aqnuep
-
@ This is the classic "brainiac vs speed-demon" thing. Better (and cheaper) to make smarter compiler than smarter hardware, I think.
5:59 PM May 16th
via Silver Bird
in reply to Icare3D
-
@ Also, I like the idea of letting the compiler to deal with instruction latencies instead of dedicated hardware.
5:58 PM May 16th
via Silver Bird
in reply to Icare3D
-
@ GK110 looks good. Dynamic parallelism sounds the most interesting to me, but the shuffle instr is also a clever & useful thing :)
5:55 PM May 16th
via Silver Bird
in reply to Icare3D
-
@ The question is not why 300 draw calls are slow in D3D11 but rather: why do you need so many?
5:45 PM May 15th
via Silver Bird
in reply to _Humus_
-
@ Because of texture proxies, which I don't understand why we keep hanging around here...
5:42 PM May 15th
via Silver Bird
in reply to g_truc
-
Insights has a cover:
11:21 AM May 14th
via web
-
An awesome reference : "Modern OpenGL" by Philip Rideout
10:22 AM May 13th
via Echofon
-
@ To be honest, I don't understand the cube map texture size limitation either. It should be 16K like any texture.
3:41 PM May 8th
via Silver Bird
in reply to g_truc
-
@ Hmm, then it seems my assumption has been falsified :)
3:36 PM May 8th
via Silver Bird
in reply to g_truc
-
@ Sorry, I meant highest common denominator. The point is Apple probably wants to support the same hw caps no matter what Mac you have
3:33 PM May 8th
via Silver Bird
in reply to g_truc
-
@ Lowest common denominator, as NVIDIA hardware only supports 512 layers ;)
3:31 PM May 8th
via Silver Bird
in reply to g_truc
-
Insights is ready for printing. On time for Siggraph release! :D
8:00 AM May 8th
via web
-
I'll finally be dropping DX9 support in the next version of C4.
5:20 PM May 7th
via web
-
@ Impressive! This means that no developer can have any more excuse not to adopt GL3/GL4 because of limited support across vendors :)
7:55 AM May 7th
via Silver Bird
in reply to g_truc
-
@ I wouldn't say Oo. It's good news! Though, I would be interested that how many of the ogl-samples-pack tests pass with it :)
6:20 AM May 3rd
via Silver Bird
in reply to g_truc
-
@ Well, that makes it a bit more complicated. DX9 hardware are not supported for a long time so no help there...
8:41 PM Apr 30th
via Silver Bird
in reply to EricLengyel
-
@ I wouldn't say it's never desired. But you can easily achieve what you want in the following way:
8:14 PM Apr 30th
via Silver Bird
in reply to EricLengyel
-
@ @ I didn't say that SI couldn't support bindless textures either :) I just thought, I try to avoid any misunderstanding...
7:25 PM Apr 30th
via Silver Bird
in reply to Icare3D
-
@ @ It is about virtual textures through partially resident image data. There are some publicly available slides about it.
7:16 PM Apr 30th
via Silver Bird
in reply to Icare3D
-
@ @ Christophe was refering to the upcoming sparse texture extension. That is not exactly the same as bindless texture.
7:14 PM Apr 30th
via Silver Bird
in reply to Icare3D
|
- Name Daniel Rákos
- Location Hungary
- Web http://www.raster...
- Bio Computer graphics enthusiast. What I publish here is my personal opinion and doesn't have anything to do with the opinions of my employer!
|
