summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* llvmpipe: TGSI -> LLVM SoA IR converstion.José Fonseca2009-08-293-0/+1588
| | | | Based on tgsi_sse2.c.
* llvmpipe: Dump only the generated function.José Fonseca2009-08-291-2/+2
|
* llvmpipe: Scalar -> vector broadcasting.José Fonseca2009-08-292-0/+23
|
* llvmpipe: More arithmetic ops.José Fonseca2009-08-292-1/+250
| | | | Mostly untested.
* llvmpipe: More intrinsic helpers.José Fonseca2009-08-292-19/+136
|
* llvmpipe: Fix typo.José Fonseca2009-08-291-2/+2
|
* util: Fix SCons build.José Fonseca2009-08-291-0/+4
|
* llvmpipe: Document the typo in llvm-c/Core.h.José Fonseca2009-08-291-0/+15
|
* util: Auto-generate u_format_table.c from SConscript.José Fonseca2009-08-291-0/+7
|
* llvmpipe: Simple description/instructions.José Fonseca2009-08-291-0/+48
|
* llvmpipe: Add missing include.José Fonseca2009-08-291-0/+1
|
* llvmpipe: Tiles in rgba8 format.José Fonseca2009-08-298-193/+171
|
* llvmpipe: Store tile color in SoA.José Fonseca2009-08-296-27/+1019
|
* llvmpipe: Code generate color masking.José Fonseca2009-08-294-94/+75
|
* llvmpipe: Use the generated SoA blending code.José Fonseca2009-08-297-780/+225
|
* llvmpipe: Handle disabled blending too.José Fonseca2009-08-293-1/+13
|
* xlib: Ensure one screen instance.José Fonseca2009-08-291-7/+2
|
* llvmpipe: Split the texture cache from the color/depth/stencil cache.José Fonseca2009-08-2913-168/+513
|
* xlib: Complete llvmpipe integration.José Fonseca2009-08-294-3/+12
|
* llvmpipe: SoA blending.José Fonseca2009-08-296-145/+635
| | | | Throughput seems to be 4x higher.
* util: Fix blend factor dumping.José Fonseca2009-08-291-0/+2
|
* llvmpipe: Update blending test.José Fonseca2009-08-291-15/+8
|
* llvmpipe: Handle more conversions.José Fonseca2009-08-293-73/+100
| | | | Not all successfully yet though..
* llvmpipe: Use full set of sse4 min/max intrinsics.José Fonseca2009-08-291-2/+20
|
* llvmpipe: Fix one const generation for some signed integers.José Fonseca2009-08-291-1/+7
|
* llvmpipe: Allow floating types without sign.José Fonseca2009-08-292-2/+3
| | | | With the meaning that all values are assumed to be positive.
* llvmpipe: Centralize lp_build_context initialization.José Fonseca2009-08-293-5/+20
|
* llvmpipe: Normalize the cycles with the number of channel.José Fonseca2009-08-292-7/+5
| | | | So that we have a comparable number for different formats.
* llvmpipe: Handle ubyte -> float conversion too.José Fonseca2009-08-293-30/+196
|
* llvmpipe: Make lp_const_offset visible.José Fonseca2009-08-292-1/+5
|
* llvmpipe: Only get the preprocessor flags from llvm-config.José Fonseca2009-08-291-1/+1
| | | | Otherwise we get -O2 optimization flag, preventing proper debugging.
* llvmpipe: Bootstrap type conversions.José Fonseca2009-08-294-0/+642
|
* llvmpipe: Use same type for reference vectors.José Fonseca2009-08-293-9/+12
|
* llvmpipe: Fix floating point const scale factor.José Fonseca2009-08-291-1/+3
|
* llvmpipe: Factor out shared test code into a separate module.José Fonseca2009-08-295-184/+598
|
* llvmpipe: Cleanup constant helpers.José Fonseca2009-08-295-40/+141
|
* llvmpipe: Move intrinsic helpers to a separate module.José Fonseca2009-08-294-42/+145
|
* llvmpipe: Collect richer blend data.José Fonseca2009-08-291-1/+10
|
* llvmpipe: Some notes about PSHUF.José Fonseca2009-08-291-4/+7
|
* llvmpipe: Optimize blend swizzles by using bitmasks instead of shuffles for ↵José Fonseca2009-08-296-55/+416
| | | | ubytes.
* llvmpipe: Write data to a tsv file for posterior analysis.José Fonseca2009-08-291-99/+170
|
* llvmpipe: Move p_build_context to lp_bld_type.hJosé Fonseca2009-08-292-26/+27
| | | | As it will be shared with more modules.
* util: New file dedicated to dump state in human/machine readable format.José Fonseca2009-08-294-0/+225
|
* llvmpipe: Don't use llvm.readcyclecounter.José Fonseca2009-08-291-27/+24
| | | | | LLVM during optimization reorders the rdtsc instructions next to each other, which makes it pretty useless.
* llvmpipe: Separate constant building.José Fonseca2009-08-296-135/+226
|
* llvmpipe: Move type support functions into a separate file.José Fonseca2009-08-297-204/+278
|
* llvmpipe: Measure the number of cycles taken for blending.José Fonseca2009-08-291-20/+90
|
* llvmpipe: Minor comments / debug mnemonics.José Fonseca2009-08-291-0/+11
|
* llvmpipe: Optimize away min/max with equal operands.José Fonseca2009-08-291-0/+14
|
* llvmpipe: Improve generated 8bit blending code.José Fonseca2009-08-291-0/+29
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-26 13:48:36 +0000