Chris Lattner
06d8aac065
Ugh, fix bugs. Ok, so the last fix wasn't as great as I thought it was.
...
Now we're back to: 131.730u, 0.330s, which is still quite an improvement,
but still quite unacceptable
llvm-svn: 8322
2003-09-02 19:14:12 +00:00
Chris Lattner
fb0b83a27b
Further simplifications
...
llvm-svn: 8321
2003-09-02 16:46:41 +00:00
Chris Lattner
dda7669997
Change the code to no longer compute the "type description" immediately when
...
the type is analyzed. Instead, only compute it when requested (with
getDescription), and cached for reuse later.
This dramatically speeds up LLVM in general because these descriptions almost
_never_ need to be constructed. The only time they are used is when a type is
<<'d. Printing of modules by themselves uses other code to print symbolic
types when possible, so these descriptions are really only used for debugging.
Also, this fixes the particularly bad case when lots of types get resolved to
each other, such as during linking of large programs. In these cases, the type
descriptions would be repeatedly recomputed and discarded even though: A. noone
reads the description before it gets resolved, and B. many many resolutions
happen at intermediate steps, causing a HUGE waste of time.
Overall, this makes the getTypeDesc function much more light-weight, and fixes
bug: Assembler/2002-07-08-HugePerformanceProblem.llx, which went from taking
1048.770u/19.150s (which is 17.5 MINUTES, on apoc), to taking 0.020u/0.000s,
which is a nice little speedup. :)
llvm-svn: 8320
2003-09-02 16:35:17 +00:00
Chris Lattner
1ae014f939
The description is no longer stored directly in the type.
...
llvm-svn: 8319
2003-09-02 16:28:03 +00:00
Brian Gaeke
5a8d15dfa1
Remove assertion which is never reached.
...
llvm-svn: 8318
2003-09-02 06:45:34 +00:00
Chris Lattner
f993c53f49
Dead header file
...
llvm-svn: 8317
2003-09-01 20:46:08 +00:00
Chris Lattner
7149fa4d4b
#include is unnecessary
...
llvm-svn: 8316
2003-09-01 20:45:46 +00:00
Chris Lattner
3fdc7bbd0d
Prototype for creator function got moved
...
llvm-svn: 8315
2003-09-01 20:45:33 +00:00
Chris Lattner
603a7af227
Add RPR prototype here
...
llvm-svn: 8314
2003-09-01 20:44:42 +00:00
Chris Lattner
8ec3783590
Remove dead file
...
llvm-svn: 8313
2003-09-01 20:41:21 +00:00
Chris Lattner
3d53c382ba
OptInfo is no longer required
...
llvm-svn: 8312
2003-09-01 20:41:07 +00:00
Chris Lattner
6d92264f5a
No longer provide an optinfo, noone uses it
...
llvm-svn: 8311
2003-09-01 20:40:59 +00:00
Chris Lattner
feebbdd6d1
No longer require an OptInfo
...
llvm-svn: 8310
2003-09-01 20:40:43 +00:00
Chris Lattner
2eb5de925f
Inline simple comparison which is sparc specific anyway
...
llvm-svn: 8309
2003-09-01 20:38:03 +00:00
Chris Lattner
5520e4dc30
Don't build dead directories
...
llvm-svn: 8308
2003-09-01 20:34:15 +00:00
Chris Lattner
87d223c70e
Remove header files that were privatized
...
llvm-svn: 8307
2003-09-01 20:33:38 +00:00
Chris Lattner
3d6b7d1b44
Move private interfaces into private .h file
...
llvm-svn: 8306
2003-09-01 20:33:07 +00:00
Chris Lattner
f2e4f5a388
Preselection is now integrated into the Sparc target library
...
llvm-svn: 8305
2003-09-01 20:30:17 +00:00
Chris Lattner
a685c382b1
Remove dead library makefile
...
llvm-svn: 8304
2003-09-01 20:29:43 +00:00
Chris Lattner
80c84d29b1
Remove makefile for dead library
...
llvm-svn: 8303
2003-09-01 20:26:52 +00:00
Chris Lattner
25669f3e09
Sparc peephole optimizer moved out of post-opts library into Sparc target library
...
llvm-svn: 8302
2003-09-01 20:26:14 +00:00
Chris Lattner
cd2980ec8c
Simplify code a bit
...
llvm-svn: 8301
2003-09-01 20:24:06 +00:00
Chris Lattner
aeb98733eb
This file is never #included
...
llvm-svn: 8300
2003-09-01 20:19:31 +00:00
Chris Lattner
049f0b8b37
LiveRange.h is now in lib/CodeGen/RegAlloc
...
llvm-svn: 8299
2003-09-01 20:17:13 +00:00
Chris Lattner
e5a47a9d95
This file is hopelessly out of date
...
