Commit Graph

789 Commits

Author SHA1 Message Date
Chris Lattner
90fb19ed2c Fix a problem brian ran into with the bytecode reader asserting. It turns
out that the problem was actually the writer writing out a 'null' value
because it didn't normalize it.  This fixes:
test/Regression/Assembler/2004-01-22-FloatNormalization.ll


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10967 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-23 00:55:21 +00:00
Misha Brukman
ce218c6492 Let subclasses implement ModuleProvider::materializeModule() which is based on
their implementation of book-keeping for which functions need to be materialized
and which don't.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10943 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-21 22:54:50 +00:00
Chris Lattner
f2d577b27a SlotCalculator.h moved
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10931 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-20 19:50:34 +00:00
Chris Lattner
68e3dbc493 Major changes. Now we only compactify individual type planes if it is in
fact "profitable" to do so.  This makes compactification "free" for small
programs (ie, it is completely disabled) and even helps large programs by
not having to encode pointless compactification planes.

On 176.gcc, this saves 50K from the bytecode file, which is, alas only
a couple percent.

This concludes my head bashing against the bytecode format, at least for
now.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10922 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-20 00:57:32 +00:00
Chris Lattner
614cdcd002 Add support for building the compactiontable for bytecode files. This shrinks
the bytecode file for 176.gcc by about 200K (10%), and 254.gap by about 167K,
a 25% reduction.  There is still a lot of room for improvement in the encoding
of the compaction table.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10913 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-18 21:07:07 +00:00
Chris Lattner
2ff95b6b4e Eliminate special case handling for CPR's
Fix some problem cases where I was building the slot calculator in bytecode
writer mode instead of asmwriter mode.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10911 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-18 21:03:06 +00:00
Chris Lattner
80b97343ec Bytecode format for LLVM 1.2 no longer explicitly encodes zeros in primitive
type planes.  This saves about 5k on 176.gcc, and is needed for a subsequent
patch of mine I'm working on.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10908 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-17 23:25:43 +00:00
Chris Lattner
d5c59d5c84 Cleanups & efficiency improvements
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10888 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-15 20:24:09 +00:00
Chris Lattner
8c202cdbda The bcwriter does not want ConstantPointerRef's to be indexed, and the asmwriter never did!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10885 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-15 18:47:15 +00:00
Chris Lattner
dcea63026f When emitting bytecode, handle strings specially. Do not add the characters
that make up the strings to the slotcalculator.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10862 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 23:34:39 +00:00
Chris Lattner
f907bac068 Hrm, apparently I missed lowering this intrinsic. :(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10858 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 20:41:29 +00:00
Chris Lattner
7aa162b29b Fix bug in previous checkin: Demorgan would be ashamed of me, I need to
lay off the crack.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10855 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 17:51:53 +00:00
Chris Lattner
13cfdea1ef Implement ConstantArray::isString
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10851 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 17:06:38 +00:00
Chris Lattner
c9b0702a89 Fix check. PHI nodes must be handled specially, of course.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10842 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 05:42:52 +00:00
Chris Lattner
30768ac3c2 Tighten up verifier checks. The result of an invoke instruction only
dominates the normal destination, not the exceptional dest (ie, the result
of a call is undefined on an exception)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10841 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 04:25:59 +00:00
Chris Lattner
8ce750145d The only clients of the slot calculator are now the asmwriter and bcwriter.
Since this really only makes sense for these two, change hte instance variable
to reflect whether we are writing a bytecode file or not.  This makes it
reasonable to add bcwriter specific stuff to it as necessary.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10837 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-14 02:49:34 +00:00
Chris Lattner
504e8fb74e Implement a bunch of symbolic constant folding opportunities. This implements
testcase test/Regression/Assembler/ConstantExprFold.llx

Note that these kinds of things only rarely show up in source code, but are
exceedingly common in the intermediate stages of algorithms like SCCP.  By
folding things (especially relational operators) that use symbolic constants,
we are able to speculatively fold more conditional branches, which can
lead to some big simplifications.

It would be easy to add a lot more special cases here, so if you notice
SCCP missing anything "obvious", you know what to make smarter.  :)


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10812 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-13 05:51:55 +00:00
Chris Lattner
3a534f200a Implement new Instruction::isRelational method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10810 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 23:18:25 +00:00
Chris Lattner
eab20b58b1 Rearrange and comment code better. No functionality changes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10808 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 22:07:24 +00:00
Chris Lattner
cbfd406541 Rename ConstantHandling.* -> ConstantFolding.*
Move a bunch of (now) private stuff from ConstantFolding.h into
ConstantFolding.cpp.

