RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-09 22:04:10 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
33,850 Commits 6 Branches 6 Tags
Commit Graph

33850 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Chris Lattner
7cdaff2298 Switch some std::sets to SmallPtrSet. This speeds up 
domtree by 10% and postdomtree by 17%

llvm-svn: 40829
 2007-08-05 00:15:57 +00:00
Chris Lattner
9c1a2dfa53 Switch DomTreeNode::assignDFSNumber from using a std::set to using 
a smallptrset.  This speeds up domtree by about 15% and postdomtree by 20%.

llvm-svn: 40828
 2007-08-05 00:10:08 +00:00
Chris Lattner
a5a692b9bb Switch the internal "Info" map from an std::map to a DenseMap. This 
speeds up idom by about 45% and postidom by about 33%.

Some extra precautions must be taken not to invalidate densemap iterators.

llvm-svn: 40827
 2007-08-05 00:02:00 +00:00
Chris Lattner
5912537997 switch the DomTreeNodes and IDoms maps in idom/postidom to a 
DenseMap instead of an std::map.  This speeds up postdomtree
by about 25% and domtree by about 23%.  It also speeds up clients,
for example, domfrontier by 11%, mem2reg by 4% and ADCE by 6%.

llvm-svn: 40826
 2007-08-04 23:48:07 +00:00
Chris Lattner
e562e9bdb0 rewrite the code used to construct pruned SSA form with the IDF method. 
In the old way, we computed and inserted phi nodes for the whole IDF of 
the definitions of the alloca, then computed which ones were dead and
removed them.

In the new method, we first compute the region where the value is live,
and use that information to only insert phi nodes that are live.  This
eliminates the need to compute liveness later, and stops the algorithm
from inserting a bunch of phis which it then later removes.

This speeds up the testcase in PR1432 from 2.00s to 0.15s (14x) in a
release build and 6.84s->0.50s (14x) in a debug build.

llvm-svn: 40825
 2007-08-04 22:50:14 +00:00
Chris Lattner
b7d4ef6ca6 Factor out a whole bunch of code into it's own method. 
llvm-svn: 40824
 2007-08-04 21:14:29 +00:00
Chris Lattner
d4a88d77d4 Use getNumPreds(BB) instead of computing them manually. This is a very small but 
measurable speedup.

llvm-svn: 40823
 2007-08-04 21:06:15 +00:00
Chris Lattner
6b9dca62dd Change the rename pass to be "tail recursive", only adding N-1 successors 
to the worklist, and handling the last one with a 'tail call'.  This speeds
up PR1432 from 2.0578s to 2.0012s (2.8%)

llvm-svn: 40822
 2007-08-04 20:40:27 +00:00
Chris Lattner
c1d2c2bdc8 cache computation of #preds for a BB. This speeds up 
mem2reg from 2.0742->2.0522s on PR1432.

llvm-svn: 40821
 2007-08-04 20:24:50 +00:00
Chris Lattner
8335a86536 reserve operand space for phi nodes when we insert them. 
llvm-svn: 40820
 2007-08-04 20:14:34 +00:00
Chris Lattner
32d9e4ba5c use continue to avoid nesting, no functionality change. 
llvm-svn: 40819
 2007-08-04 20:07:06 +00:00
Chris Lattner
a97ceae263 Promoting allocas with the 'single store' fastpath is 
faster than with the 'local to a block' fastpath.  This speeds
up PR1432 from 2.1232 to 2.0686s (2.6%)

llvm-svn: 40818
 2007-08-04 20:03:23 +00:00
Chris Lattner
479e3fa267 When PromoteLocallyUsedAllocas promoted allocas, it didn't remember 
to increment NumLocalPromoted, and didn't actually delete the
dead alloca, leading to an extra iteration of mem2reg.

llvm-svn: 40817
 2007-08-04 20:01:43 +00:00
Chris Lattner
bd506a8e12 std::map -> DenseMap 
llvm-svn: 40816
 2007-08-04 19:52:20 +00:00
Nick Lewycky
4e96222245 Clean up comments, fix up some confusing code logic. 
Predsimplify fails llvm-gcc bootstrap.

llvm-svn: 40815
 2007-08-04 18:45:32 +00:00
Chris Lattner
9748fa5c6f fix a logic bug where we wouldn't promote single store allocas if the 
stored value was a non-instruction value.  Doh.

