Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | llvmpipe: added debug option to disable LLVM optimization passes | Brian Paul | 2010-01-27 | 3 | -13/+19 |
| | |||||
* | llvmpipe: count/report time spent in LLVM compilations | Brian Paul | 2010-01-27 | 3 | -4/+24 |
| | |||||
* | llvmpipe: use some local vars to index step arrays | Brian Paul | 2010-01-21 | 1 | -4/+6 |
| | | | | Saves a few more cycles. | ||||
* | llvmpipe: s/inline/INLINE/ | Brian Paul | 2010-01-21 | 1 | -1/+1 |
| | |||||
* | llvmpipe: consolidate lp_scene_alloc_aligned() calls | Brian Paul | 2010-01-21 | 1 | -13/+41 |
| | | | | Use just one call instead of four. Good for a few more fps. | ||||
* | llvmpipe: re-use a1 var in linear_coef() | Brian Paul | 2010-01-21 | 1 | -1/+1 |
| | |||||
* | llvmpipe: area is an int here, not float | Brian Paul | 2010-01-21 | 1 | -4/+4 |
| | |||||
* | llvmpipe: manually unroll the inputs.step[] setup code | Brian Paul | 2010-01-21 | 1 | -18/+31 |
| | | | | Good for a few more fps in some tests. | ||||
* | llvmpipe: added simple perf/statistics counting facility | Brian Paul | 2010-01-21 | 11 | -6/+189 |
| | | | | | | | | Currently counting number of tris, how many tiles of each size are fully covered, partially covered or empty, etc. Set LP_DEBUG=counters to enable. Results are printed upon context destruction. | ||||
* | llvmpipe: optimize tile writing code | Brian Paul | 2010-01-20 | 1 | -18/+109 |
| | | | | | | | The code which converts/copies color tiles to the linear layout has been rewritten. There's less arithmetic and better loop unrolling, and possibly a better memory access pattern. Some demos, like gears, are about 20% faster now. | ||||
* | llvmpipe: remove tile clipping code | Brian Paul | 2010-01-20 | 1 | -20/+0 |
| | | | | The surface is always a multiple of the tile size now. | ||||
* | llvmpipe: align display target size to multiple of tile size | Brian Paul | 2010-01-20 | 1 | -7/+13 |
| | | | | This will allow us to skip clipping tiles to surface bounds. | ||||
* | llvmpipe: put TILE_SIZE #define in its own header | Brian Paul | 2010-01-20 | 2 | -11/+43 |
| | | | | To avoid including lp_tile_soa.h in other places. | ||||
* | llvmpipe: add makefile rule for generating .s files | Brian Paul | 2010-01-20 | 1 | -0/+5 |
| | |||||
* | llvmpipe: updated comment | Brian Paul | 2010-01-20 | 1 | -0/+2 |
| | |||||
* | llvmpipe: asst. task-related clean-ups | Brian Paul | 2010-01-19 | 1 | -22/+29 |
| | |||||
* | llvmpipe: fix-up comment | Brian Paul | 2010-01-19 | 1 | -1/+2 |
| | |||||
* | llvmpipe: re-order file list, fix indentation | Brian Paul | 2010-01-19 | 2 | -9/+9 |
| | |||||
* | llvmpipe: remove unneeded DEBUG checks, use step var | Brian Paul | 2010-01-19 | 1 | -5/+1 |
| | |||||
* | llvmpipe: updated comments | Brian Paul | 2010-01-19 | 1 | -2/+4 |
| | |||||
* | llvmpipe: tweak a comment | Brian Paul | 2010-01-18 | 1 | -1/+1 |
| | |||||
* | llvmpipe: added show_tiles and show_subtiles debug options | Brian Paul | 2010-01-18 | 3 | -0/+65 |
| | | | | | These options draw lines over the tiles to show the 64x64 tile bounds and 16x16 sub-tile bounds. For debugging/visualization. | ||||
* | llvmpipe: use new u_ringbuffer for scene queue | Keith Whitwell | 2010-01-17 | 4 | -89/+46 |
| | |||||
* | Merge remote branch 'origin/master' into lp-binning | José Fonseca | 2010-01-16 | 11 | -49/+59 |
|\ | | | | | | | | | | | Conflicts: src/gallium/drivers/llvmpipe/lp_quad.h src/gallium/drivers/llvmpipe/lp_setup.c | ||||
| * | Merge branch 'gallium-noconstbuf' | Roland Scheidegger | 2010-01-15 | 4 | -9/+8 |
| |\ | | | | | | | | | | | | | | | | Conflicts: src/gallium/drivers/softpipe/sp_draw_arrays.c src/mesa/state_tracker/st_draw_feedback.c | ||||
| | * | gallium: remove const qualifier from pipe_buffer argument in set_constant_buffer | Roland Scheidegger | 2010-01-11 | 2 | -2/+2 |
| | | | |||||
| | * | gallium: adapt drivers to pipe_constant_buffer removal | Roland Scheidegger | 2009-12-24 | 4 | -9/+8 |
| | | | |||||