This _finally_ gets us to a place where we have a sane constant folder.  The
rules are:

1. LLVM clients now use ConstantExpr::get* methods to fold constants.  If they
   cannot be folded, a constantexpr is created, so these methods always return
   valid Constant*'s.
2. The implementation of ConstantExpr::get* uses the functions exposed by
   ConstantFolding.h to try to fold constants.  If they cannot be folded,
   they should return a null pointer.
3. The implementation of ConstantFolding can do whatever it wants, and only
   has one client (Constants.cpp)

This cuts down on the wierd dependencies, and eliminates the two interfaces.
The old constanthandling interface was especially bad for clients to use
because almost none of them took the failure condition into consideration,
thus leading to obscure problems.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10807 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 21:13:12 +00:00
Chris Lattner
8b0f0cb908 Remove a whole bunch more ugliness. This is actually getting to the point of
this whole refactoring: allow constant folding methods to return something
other than predefined classes, allow them to return generic Constant*'s.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10806 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 21:02:29 +00:00
Chris Lattner
6b7b91ab03 Move ConstantHandling.h into lib/VMCore and out of include/llvm
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10805 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 20:48:11 +00:00
Chris Lattner
731ba7f032 Eliminate a lot of out-of-date comments, and all of the wierd overloaded
operator constant folding stuff.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10803 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 20:43:44 +00:00
Chris Lattner
4a0ecc8c0c Eliminate usage of the wierd overloaded operator constant folders
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10802 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 20:41:05 +00:00
Chris Lattner
d4403b4c3a Eliminate ConstantFoldShiftInstruction reference
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10801 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 20:40:42 +00:00
Chris Lattner
67bb7603ea Remove use of ConstantHandling itf
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10800 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 20:13:04 +00:00
Chris Lattner
69b9e892b6 Fix out of date comment, remove use of ConstantExpr::getShift
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10794 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 19:12:58 +00:00
Chris Lattner
c6646ebdb9 Remove use of ConstantExpr::getShift
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10792 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 19:10:58 +00:00
Chris Lattner
3edd3979e1 Make ConstantExpr::get work for shifts as well
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10790 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 19:04:55 +00:00
Chris Lattner
c4a1509829 Move prototype of llvm::ConstantFoldInstruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10786 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 18:25:44 +00:00
Chris Lattner
28aabc34f7 Move llvm::ConstantFoldInstruction from here to Transforms/Utils
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10784 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-12 18:25:03 +00:00
Chris Lattner
952454dbff Implement: Assembler/2004-01-11-getelementptrfolding.llx
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10759 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-11 23:56:33 +00:00
Chris Lattner
92fa44c1cf Urg, remove testing code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10757 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-11 23:30:03 +00:00
Chris Lattner
9380297bc0 Fix a regression that I introduced yesterday. :(
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10756 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-11 23:29:26 +00:00
Chris Lattner
a14b0d441d Check in two changes:
The first change (which is disabled) compactifies all of the function constant
pools into the global constant pool, in an attempt to reduce the amount of
duplication and overhead.  Unfortunately, as the comment indicates, this is
not yet a win, so it is disabled.

The second change sorts the typeid's so that those types that can be used
by instructions in the program appear earlier in the table than those that
cannot (such as structures and arrays).  This causes the instructions to
be able to use the dense encoding more often, saving about 5K on 254.gap.
This is only a .65% savings though, unfortunately. :(


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10754 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-10 23:46:13 +00:00
Chris Lattner
91d7efbf4b Hrm, another minor cleanup, which I missed before
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10753 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-10 21:42:24 +00:00
Chris Lattner
9a0817971a Minor cleanup
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10752 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-10 21:40:29 +00:00
Chris Lattner
f7703df496 Finegrainify namespacification
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10727 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-09 06:12:26 +00:00
Chris Lattner
90ab32f9f0 Add support for new intrinsic
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10701 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-06 05:33:02 +00:00
Chris Lattner
77b1330ece Add VMCore and code generator support for debugging intrinsics. By default
code generators completely ignore them.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10691 91177308-0d34-0410-b5e6-96231b3b80d8
2004-01-05 05:36:30 +00:00
Chris Lattner
60837821e2 Add some comments, add new getGlobalVariable method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10671 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-31 08:43:01 +00:00
Chris Lattner
6d28a268c3 * Make Module::getTypeName const
* Add new Module::getTypeByName method
* Group methods in Module.cpp better


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10668 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-31 07:09:33 +00:00
Chris Lattner
69284b03d1 Make the lookup method const.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10667 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-31 07:08:19 +00:00
Chris Lattner
3b41e0e6f4 Fix PR198
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10659 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-31 03:19:37 +00:00
Chris Lattner
4b1be101fb Add comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10658 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-31 02:50:02 +00:00
Chris Lattner
f70e0c216c Clean up a lot of the code I added yesterday by exposing the IntrinsicLowering
implementation from the TargetMachine directly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10636 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-28 21:23:38 +00:00
Chris Lattner
42450d8a5f Move into the VMCore library
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10623 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-28 08:30:20 +00:00
Chris Lattner
3b66ecb05f Implement the default implementation of the intrinsic lowering class
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10621 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-28 08:19:41 +00:00
Chris Lattner
2cdd21c2e4 Finegrainify namespacification
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10464 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-14 21:35:53 +00:00
Chris Lattner
eae45cf44b The recalclulate method was a nasty hack that was once used by the -cee pass,
which never worked itself.  The cee pass still doesn't work, but it doesn't use
this method anymore anyway, so eliminate the method.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@10302 91177308-0d34-0410-b5e6-96231b3b80d8
2003-12-07 00:55:32 +00:00