This increase the # single store allocas from 8982 to 9026, and
speeds up mem2reg on the testcase in PR1432 from 2.17 to 2.13s.

llvm-svn: 40813
 2007-08-04 02:45:02 +00:00
Chris Lattner
3f971fdbd5 When we do the single-store optimization, delete both the store 
and the alloca so they don't get reprocessed.

This speeds up PR1432 from 2.20s to 2.17s.

llvm-svn: 40812
 2007-08-04 02:38:38 +00:00
Chris Lattner
c38b2a2473 Three improvements: 
1. Check for revisiting a block before checking domination, which is faster.
  2. If the stored value isn't an instruction, we don't have to check for domination.
  3. If we have a value used in the same block more than once, make sure to remove the
     block from the UsingBlocks vector.  Not doing so forces us to go through the slow
     path for the alloca.

The combination of these improvements increases the number of allocas on the fastpath
from 8935 to 8982 on PR1432.  This speeds it up from 2.90s to 2.20s (31%)

llvm-svn: 40811
 2007-08-04 02:32:22 +00:00
Chris Lattner
fe6a3e2fb4 switch from using a std::set to using a SmallPtrSet. This speeds up the 
testcase in PR1432 from 6.33s to 2.90s (2.22x)

llvm-svn: 40810
 2007-08-04 02:21:22 +00:00
Chris Lattner
9b45ad1f5c In mem2reg, when handling the single-store case, make sure to remove 
a using block from the list if we handle it.  Not doing this caused us
to not be able to promote (with the fast path) allocas which have uses (whoops).

This increases the # allocas hitting this fastpath from 4042 to 8935 on the
testcase in PR1432, speeding up mem2reg by 2.6x

llvm-svn: 40809
 2007-08-04 02:15:24 +00:00
Chandler Carruth
d76683caee Regenerating. 
llvm-svn: 40808
 2007-08-04 01:56:21 +00:00
Chandler Carruth
00e56b0e81 This is the patch to provide clean intrinsic function overloading support in LLVM. It cleans up the intrinsic definitions and generally smooths the process for more complicated intrinsic writing. It will be used by the upcoming atomic intrinsics as well as vector and float intrinsics in the future. 
This also changes the syntax for llvm.bswap, llvm.part.set, llvm.part.select, and llvm.ct* intrinsics. They are automatically upgraded by both the LLVM ASM reader and the bitcode reader. The test cases have been updated, with special tests added to ensure the automatic upgrading is supported.

llvm-svn: 40807
 2007-08-04 01:51:18 +00:00
Chris Lattner
b5af2cf90d split rewriting of single-store allocas into its own 
method.

llvm-svn: 40806
 2007-08-04 01:47:41 +00:00
Chris Lattner
f8b22a3568 refactor some code to shrink PromoteMem2Reg::run a bit 
llvm-svn: 40805
 2007-08-04 01:41:18 +00:00
Chris Lattner
2d4d557c80 add a typedef, no other change. 
llvm-svn: 40804
 2007-08-04 01:19:38 +00:00
Chris Lattner
65181deba3 avoid an unneeded vector copy. This speeds up mem2reg on the testcase 
in PR1432 by 6%

llvm-svn: 40803
 2007-08-04 01:07:49 +00:00
Chris Lattner
34cf111957 make RenamePassWorkList a local var instead of an ivar. 
llvm-svn: 40802
 2007-08-04 01:04:40 +00:00
Dale Johannesen
c8462173c5 Make x86 long double alignment 32 for everything but 
Darwin (which makes size within a struct==96)

llvm-svn: 40796
 2007-08-03 22:46:15 +00:00
Dale Johannesen
976a1cf6cf long double patch 3 of N. Add to MVT. 
llvm-svn: 40793
 2007-08-03 20:51:37 +00:00
Dale Johannesen
4b91728849 long double patch 2 of N. Handle it in TargetData. 
(I've tried to get the info right for all targets,
but I'm not expert on all of them - check yours.)

