[Mesa-dev] [PATCH] Reorder LLVM passes, running mem2reg earlier.

[José Fonseca]

Török,

Thanks.

I didn't see as much improvement (most of the stuff I've been playing
with has actually simple shaders), but I saw no regression so I've
commited it. We have more benchmarks running continuously from git so
once the commit goes through them we should have more data.

Also, do you know any good piece of documentation describing the good
ordering of passes, or is it just trial/error?

Jose 


On Mon, 2010-05-03 at 07:43 -0700, Török Edwin wrote:
> This gives a ~30% shader optimization time improvement on blender.
> Tested by comparing the dumped LLVM modules.
> Current ordering:
> time ~/llvm-git/obj/Release-Asserts/bin/opt l.bc  -constprop -instcombine
> -mem2reg -gvn  -simplifycfg
> real    0m1.126s
> user    0m1.108s
> sys     0m0.012s
> 
> With this patch:
> time ~/llvm-git/obj/Release-Asserts/bin/opt l.bc -mem2reg -constprop -instcombine   -gvn  -simplifycfg
> real    0m0.885s
> user    0m0.880s
> sys     0m0.000s
> 
> The overall improvement in blender is ~15%.
> Blender without the patch takes 1m13s:
> edwin     5934 87.6 11.5 729440 458296 pts/5   SLl+ 17:35   1:13 blender
> 
> Blender with the patch takes 1m3s:
> edwin     5726 94.2 11.2 716424 446168 pts/5   SLl+ 17:32   1:03 blender
> 
> It is still slow with the patch, but better (most of the optimization time is
> taken up by GVN, see LLVM PR7023).
> 
> Signed-off-by: Török Edwin <edwintorok at gmail.com>
> ---
>  src/gallium/auxiliary/draw/draw_llvm.c |    4 ++--
>  src/gallium/drivers/llvmpipe/lp_jit.c  |    4 ++--
>  2 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/src/gallium/auxiliary/draw/draw_llvm.c b/src/gallium/auxiliary/draw/draw_llvm.c
> index 2c23428..ea9b7c9 100644
> --- a/src/gallium/auxiliary/draw/draw_llvm.c
> +++ b/src/gallium/auxiliary/draw/draw_llvm.c
> @@ -182,6 +182,8 @@ draw_llvm_create(struct draw_context *draw)
>     /* These are the passes currently listed in llvm-c/Transforms/Scalar.h,
>      * but there are more on SVN. */
>     /* TODO: Add more passes */
> +   LLVMAddCFGSimplificationPass(llvm->pass);
> +   LLVMAddPromoteMemoryToRegisterPass(llvm->pass);
>     LLVMAddConstantPropagationPass(llvm->pass);
>     if(util_cpu_caps.has_sse4_1) {
>        /* FIXME: There is a bug in this pass, whereby the combination of fptosi
> @@ -190,9 +192,7 @@ draw_llvm_create(struct draw_context *draw)
>         */
>        LLVMAddInstructionCombiningPass(llvm->pass);
>     }
> -   LLVMAddPromoteMemoryToRegisterPass(llvm->pass);
>     LLVMAddGVNPass(llvm->pass);
> -   LLVMAddCFGSimplificationPass(llvm->pass);
>  
>     init_globals(llvm);
>  
> diff --git a/src/gallium/drivers/llvmpipe/lp_jit.c b/src/gallium/drivers/llvmpipe/lp_jit.c
> index 466a2f5..30e206a 100644
> --- a/src/gallium/drivers/llvmpipe/lp_jit.c
> +++ b/src/gallium/drivers/llvmpipe/lp_jit.c
> @@ -185,6 +185,8 @@ lp_jit_screen_init(struct llvmpipe_screen *screen)
>        /* These are the passes currently listed in llvm-c/Transforms/Scalar.h,
>         * but there are more on SVN. */
>        /* TODO: Add more passes */
> +      LLVMAddCFGSimplificationPass(screen->pass);
> +      LLVMAddPromoteMemoryToRegisterPass(screen->pass);
>        LLVMAddConstantPropagationPass(screen->pass);
>        if(util_cpu_caps.has_sse4_1) {
>           /* FIXME: There is a bug in this pass, whereby the combination of fptosi
> @@ -193,9 +195,7 @@ lp_jit_screen_init(struct llvmpipe_screen *screen)
>            */
>           LLVMAddInstructionCombiningPass(screen->pass);
>        }
> -      LLVMAddPromoteMemoryToRegisterPass(screen->pass);
>        LLVMAddGVNPass(screen->pass);
> -      LLVMAddCFGSimplificationPass(screen->pass);
>     }
>  
>     lp_jit_init_globals(screen);