<div dir="ltr">I also got myself some benchmark data! I wrote a python script to which generates a shader_test file with 5000 back-to-back random integer division operations. The compiler takes a long time to compile the shader but, running it with shader_time, I see about a 2x improvement with my pass. </div> <div class="gmail_quote"><div dir="ltr">On Thu, Sep 13, 2018 at 2:41 PM Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>> wrote: </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Shader-db results on Sky Lake: total instructions in shared programs: 15105795 -> 15111403 (0.04%) instructions in affected programs: 72774 -> 78382 (7.71%) helped: 0 HURT: 265 Note that hurt here actually means helped because we're getting rid of integer quotient operations (which are a send on some platforms!) and replacing them with fairly cheap ALU ops. --- src/intel/compiler/brw_nir.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/src/intel/compiler/brw_nir.c b/src/intel/compiler/brw_nir.c index b38c3ba383d..4de0a6c44d4 100644 --- a/src/intel/compiler/brw_nir.c +++ b/src/intel/compiler/brw_nir.c @@ -569,6 +569,7 @@ brw_nir_optimize(nir_shader *nir, const struct brw_compiler *compiler, OPT(nir_opt_cse); OPT(nir_opt_peephole_select, 0); OPT(nir_opt_intrinsics); + OPT(nir_opt_idiv_const, 0); OPT(nir_opt_algebraic); OPT(nir_opt_constant_folding); OPT(nir_opt_dead_cf); @@ -675,7 +676,8 @@ brw_preprocess_nir(const struct brw_compiler *compiler, nir_shader *nir) */ nir_lower_int64(nir, nir_lower_imul64 | nir_lower_isign64 | - nir_lower_divmod64); + nir_lower_divmod64 | + nir_lower_imul_high64); nir = brw_nir_optimize(nir, compiler, is_scalar, true); -- 2.17.1 </blockquote></div>