[Mesa-dev] [PATCH 2/4] gallivm: remove LICM pass

sroland at vmware.com sroland at vmware.com
Mon Apr 23 04:24:36 UTC 2018


From: Roland Scheidegger <sroland at vmware.com>

LICM is simply too expensive, even though it presumably can help quite
a bit in some cases.
It was definitely cheaper in llvm 3.3, though as far as I can tell with
llvm 3.3 it failed to do anything in most cases. early-cse also actually
seems to cause licm to be able to move things when it previously couldn't,
which causes noticeable compile time increases.
There's more loop passes in llvm, but I'm not sure which ones are helpful,
and I couldn't find anything which would roughly do what the old licm in
llvm 3.3 did, so ditch it.
---
 src/gallium/auxiliary/gallivm/lp_bld_init.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/src/gallium/auxiliary/gallivm/lp_bld_init.c b/src/gallium/auxiliary/gallivm/lp_bld_init.c
index 05a74a0..abca624 100644
--- a/src/gallium/auxiliary/gallivm/lp_bld_init.c
+++ b/src/gallium/auxiliary/gallivm/lp_bld_init.c
@@ -145,7 +145,15 @@ create_pass_manager(struct gallivm_state *gallivm)
       LLVMAddScalarReplAggregatesPass(gallivm->passmgr);
       LLVMAddEarlyCSEPass(gallivm->passmgr);
       LLVMAddCFGSimplificationPass(gallivm->passmgr);
-      LLVMAddLICMPass(gallivm->passmgr);
+      /* 
+       * FIXME: LICM is potentially quite useful. However, for some
+       * rather crazy shaders the compile time can reach _hours_ per shader,
+       * due to licm implying lcssa (since llvm 3.5), which can take forever.
+       * Even for sane shaders, the cost of licm is rather high (and not just
+       * due to lcssa, licm itself too), though mostly only in cases when it
+       * can actually move things, so having to disable it is a pity.   
+       * LLVMAddLICMPass(gallivm->passmgr);
+       */
       LLVMAddReassociatePass(gallivm->passmgr);
       LLVMAddPromoteMemoryToRegisterPass(gallivm->passmgr);
       LLVMAddConstantPropagationPass(gallivm->passmgr);
-- 
2.7.4



More information about the mesa-dev mailing list