Graphics card compatibility
-
A while back, somewhere on this forum, I came across a post concerning graphics card compatibility with SU, had at least a partial list of known problems with certain cards and SU. I've just built a new computer and am having problems with my card, occasionally gives error msg below. When this happens any instance of SU I have running closes and I lose all my work. Very frustrating. Any idea where this post might be found? (and of course visiting the site mentioned provides absolutely no help at all)
-
CGarchitect Graphics Card Review Roundup: http://www.cgarchitect.com/news/Reviews/Review076_1.asp
GPU feedback - SketchUp Sage:
http://sites.google.com/site/sketchupsage/problems/gpu/gpu-feedback -
Have you tried updating your graphic card drivers?
-
Thanks for all your suggestions. Yes, I did update my drivers, it's the first thing I do when I get a new piece of hardware. I got an RMA for my card and ordered a new one. Unfortunately I did that before I read about all the problems people have been having with ATI cards. Guess what kind of card I ordered. Crap. Well, I'll give it a try and see what happens.
I've also been reading about the suggestions people have been giving Google for the next version of SU, and was wondering if there have been any new developments as far as support for multiple core CPU's. It's such a shame that we are limited to just one core, so much could be done if we could use all the power of our computers.
Well, anyhow, I'll let you know how it goes with my new card. -
@hellnbak said:
It's such a shame that we are limited to just one core, so much could be done if we could use all the power of our computers.
slicing a 'dumb' bitmap operation as e.g. rendering/raytracing to multiple bands on multiple cores is much more easy than a vector based operation as e.g. modeling curves/faces. Actually I do not know any modeler capable of doing this, therefore blaming Google for carelessly neglecting this putately simple enhancement can be negated...
hth,
Norbert -
@sketch3d.de said:
splicing a 'dumb' bitmap operation as e.g. rendering/raytracing to multiple bands on multiple cores is much more easy than a vector based operation as e.g. modeling curves/faces. Actually I do not know any modeler capable of doing this, therefore blaming Google for carelessly neglecting this putately simple enhancement can be negated...
hth,NorbertNot sure I understand what you are trying to say here. Actually I'm sure I don't understand, my knowledge of such things is all but nonexistent. Do you mean that it would not be possible for Google to modify SU to take advantage of multiple core CPU's, or that it would be too difficult for them to do so, or that even if they did it would not be of benefit to us?
And I'm not "blaming" Google for anything. Like many others, I'm simply wondering if certain improvements to SU are in the works.
-
@hellnbak said:
Not sure I understand what you are trying to say here. Actually I'm sure I don't understand, my knowledge of such things is all but nonexistent. Do you mean that it would not be possible for Google to modify SU to take advantage of multiple core CPU's, or that it would be too difficult for them to do so, or that even if they did it would not be of benefit to us?
And I'm not "blaming" Google for anything. Like many others, I'm simply wondering if certain improvements to SU are in the works.
Yes, some types computation can not be split into multiple processes. It's far from trivial to make use of multiple processes - only some types of jobs can be done like that. Most thing requires to be done linearly instead of in parallel.
When you see various 3d packages advertises they are multi-core aware it's usually something like the rendering process that is a multi-core process. But the actual modelling and displaying the model to the viewport is single process. -
@hellnbak said:
Do you mean that it would not be...
would be of great benefit but very difficult for modeling/display operations if not impossible.
Thomas has elaborated already, extended information concerning this feature request can be found in the appropriate Google Group.
@hellnbak said:
It's such a shame... And I'm not "blaming" Google for anything.
of course
Norb.
Advertisement