llvm-svn: 40792
 2007-08-03 20:20:50 +00:00
Owen Anderson
891f5b4d28 Fix a subtle miscompilation. This allows 197.parser to be compiled correctly. 
llvm-svn: 40791
 2007-08-03 19:59:35 +00:00
Owen Anderson
d1063b7a3c Fix a subtle iterator invalidation bug in a recursive algorithm. 
llvm-svn: 40776
 2007-08-03 11:03:26 +00:00
Reid Spencer
d1b39fc597 Prepare for "core" website. 
llvm-svn: 40775
 2007-08-03 05:43:35 +00:00
Dale Johannesen
dda00eb44d Long double, part 1 of N. Support in IR. 
llvm-svn: 40774
 2007-08-03 01:03:46 +00:00
Chris Lattner
a9026e5f12 add an observation 
llvm-svn: 40772
 2007-08-03 00:17:42 +00:00
Chris Lattner
cf6dd53ad8 Fix an accidental commit. 
llvm-svn: 40758
 2007-08-02 21:33:36 +00:00
Dan Gohman
352de5aeb9 More explicit keywords. 
llvm-svn: 40757
 2007-08-02 21:21:54 +00:00
Dan Gohman
1afde4166e Fix the alignment requirements of several unpck and shuf instructions. 
Generalize isPSHUFDMask and add a unary SHUFPD pattern so that SHUFPD's
memory operand alignment can be tested as well, with a fix to avoid
breaking MMX's use of isPSHUFDMask.

llvm-svn: 40756
 2007-08-02 21:17:01 +00:00
Dan Gohman
7d1e35e6d1 Fix pastos in vector arithmetic intrinsics. 
llvm-svn: 40754
 2007-08-02 21:06:40 +00:00
Owen Anderson
414247c075 Fix 80 col. violations. 
llvm-svn: 40751
 2007-08-02 18:20:52 +00:00
Owen Anderson
ad74eeed3d Fix 80 col. violations. 
llvm-svn: 40750
 2007-08-02 18:16:06 +00:00
Owen Anderson
ae769ac824 Fix 80 col. violations. 
llvm-svn: 40749
 2007-08-02 18:11:11 +00:00
Chris Lattner
9319dfc93a don't redefine a parameter 
llvm-svn: 40748
 2007-08-02 18:08:16 +00:00
Owen Anderson
d617e4c963 Fix a bug that was causing several miscompilations on SPEC. 
llvm-svn: 40746
 2007-08-02 17:56:05 +00:00
Christopher Lamb
f27d856acd Implement review feedback. 
llvm-svn: 40745
 2007-08-02 17:52:00 +00:00
Chris Lattner
4670f15d73 I don't have time to restore this functionality right now. 
llvm-svn: 40743
 2007-08-02 17:43:39 +00:00
Chris Lattner
5ca7348fc4 Replacing a cast with another one does not reduce the number of 
casts in the input.

llvm-svn: 40741
 2007-08-02 17:23:38 +00:00
Chris Lattner
f0f4024c46 Reduced testcase for PR1594 
llvm-svn: 40740
 2007-08-02 17:11:24 +00:00
Chris Lattner
7c6e8f735f Disable an xform that causes an infinite loop. This fixes PR1594 
llvm-svn: 40739
 2007-08-02 16:56:32 +00:00
Chris Lattner
25a8bfdedb wrap some long lines. Major offenders that are left include 
gvn, gvnpre, dse, and predsimplify.  To see these, use:

  make check-line-length

llvm-svn: 40738
 2007-08-02 16:53:43 +00:00