would be arbitrary, and temps graph would not really mean the very same thing if people could scale it up/down. As i do think, personally, that there are more cases where the temps reading are accurate rather than inaccurate, and as idon't know of people would calculate this offset other by arbitrary decision, i am choosing to rely on the monitoring readings, directly. It ensures that the graphs is made out of the values read by the monitoring software, and that people won't understimate their temp readings with a "-20

*grin*

Unfortunatly, i don't think that's going to happen... OCCT is already pretty reliable as it is, but having different algorithm can only be a good thing, and i can't see OCCT gathering them all As for the GPU, i'm still working on it

Expect new betas soon : i'll try to add Core Temp support this week, and correct minor quirks as well, preparing 1.1.1 release

I may as well post my modest overclock here C2D E4300 @ 3.0Ghz (9*335, don't ask me why it doesn't supports 333Mhz bus). Asus P5N-E SLI 2x1Gb G.Skill PC6400HK @ 3-3-3-5-3-1T synced (i.e. 375Mhz, to get max performance with reduced timings) Seasonic S12-500W 2x7900GTX (Sli) In terms of cooling : Zalman CNPS-9500 @7V ThermalRight HR-05-SLI with a Noiseblocker fan @5V(800rpm) on the northbridge 1x120mm fan extracting the hot air 12 hours OCCT stable

What is, in your opinion, more important : A memtest-like under windows in the memory left free by the OS A GPU test to load both CPU and GPU to stress the power supply much more (no error checking on the GPU) A programmable OCCT where you can build your custom tests (like "myowntest=1h CPU&RAM, 30minCPU, 2h RAM", with graph generation Thanks for your opinion

Almost ! (Tr

Some news : i could do a somewhat efficient GPU loading shader (no error checking, mind you), and it is loading my 7900GTX to 67

Some news about the 8-cores server : OCCT detects an error fast in "RAM" mode (below 10mins), between 15~1h in CPU&RAM mode, and runs flawlessly in CPU mode : The server is unstable, and of all the tests i know of (orthos, superpi, and such), only OCCT reported the server being unstable. But, as OCCT is running flawlessly in CPU mode, it *cannot* be a bug in OCCT (all OCCT modes share the same algorithm, only the data changes. As the CPU mode is running fine, it meas OCCT can handle 8-cores... and thus... OCCT is fine here). ==> OCCT PT 1.1.0 validated on a bi-quad-core And more efficient than anything else

Thanks I will try really hard The memtest option is almost there (probably 1.1.2 or something like that). As for the GPU, i'm looking for somebody with 3d programming language... and if i can't find anybody, i'll do it myself, but it'll take much more time, unfortunatly... i have a bit of knowledge in 3d programming, but nothing that goes beyond modeling a planetarium 4 years ago in OpenGL... That's an "of course"

It's just a server we received at work It's supposed to be for testing... they didn't think of those kind of tests, i'm sure An error was detected after 45min... Personally, i don't even have a Dual-Core CPU (but that's going to change soon).