On Thu, 3 Feb 2005, Rune Petersen wrote:


You are quite right - also the following instruction:

{
EASY_PFS_INSTR0(MAD, SRC0C_XYZ, ONE, ZERO),
EASY_PFS_INSTR1(0, 0, 0 | PFS_FLAG_CONST, 0 | PFS_FLAG_CONST, NONE, ALL),
EASY_PFS_INSTR2(MAD, SRC0A, ONE, ZERO),
EASY_PFS_INSTR3(0, 0, 0 | PFS_FLAG_CONST, 0 | PFS_FLAG_CONST, OUTPUT)
}


looks like a NOP ( it basically does reg0<-reg0*1+0.0 ).

with some more testing I found:
- FLAT_COLOR_PIXEL_SHADER needs 2 NOP when resizing window->full screen to avoid corruption. Can you add these NOPs only when resizing? =)

Not really, but this might indicate a sort of bandwidth issue - one part is issuing pixels faster than the other part can receive. What resolution are you running at ?



- SINGLE_TEXTURE_VERTEX_SHADER needs 1 NOP to get q3a to work for me. else get a blank screen. The game runs fine I just can't see anything. The NOP might be needed for the initial setup.

This can also be fixed by setting alu_offset field to 0 instead of 1 - apparently it is an offset into alu program.


I missed it on my last commit, as reducing the length of the program did not overwrite previous instruction 1.

Also, if one believes alu_end then it says that the length of the program
is 1 in both cases and thus PFS_NOP pruning has mostly cosmetic effect.

                          best

                            Vladimir Dergachev


Rune Petersen



-------------------------------------------------------
This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
Tool for open source databases. Create drag-&-drop reports. Save time
by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
Download a FREE copy at http://www.intelliview.com/go/osdn_nl
--
_______________________________________________
Dri-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to