llvm-svn: 8298
2003-09-01 20:14:00 +00:00
Chris Lattner
74e93a318b
LiveRangeInfo got moved into the lib/CodeGen/RegAlloc directory
...
llvm-svn: 8297
2003-09-01 20:12:17 +00:00
Chris Lattner
a591134169
PhyRegAlloc.h got moved to lib/CodeGen/RegAlloc
...
llvm-svn: 8296
2003-09-01 20:09:04 +00:00
Chris Lattner
e3fcb37a20
Move IGNode from public include directory to here. Minor cleanups like adding std:: namespace qualifiers
...
llvm-svn: 8295
2003-09-01 20:05:47 +00:00
Chris Lattner
8ebc13b6ca
IGNode got moved to lib/CodeGen/RegAlloc
...
llvm-svn: 8294
2003-09-01 20:00:08 +00:00
Chris Lattner
ee413804c7
This file just needs LiveRange.h not IGNode.h
...
llvm-svn: 8293
2003-09-01 19:58:02 +00:00
Chris Lattner
a43cd2a9c9
No longer include IGNode.h in the Sparc global header
...
llvm-svn: 8292
2003-09-01 19:56:48 +00:00
Chris Lattner
16149f6b85
Final cleanup: remove dead code
...
llvm-svn: 8291
2003-09-01 18:01:36 +00:00
Chris Lattner
b70da7293b
Not only is this a lot smaller, it actually works if there is already a
...
function with the right name in the module.
llvm-svn: 8290
2003-09-01 16:53:46 +00:00
Chris Lattner
3bef1d6ae4
Dead files
...
llvm-svn: 8289
2003-09-01 16:49:57 +00:00
Chris Lattner
8fd6602748
If "These should be used only by the auto-parallelization pass", we might as
...
well put them INTO the auto-par pass.
llvm-svn: 8288
2003-09-01 16:49:38 +00:00
Chris Lattner
d5f57a52db
Moved all of the cilkifier stuff into lib/Transforms/IPO, as it really is not
...
support stuff.
llvm-svn: 8287
2003-09-01 16:45:30 +00:00
Chris Lattner
ed0dcf3cf0
This file is just a subset of Cilkifier.h
...
llvm-svn: 8286
2003-09-01 16:42:43 +00:00
Chris Lattner
4ee52b7154
Minor cleanups
...
Do not #include Parallelize.h, it's just a subset of Cilkifier.h
llvm-svn: 8285
2003-09-01 16:42:16 +00:00
Chris Lattner
ef379b667e
Other minor cleanups while I'm in the area
...
llvm-svn: 8284
2003-09-01 16:38:43 +00:00
Chris Lattner
eaed5602a7
Fix spell-o
...
llvm-svn: 8283
2003-09-01 16:35:30 +00:00
Chris Lattner
4252e57c90
Remove gross old hacky code that was in there for backwards compatibility
...
1 year is plenty of migration time!
llvm-svn: 8282
2003-09-01 16:31:28 +00:00
Chris Lattner
2cdcc41517
Change the RaiseAllocations pass to be a Pass instead of a BasicBlock pass.
...
This makes it more efficient: it doesn't have to scan the whole program, so
it performs work proportional to the number of malloc/free calls in the
program, not the size of the program.
llvm-svn: 8280
2003-09-01 03:14:56 +00:00
Chris Lattner
2ef2a63728
Change the RaiseAllocations pass from being a BasicBlockPass to being a Pass
...
llvm-svn: 8279
2003-09-01 03:14:00 +00:00
Chris Lattner
1f9495ceab
Remove the -stopAfterNPasses option, which has been long obsoleted by bugpoint
...
llvm-svn: 8278
2003-08-31 21:47:24 +00:00
Chris Lattner
c2f8ebd598
* move the dead-type-eliminate passes to the end, where they are more useful
...
* add prune-eh and inlining passes
* other minor pass reorganizations
llvm-svn: 8277
2003-08-31 21:45:55 +00:00
Chris Lattner
c3019e3d22
Fix bug: TailDuplicate/2003-08-31-UnreachableBlocks.ll
...
llvm-svn: 8276
2003-08-31 21:17:44 +00:00
Chris Lattner
59db72924e
New testcase
...
llvm-svn: 8275
2003-08-31 21:05:39 +00:00
Chris Lattner
0620f3d2fa
Replace M with F when refering to functions
...
llvm-svn: 8274
2003-08-31 20:36:52 +00:00
Chris Lattner
f72da72785
Rename TarjanSCCIterator -> scc_iterator
...
* Increases consistency with other iterators (e.g. df_iterator, po_iterator...)
* It's shorter
* We don't name classes by the implementation, we name it for the interface!
llvm-svn: 8273
2003-08-31 20:01:57 +00:00
Chris Lattner
076533d11b
ELIMINATE the SCC class completely. One less thing deriving from std::vector
...
llvm-svn: 8272
2003-08-31 19:55:31 +00:00