LLVMPIPE -- a fork of softpipe that employs LLVM for code generation.
Status
======
Done so far is:
- the whole fragment pipeline is code generated in a single function
- input interpolation
- depth testing
- texture sampling (not all state/formats are supported)
- fragment shader TGSI translation
- same level of support as the TGSI SSE2 exec machine, with the exception
we don't fallback to TGSI interpretation when an unsupported opcode is
found, but just ignore it
- done in SoA layout
- input interpolation also code generated
- alpha testing
- blend (including logic ops)
- both in SoA and AoS layouts, but only the former used for now
- code is generic
- intermediates can be vectors of floats, ubytes, fixed point, etc, and of
any width and length
- not all operations are implemented for these types yet though
Most mesa/progs/demos/* work.
To do (probably by this order):
- code generate stipple and stencil testing
- translate the remaining bits of texture sampling state
- translate TGSI control flow instructions, and all other remaining opcodes
- integrate with the draw module for VS code generation
- code generate the triangle setup and rasterization
Requirements
============
- Linux
- udis86, http://udis86.sourceforge.net/ . Use my repository, which decodes
opcodes not yet supported by upstream.
git clone git://people.freedesktop.org/~jrfonseca/udis86
cd udis86
./configure --with-pic
make
sudo make install
- LLVM 2.5. On Debian based distributions do:
aptitude install llvm-dev
There is a typo in one of the llvm-dev 2.5 headers, that causes compilation
errors in the debug build:
--- /usr/include/llvm-c/Core.h.orig 2009-08-10 15:38:54.000000000 +0100
+++ /usr/include/llvm-c/Core.h 2009-08-10 15:38:25.000000000 +0100
@@ -831,7 +831,7 @@
template<typename T>
inline T **unwrap(LLVMValueRef *Vals, unsigned Length) {
#if DEBUG
- for (LLVMValueRef *I = Vals, E = Vals + Length; I != E; ++I)
+ for (LLVMValueRef *I = Vals, *E = Vals + Length; I != E; ++I)
cast<T>(*I);
#endif
return reinterpret_cast<T**>(Vals);
- A x86 or amd64 processor with support for sse2, sse3, and sse4.1 SIMD
instructions. This is necessary because we emit several SSE intrinsics for
convenience. See /proc/cpuinfo to know what your CPU supports.
- scons
Building
========
To build everything invoke scons as:
scons debug=yes statetrackers=mesa drivers=llvmpipe winsys=xlib dri=false -k
Alternatively, you can build it with GNU make, if you prefer, by invoking it as
make linux-llvm
but the rest of these instructions assume that scons is used.
Using
=====
Building will create a drop-in alternative for libGL.so. To use it set the
environment variables:
export LD_LIBRARY_PATH=$PWD/build/linux-x86_64-debug/lib:$LD_LIBRARY_PATH
or
export LD_LIBRARY_PATH=$PWD/build/linux-x86-debug/lib:$LD_LIBRARY_PATH
For performance evaluation pass debug=no to scons, and use the corresponding
lib directory without the "-debug" suffix.
Unit testing
============
Building will also create several unit tests in
build/linux-???-debug/gallium/drivers/llvmpipe:
- lp_test_blend: blending
- lp_test_conv: SIMD vector conversion
- lp_test_format: pixel unpacking/packing
Some of this tests can output results and benchmarks to a tab-separated-file
for posterior analysis, e.g.:
build/linux-x86_64-debug/gallium/drivers/llvmpipe/lp_test_blend -o blend.tsv
Development Notes
=================
- When looking to this code by the first time start in lp_state_fs.c, and
then skim through the lp_bld_* functions called in there, and the comments
at the top of the lp_bld_*.c functions.
- All lp_bld_*.[ch] are isolated from the rest of the driver, and could/may be
put in a stand-alone Gallium state -> LLVM IR translation module.
- We use LLVM-C bindings for now. They are not documented, but follow the C++
interfaces very closely, and appear to be complete enough for code
generation. See
http://npcontemplation.blogspot.com/2008/06/secret-of-llvm-c-bindings.html
for a stand-alone example.
659609e0ae