| * | | gallium: Rename ALIGN_STACK -> PIPE_ALIGN_STACK for consistency. | José Fonseca | 2010-01-12 | 4 | -4/+4 |
| | | | |||||
| * | | gallium: Simplify PIPE_ALIGN_VAR. | José Fonseca | 2010-01-12 | 4 | -17/+17 |
| | | | | | | | | | | | | | | | | | | gcc allows pre-fix variable attributes. Suggested by Ian Romanick. | ||||
| * | | gallium: Generalize the alignment macros to other compilers and any alignment. | José Fonseca | 2010-01-12 | 4 | -17/+18 |
| | | | |||||
| * | | llvmpipe: Update instructions. | José Fonseca | 2010-01-10 | 1 | -21/+18 |
| | | | | | | | | | | | | | | | | | | | | | Explain how to build for windows. Both MSVC 9 and cross MinGW supported. Stop documenting LLVM 2.5 as supported. It still supported at the moment but it will soon stop being. | ||||
| * | | llvmpipe: Yet another hack to get release LLVM static libraries to link in ↵ | José Fonseca | 2010-01-10 | 1 | -0/+14 |
| | | | | | | | | | | | | debug builds. | ||||
* | | | llvmpipe: implement scissor test in triangle setup | Brian Paul | 2010-01-15 | 5 | -3/+15 |
| | | | |||||
* | | | llvmpipe: remove lp_rast_triangle::min/max fields | Brian Paul | 2010-01-15 | 2 | -16/+10 |
| | | | | | | | | | | | | These values aren't needed outside the do_triangle_ccw() function. | ||||
* | | | llvmpipe: skip 4x4 in/out test code | Brian Paul | 2010-01-15 | 1 | -14/+4 |
| | | | | | | | | | | | | | | | It's a litte faster to just do the in/out testing in the shader jit code. | ||||
* | | | llvmpipe: added comment about lookup-tables vs. computation | Brian Paul | 2010-01-15 | 1 | -0/+3 |
| | | | |||||
* | | | llvmpipe: change 'in' to boolean, add comments | Brian Paul | 2010-01-15 | 1 | -4/+4 |
| | | | |||||
* | | | llvmpipe: generate two shader varients, one omits triangle in/out testing | Brian Paul | 2010-01-15 | 8 | -52/+142 |
| | | | | | | | | | | | | | | | | | | | | | When we know that a 4x4 pixel block is entirely inside of a triangle use the jit function which omits the in/out test code. Results in a few percent speedup in many tests. | ||||
* | | | llvmpipe: asst code changes in lp_state_fs.c | Brian Paul | 2010-01-15 | 1 | -27/+32 |
| | | | |||||
* | | | llvmpipe: put labels on some value refs | Brian Paul | 2010-01-15 | 1 | -1/+5 |
| | | | |||||
* | | | llvmpipe: add extra flags to DEFINES, no CFLAGS | Brian Paul | 2010-01-15 | 1 | -1/+1 |
| | | | | | | | | | | | | | | | Doesn't make any real difference but -D flags are put into DEFINES everywhere else. | ||||
* | | | llvmpipe: implement scissor testing | Brian Paul | 2010-01-14 | 8 | -7/+134 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The scissor test is implemented as another per-quad operation in the JIT code. The four scissor box params are passed via the lp_jit_context. In the JIT code we compare the quad's x/y coords against the clip bounds and create a new in/out mask that's AND'd with the main quad mask. Note: we should also do scissor testing in the triangle setup code to improve efficiency. That's not done yet. | ||||
* | | | llvmpipe: minor comment update | Brian Paul | 2010-01-14 | 1 | -1/+1 |
| | | | |||||
* | | | llvmpipe: clean-up, fixing up frame dump/debug code | Brian Paul | 2010-01-14 | 1 | -7/+12 |
| | | | |||||
* | | | llvmpipe: remove redundant code in llvmpipe_set_blend_color() | Brian Paul | 2010-01-14 | 1 | -3/+0 |
| | | | |||||
* | | | llvmpipe: assorted clean-ups in texture code | Brian Paul | 2010-01-14 | 3 | -33/+29 |
| | | | |||||
* | | | llvmpipe: use one loop in lp_rast_clear_zstencil() | Brian Paul | 2010-01-13 | 1 | -4/+3 |
| | | | | | | | | | | | | This is just a tiny bit faster. | ||||
* | | | llvmpipe: optimize lp_rast_clear_color() for non-gray colors | Brian Paul | 2010-01-13 | 1 | -6/+22 |
| | | | | | | | | | | | | | | | This makes a big difference in progs that clear to a non-gray color. Some demos are 30-50% faster. | ||||
* | | | llvmpipe: minor assorted clean-ups | Brian Paul | 2010-01-13 | 1 | -10/+12 |
| | | | |||||
* | | | llvmpipe: check render targets before other textures | Brian Paul | 2010-01-13 | 1 | -6/+9 |
| | | |