Age | Commit message (Collapse) | Author | |
---|---|---|---|
2010-09-23 | mesa: Remove SGI_color_matrix. | Eric Anholt | |
Another optional ARB_imaging subset extension. | |||
2010-08-25 | i965: Remove unnecessary header. | Vinson Lee | |
2010-08-22 | i965: Fix up WM push constant setup on gen6. | Eric Anholt | |
Fixes glsl-algebraic-add-add-1. | |||
2010-08-22 | i965: Use intel->gen >= 6 instead of IS_GEN6. | Eric Anholt | |
2010-06-12 | i965: Update gen6 paths for the streaming rework. | Eric Anholt | |
2010-06-11 | i965: Remove caching of surface state objects. | Eric Anholt | |
It turns out that computing a 56 byte key to look up a 20-byte object out of a hash table was some sort of a bad idea. Whoops. before: [ # ] backend test min(s) median(s) stddev. count [ 0] gl firefox-talos-gfx 37.799 38.203 0.39% 6/6 after: [ 0] gl firefox-talos-gfx 34.761 34.784 0.17% 5/6 | |||
2010-06-11 | i965: Convert the binding table to streamed indirect state. | Eric Anholt | |
This slightly reduces reduces cairo-gl firefox-talos-gfx runtime on my Ironlake: before: [ # ] backend test min(s) median(s) stddev. count [ 0] gl firefox-talos-gfx 38.236 38.383 0.43% 5/6 after: [ 0] gl firefox-talos-gfx 37.799 38.203 0.39% 6/6 It turns out the cost of caching these objects and looking them up in the cache again is greater than the cost of just computing the object again, particularly when the overhead of having a separate BO to pin is removed. (Those that are paying close attention will note that this is a reversal of the path I was moving the driver in a couple of years ago. The major thing that has changed is that back then all state was recomputed when we wrapped the streaming state buffer, including recompiling our precious programs. Now, we're uncaching just the objects that are cheap to compute, and retaining caching of expensive objects) | |||
2010-06-11 | i965: Split constant buffer setup from its surface state/binding state. | Eric Anholt | |
This was bothering me when redoing the binding tables. | |||
2010-06-11 | i965: Set the CC VP state immediately on state change. | Eric Anholt | |
The cache lookup of these two little floats was .12% of total CPU time on firefox-talos-gfx because we did it any time commonly-changed state changed. On the other hand, updating the CC VP bo immediately whenver CC VP state changes is a .07% overhead due to putting a driver hoook in glEnable(). | |||
2010-06-08 | intel: Change dri_bo_* to drm_intel_bo* to consistently use new API. | Eric Anholt | |
The slightly less mechanical change of converting the emit_reloc calls will follow. | |||
2010-02-25 | i965: Set up the SNB sampler state pointers. | Eric Anholt | |
2010-02-25 | i965: Move PIPELINE_SELECT to the top of gen6 3d pipeline setup. | Eric Anholt | |
2010-02-25 | i965: Untested Sandybridge WM packets. | Eric Anholt | |
2010-02-25 | i965: Hook up remaining Sandybridge state packets besides WM. | Eric Anholt | |
2010-02-25 | i965: Untested Sandybridge SF setup. | Eric Anholt | |
2010-02-25 | i965: Add Sandybridge viewport setup. | Eric Anholt | |
2010-02-25 | i965: Enable DRAWING_RECTANGLE emit on Sandybridge. | Eric Anholt | |
2010-02-25 | i965: Add Sandybridge scissor state. | Eric Anholt | |
2010-02-25 | i965: Set the state base address on Sandybridge. | Eric Anholt | |
2010-02-25 | i965: Reconnect the index/vertex setup. | Eric Anholt | |
2010-02-25 | i965: Set up the SNB URB. | Eric Anholt | |
even with vs disabled, still doesn't work. | |||
2010-02-25 | i965: Get vp-tri batchbuffers running (no rendering). | Eric Anholt | |
2010-02-25 | i965: Add untested REJECT_ALL clip state. | Eric Anholt | |
2010-02-25 | i965: Add untested passthrough GS setup. | Eric Anholt | |
2010-02-25 | i965: Add untested Sandybridge passthrough VS setup. | Eric Anholt | |
2010-02-25 | i965: Start adding support for the Sandybridge CC unit. | Eric Anholt | |
