1 (edited by Gregole 03-04-2012 09:44:39)

Topic: [FIXED] GTX680: Problems with SVPmark

It's not really that big of a problem but I was wondering,  why does my Real-Time test shows up as N/A?  Is it because I changed my videocard or something else?

Post's attachments

svpmark_0304_0238.log 1.16 kb, 543 downloads since 2012-04-03 

Re: [FIXED] GTX680: Problems with SVPmark

Gregole
May be it is protection from very fast results big_smile

CPU:  Intel Core i7-2600K @3400 MHz [8 threads]
GPU:  NVIDIA GeForce GTX 680 [ver.301.10]

Re: [FIXED] GTX680: Problems with SVPmark

This's obviously a bug  hmm
Are the failed tests list changing from run to run?

4 (edited by Gregole 03-04-2012 16:41:43)

Re: [FIXED] GTX680: Problems with SVPmark

The test that always fail are:
48 fps - vectors search
48 fps - [SVP] fastest
48 fps - [SVP] simple 1
60 fps - [SVP] good
72 fps - [SVP] simple 2

MAG79 wrote:

Gregole
May be it is protection from very fast results big_smile

CPU:  Intel Core i7-2600K @3400 MHz [8 threads]
GPU:  NVIDIA GeForce GTX 680 [ver.301.10]

Weird. . . its showing my cpu as 3.4ghz even though I overclocked it to 4.8ghz.

Re: [FIXED] GTX680: Problems with SVPmark

Noone can read overclocked Sandy clocks right without kernel-mode driver and admin rights sad

Re: [FIXED] GTX680: Problems with SVPmark

Now we know that it's a bug (or a "feature"  big_smile) either in 301.10 OpenCL compiler for GTX680 or in SVP code.
I hope the next WHQL release will show the answer.

Re: [FIXED] GTX680: Problems with SVPmark

Fixed in SVP pre-3.1.2 patch, SVPflow 1.0.5 and SVPmark 3.0.3a