graphic question - notebook & computer

How to noticegraphic problems, besides overheating and exhaustive
tests of graphic card/chip ?

is there a possible way to notice such failures in some kind of
log,, like card producers log if they make them ???

or what event IDs would i have to search for to look for in the
event viewer ?
(are there any event IDs which are known as graphic card,
motherboard failure IDs which are sign of mailfunction of a working
device ?)


and what if the user has put a very new Operation system, then any
log or any event ID is empty as the system is fresh, what then ???



thanks.

comitter wrote:
> How to noticegraphic problems, besides overheating and exhaustive
> tests of graphic card/chip ?
>
> is there a possible way to notice such failures in some kind of
> log,, like card producers log if they make them ???
>
> or what event IDs would i have to search for to look for in the
> event viewer ?
> (are there any event IDs which are known as graphic card,
> motherboard failure IDs which are sign of mailfunction of a working
> device ?)
>
>
> and what if the user has put a very new Operation system, then any
> log or any event ID is empty as the system is fresh, what then ???
>
>
>
> thanks.

If the manufacturer would share their testing code with us,
we'd be able to test some things, like they do.

But as long as the manufacturer test details are unknown,
all we can do is "play 3D games" as a test, or use tools
like Furmark or ATI Tool. Furmark seems to exist, mainly
to generate heat, so may not be the best test. And Ati Tool
may not work with all operating systems.

http://files.extremeoverclocking.com/file.php?f=204

http://en.wikipedia.org/wiki/Atitool

http://www.techpowerup.com/atitool/

Using GPGPU code, CUDA or OpenCL, might be another way to do
testing, but then that will only be testing programmable
shaders, and not give good test coverage for other component
parts of the GPU.

In short, we don't have good tests. Only partial tests.

Paul

yeah, today you have a ton of instructions and new capabilities which
some programs do not handle,
and then you aren't sure so well was the program well efficient or
not....

Like i'm using VMT for testing RAM memory at Graphic cards... and you
newer know was the test handled as it should or haven't i forgot to
pick some extra functions... only program maker knows what it really
does including the manging elite and engineers in that stuff...

i don't buy it again M;) i just think of what kind of tests i do with
the graphic cards, 3-5 test programs which are the base, basics for
testing... i don't know about high-grade-gamer cards of this
days...that amount of huge memory compartment like 1-3GB GDDR5... i
really doubt this programs are good written for every kind of graphic
cards, yet there are numerous variations in options and functions when
starting a test...

yeah, i still haven't got my claw on that Cuda, wonder stuff... it's a
newer story...