2010-02-25 | i965: Update WM surface state setup for sandybridge's new BLEND_STATE. | Eric Anholt | |
2010-02-25 | i965: Set up sandybridge depthbuffer. | Eric Anholt | |
2010-02-25 | intel: Start adding defines and some bits for sandybridge bringup. | Eric Anholt | |
2010-02-19 | Remove _mesa_memset in favor of plain memset. | Kenneth Graunke | |
This may break the SUNOS4 build, but it's no longer relevant. | |||
2010-01-19 | intel: Remove dead note_fence vtbl hook. | Eric Anholt | |
2010-01-19 | i965: Remove obsolete comment about the state atoms. | Eric Anholt | |
2009-11-21 | i965: Fix several memory leaks on exit. | Eric Anholt | |
Bug #25194. | |||
2009-11-03 | intel: avoid unnecessary front buffer flushing/updating | Brian Paul | |
Before, if we just called glXMakeCurrent() and didn't render anything we'd still trigger a flushFrontBuffer() call. Now only set the intel->front_buffer_dirty field at state validation time just before we draw something. NOTE: additional calls to intel_check_front_buffer_rendering() might be needed if I missed some rendering paths. | |||
2009-10-27 | i965: be clear that the Fallback field is a boolean, not a bitfield | Brian Paul | |
2009-09-24 | i965: Clean up some mess with the batch cache. | Eric Anholt | |
Its flagging of extra state that's already flagged by the vtbl new_batch when appropriate was confusing my tracking down of the OA clear bug. | |||
2009-08-26 | i965: added texture unit sanity check | Brian Paul | |
Check that all the textures needed by the current fragment program actually exist and are valid. | |||
2009-08-12 | i965: Avoid re-uploading the index buffer when we don't need to. | Eric Anholt | |
No performance difference proven at 95% confidence with my GLSL demo (n=10). | |||
2009-07-07 | i965: Remove BRW_NEW_INPUT_VARYING | Eric Anholt | |
This state flag has been unused since the ffvertex_prog move to core. | |||
2009-05-06 | i965: Split WM constant buffer update from other WM surfaces. | Eric Anholt | |
This can avoid re-uploading constant data when it isn't necessary, and is a step towards not updating other surfaces just because constants change. It also brings the upload of the constant buffer next to the creation. This brings openarena performance up another 4%, to 91% of the Mesa 7.4 branch. | |||
2009-05-06 | i965: Disentangle VS constant surface state from WM surface state. | Eric Anholt | |
Also, only create VS surface state if there's a VS constant buffer to be uploaded, and set the contents of the buffer at the same time as creation. | |||
2009-04-22 | i965: checkpoint commit: use two state caches instead of one | Brian Paul | |
The new, second cache will only be used for surface-related items. Since we can create many surfaces the original, single cache could get filled quickly. When we cleared it, we had to regenerate shaders, etc. With two caches, we can avoid doing that. | |||
2009-04-22 | i965: remove unused state atom entries | Brian Paul | |
2009-04-22 | i965: the brw_constant_buffer state atom is no longer dynamic | Brian Paul | |
No more dynamic atoms so we can simplify the state validation code a little. | |||
2009-04-22 | i965: add _NEW_PROGRAM_CONSTANTS to mesa_bits[] list | Brian Paul | |
2009-02-02 | i965: Remove brw->attribs now that we can just always look in the GLcontext. | Eric Anholt | |
2009-02-02 | i965: Delete old metaops code now that there are no remaining consumers. | Eric Anholt | |
2009-01-07 | i965: Remove worrisome comment about _NEW_PROGRAM signaling fp change. | Eric Anholt | |
Everything now depends on either BRW_NEW_FRAGMENT_PROGRAM or BRW_NEW_VERTEX_PROGRAM. | |||
2008-11-28 | i965: Remove BRW_WM_LOCK dirty bit, introduced to work around lack of relocs. | Eric Anholt | |
This was causing a prepare of wm state at every primitive emit. | |||
2008-11-28 | i965: Add debug code for dumping how frequently different dirty bits are set. | Eric Anholt | |