New version OCCT stability test

[003]

Thanks

I'll try to give a shot at finding dead pixels in fullscreen mode. Maybe it'll :
1- make the driver switch to 3d mode
2- use all the GPUs

However, it'll take me some time to achieve, as i'm deep into the new user interface of 3.0.0 as of now, and i don't want to leave that until finished. Be sure i'll be posting here once i do it to see how your config behaves

Yeah that sounds great! And you can still use a low resolution in full screen, the same as it is windowed.

[TunedFish]

I'm using OCCT GPU v0.41 and was wondering what "Coef:xx" means.

[stasio]

OCCT 2.0.1-13.10.2008



.EXE FILE
.ZIP FILE

[emoners]

thanks stasio for the update!

[Tetedeiench]

I'm using OCCT GPU v0.41 and was wondering what "Coef:xx" means.

It's the complexity. The higher the coef, the higher the complexity. I use this parameter to adjust the frames calculated per second, as no known CPU can check 200 frames per second to see if there's any error. That would mean checking 200*512*512 pixels, which is truly... enormous. I had to find a way around that issue.

[Tetedeiench]

OCCT 2.0.1-13.10.2008



.EXE FILE
.ZIP FILE

Faster than me

[gymenii]

Thanks Tetedeiench!
Keep up the great work

[Tetedeiench]

Thanks Tetedeiench!
Keep up the great work

^^;

As i said, i'm working on 3.0.0. I spent 8 hours on it today, mostly on the interface. I can tell you what OCCT PT 3.0.0 will add :
* 2 CPU tests : the regular OCCT and one based on Intel's Linpack libraries
* 1 GPU test (the one you currently know) with error checking (and yes, i'll try to do that in fullscreen mode)... or pure burn
* 1 GPU memory test written in CUDA (similar to MemTest86, but for Graphic cards)
* 1 Power Supply test (which is OCCT CPU Test + GPU test)
* New interface (will have some common traits with the old one though, but it had to be modified to comply with all the tests i'm introducing recently

I've already worked about 3 months on that (GPU test and CUDA took a long, long time), i expect i still have about 2 months of work, but it should be ready soon enough

[jmke]

very much appreciate your work Tetedeiench, looking forward to Linpack integration
the GPU tests are very interesting too, a load test ALA CPU will definitely be put to good use by many

[cantankerous]

This program is sounding like a beast! Definitely have some amazing additions lined up. Thanks soo much for all your time and effort in coding this amazing app for us. I have been using it for years, and though I am done clocking my current system I look forward to giving the new version a try when it's released.

[Think]

using OCCT 2.1.0.b05 on vista x64 Q9650 and I'm getting an error after 1minute duo to a ripple in voltage drops of barely .5volts. Crashed Code 1

[Tetedeiench]

using OCCT 2.1.0.b05 on vista x64 Q9650 and I'm getting an error after 1minute duo to a ripple in voltage drops of barely .5volts. Crashed Code 1

Sorry for the long reply delay.

The Crash code 1 is always accompanied by a code 4 somewhere. Code 1 just means that the thread stopped because another thread reported an error. This is due to the behaviour of OCCT i wanted : one error detected on a core = we stop everything. Code 4 means an error was detected, that the result computed was different than the one expected.

It should be much more explicit on 3.0.0.

I have to say i'm really busy these days with 3.0, and that i'm getting closer every day to release. There's still some way to go before beta1 though. I'd say it should be out within a month.

[Praz]

I have to say i'm really busy these days with 3.0, and that i'm getting closer every day to release. There's still some way to go before beta1 though. I'd say it should be out within a month.

Looking forward to it.

[dengyong]

I have to say i'm really busy these days with 3.0, and that i'm getting closer every day to release. There's still some way to go before beta1 though. I'd say it should be out within a month.

In all versions so far,5v and 12v are incorrect when using the evga 790i ftw.

Great test, thanks for all your hard work.

[RejZoR]

This latest thing is broken. CPU usage percentage is all borked up (just look the image). And also doesn't seem to kick CPU to full speed. Instead it's "idling" in power saving state (1,8GHz instead at full 3,7GHz).

[Tetedeiench]

This latest thing is broken. CPU usage percentage is all borked up (just look the image). And also doesn't seem to kick CPU to full speed. Instead it's "idling" in power saving state (1,8GHz instead at full 3,7GHz).

That's because you enabled Power saving features, that reduce your CPU speed when it's idling.

OCCT displays the real frequency of your CPU. You should see it kick up once the test is started, and the "real" one is recorded in the graphs

Everybody else > Thanks for your cheers

[RejZoR]

Erm, no. Last version 2.0.0a kicked in from power saving state to full throttle just fine. This one just stays in power saving mode even after the testing actually starts.
Default clock is for processors original clock, CPU clock is the actual current clock which always reflects the current clock, even when changing power saving states.
I even checked the CPU cycles with perfmonitor and it was showing 3750 cycles (menaing 3750MHz) while OCCT was showing 1800MHz clock (like on image above).

[Tetedeiench]

Erm, no. Last version 2.0.0a kicked in from power saving state to full throttle just fine. This one just stays in power saving mode even after the testing actually starts.
Default clock is for processors original clock, CPU clock is the actual current clock which always reflects the current clock, even when changing power saving states.
I even checked the CPU cycles with perfmonitor and it was showing 3750 cycles (menaing 3750MHz) while OCCT was showing 1800MHz clock (like on image above).

Erm, true, my bad.

It seems the test didn't start though. The CPU usage % is just not right as well. As the test isn't started, the frequency doesn't jump up - there's no need for that.

I'll look into it ASAP.

[philbrown23]

isnt there an occt gpu test also?

[Tetedeiench]

isnt there an occt gpu test also?

Yes, and it will be included in v3.0.0 .

Here's a screenshot of the Beta version of OCCT 3.0 - it's still under development, but the general look is there :



The next version will include all those tests : OCCT CPU (classic), Linpack (same as IntelBurnTest), OCCT GPU, OCCT memtest (based on CUDA), and a Power supply test (which is really OCCT CPU + GPU at the moment, i have to look into Linpack + OCCT GPU).

Of course, all those tests come with the features of OCCT (graphs, custom tests, auto tests, and such).

I'm even thinking about a few new features that i'm pretty sure everybody will like

[dengyong]

Yes, and it will be included in v3.0.0 .

Here's a screenshot of the Beta version of OCCT 3.0 - it's still under development, but the general look is there :

I'm even thinking about a few new features that i'm pretty sure everybody will like

Will the new version support the evga 790i FTW ?

[gymenii]

Yes, and it will be included in v3.0.0 .

Here's a screenshot of the Beta version of OCCT 3.0 - it's still under development, but the general look is there :

I'm even thinking about a few new features that i'm pretty sure everybody will like

I like it already, looks amazing!

[DJSUB]

can't wait to have the version 3
keep up the good work Tetedeiench

[Tetedeiench]

Will the new version support the evga 790i FTW ?

Most likely

The voltages are screwed up on the screen on 3.0 because the ini file i used is for test purposes Don't worry, this screenshot is about 2 weeks old, and it's working flawlessly.

For 2.0.1, as on your screen, you can always tune up the voltages in the options, and select the good sensors if they aren't detected automatically, or screwed up.

i'll post b01 once it's done here. I expect t to be ready in a week or two...

[pest]

The CUDAMemTest isn't really usefull with my 8800GT,
it completes the Test at 1170MHz, this is barely stable for a 3DMark, let alone gaming, perhaps more passes (infinite loop) will help

anyway thanks for OCCT