aboutsummaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/llvmpipe/lp_tile_cache.c
Commit message (Collapse)AuthorAgeFilesLines
* move assert to avoid crash in debug build.Alan Hourihane2009-12-081-1/+2
|
* llvmpipe: Ensure transfers are mapped.José Fonseca2009-12-041-0/+4
| | | | | This shouldn't happen but it does by some misterious reason. Fail the assertion but at least do not segfault on release builds.
* gallium: adapt drivers to interface cleanupsRoland Scheidegger2009-12-021-3/+3
|
* llvmpipe: Ensure tile cache transfers are mapped before flushing it.José Fonseca2009-10-041-0/+2
|
* llvmpipe: Autogenerate lp_tile_soa.c from u_format.csv.José Fonseca2009-10-041-37/+40
| | | | | | | | | This is just a temporary change until we code generate the tile read/write functions in runtime. The new code avoids an extra memcpy that exists in u_tile.c functions, from which lp_tile_soa.c was originally based. This achieves up to 5% improvement, particularly in frames with little geometry overlap.
* llvmpipe: Fetch tile only if a color buffer is bound.José Fonseca2009-10-031-0/+3
|
* llvmpipe: move tile cache datatypes into .c file since they're privateBrian Paul2009-09-231-0/+37
|
* llvmpipe: added max texture/surface size sanity checkBrian Paul2009-09-231-0/+6
| | | | Carried over from softpipe driver.
* llvmpipe: Fix lp_get_cached_tile.José Fonseca2009-09-201-1/+1
| | | | Align coordinates to tile boundaries.
* llvmpipe: Update tile status on flush.José Fonseca2009-09-201-0/+3
|
* llvmpipe: Update for changes in master.José Fonseca2009-08-291-1/+1
|
* llvmpipe: Don't limit the number of tiles in the cache.José Fonseca2009-08-291-172/+84
|
* llvmpipe: Clean dead tile cache code.José Fonseca2009-08-291-21/+0
|
* llvmpipe: Simplify tile clears. Use pipe_fill_rect instead of scratch tile.José Fonseca2009-08-291-53/+35
|
* llvmpipe: Start hiding llvmpipe_cached_tile.José Fonseca2009-08-291-2/+3
|
* llvmpipe: Drop depth/stencil support from tile cache.José Fonseca2009-08-291-79/+18
|
* llvmpipe: use align_malloc for all structs containing ALIGN16 membersKeith Whitwell2009-08-291-2/+3
| | | | | Unless the struct is allocated aligned, aligning the members isn't very helpful.
* llvmpipe: Tile in SoA format identical to the blender function output.José Fonseca2009-08-291-4/+7
|
* llvmpipe: Add missing include.José Fonseca2009-08-291-0/+1
|
* llvmpipe: Tiles in rgba8 format.José Fonseca2009-08-291-1/+1
|
* llvmpipe: Store tile color in SoA.José Fonseca2009-08-291-24/+18
|
* llvmpipe: Split the texture cache from the color/depth/stencil cache.José Fonseca2009-08-291-127/+0
|
* llvmpipe: also shortcircuit non-texture tile lookupsKeith Whitwell2009-08-291-17/+16
|
* llvmpipe: shortcircuit repeated lookups of the same tileKeith Whitwell2009-08-291-88/+65
| | | | | | | | | The lp_tile_cache is often called repeatedly to look up the same tile. Add a cache (to the cache) of the single tile most recently retreived and make a quick inline check to see if this matches the subsequent request. Add a tile_address bitfield struct to make this check easier.
* llvmpipe: make some small steps to flush texture cache less frequentlyKeith Whitwell2009-08-291-32/+48
| | | | No performance gain yet, but the code is a bit cleaner.
* llvmpipe: remove backwards dependency from tilecache to llvmpipeKeith Whitwell2009-08-291-12/+8
| | | | | | | | The tile cache is a utility, it shouldn't know anything about the entity which is making use of it (ie llvmpipe). Remove llvmpipe parameter to all the tilecache function calls, and also remove the need to keep a llvmpipe pointer in the sampler structs.
* llvmpipe: Fork softpipe for experimentation with llvm.José Fonseca2009-08-291-0/+626