Commit Graph

274 Commits

Author SHA1 Message Date
Reid Spencer
25319569fb - Type::TypeTyID doesn't exist any more (bug 122)
- Types don't have names any more, just write them on ostream directly

llvm-svn: 14606
2004-07-04 11:37:54 +00:00
Reid Spencer
5ef4237089 Move SlotCalculator.h to lib/Bytecode/Writer since that is the only place
that needs it after the Type != Value changes (bug 122).

llvm-svn: 14605
2004-07-04 11:36:31 +00:00
Chris Lattner
4ba8a8d1ef No functionality changes here:
* Some warning fixes for MSVC
  * Minor simplification to the deque scanning code

llvm-svn: 14417
2004-06-25 20:52:10 +00:00
Chris Lattner
ccd8ed145a Fix more warnings building with VC++
llvm-svn: 14391
2004-06-25 00:35:55 +00:00
Chris Lattner
521eb8752b fix warnings
llvm-svn: 14388
2004-06-25 00:11:25 +00:00
Chris Lattner
6b7275996c Rename Type::PrimitiveID to TypeId and ::getPrimitiveID() to ::getTypeID()
llvm-svn: 14201
2004-06-17 18:19:28 +00:00
Reid Spencer
e73febc64a Cleaned up a dead header file to prevent duplicate definition warnings
in doxygen.

llvm-svn: 14084
2004-06-09 04:38:34 +00:00
Reid Spencer
9163bcdf0f Moved this file to lib/Bytecode/Writer because its used there only.
llvm-svn: 13900
2004-05-30 00:09:32 +00:00
Reid Spencer
f2f34b3052 Use the SymbolTable::isEmpty() method instead of checking for no value
planes. A SymbolTable could still have types in it! This fixes problems
with two regression tests that failed because a symbol table that only
contained types was being omitted from bytecode files. Thanks to Chris
for the reduced test case that helped find this immediately.

llvm-svn: 13842
2004-05-27 20:18:51 +00:00
Reid Spencer
0aff01a239 Part of bug 122:
This change removes the BuildBytecodeInfo flag from the SlotCalculator
class. This flag was needed to distinguish between the Bytecode/Writer
and the AsmWriter. Now that AsmWriter doesn't use SlotCalculator, we can
remove this flag and simplify some code. Also, some minor name changes
to CachedWriter.h needed to be committed (missed in previous commit).

llvm-svn: 13785
2004-05-26 07:37:11 +00:00
Reid Spencer
abac0ef0b7 Put SlotTable.h inclusion back at front of list to be coding standards
compliant. Thanks, Chris.

llvm-svn: 13771
2004-05-25 20:09:05 +00:00
Reid Spencer
6244bc7831 Make the constructor explicit so we can't implicitly convert bool to
SlotTable.

llvm-svn: 13766
2004-05-25 19:09:25 +00:00
Reid Spencer
37dd30c00a Make some improvements suggested by Chris.
llvm-svn: 13765
2004-05-25 19:03:21 +00:00
Reid Spencer
4feb245a31 Adding the initial implementation of the SlotTable class. This class is
the Abstract Data Type that holds slot number values and associates them
with Type* and Value*. The SlotTable is simply the holder of the slot
numbers and provides a controlled interface for building the table. It does
not enforce any particular idiom or functionality for manipulating the slot
numbers.

This is part of bug_122. The SlotCalculator and SlotMachine classes will
follow.

llvm-svn: 13764
2004-05-25 18:44:51 +00:00
Reid Spencer
660ea5fadb Changed to use SymbolTable's new iteration interfaces.
llvm-svn: 13759
2004-05-25 17:29:59 +00:00
Chris Lattner
15701e84d1 Implement support for a new LLVM 1.3 bytecode format, which uses uint's
to index into structure types and allows arbitrary 32- and 64-bit integer
types to index into sequential types.

llvm-svn: 12651
2004-04-05 01:27:26 +00:00
Chris Lattner
e7c9f6f7ae Write select instructions to bytecode
llvm-svn: 12315
2004-03-12 05:52:01 +00:00
Chris Lattner
ac6db755c3 Adjust to the changed StructType interface. In particular, getElementTypes() is gone.
llvm-svn: 11228
2004-02-09 04:37:31 +00:00
Chris Lattner
fa829be4d3 Start using the new and improve interface to FunctionType arguments
llvm-svn: 11224
2004-02-09 04:14:01 +00:00
Chris Lattner
8c285b20b7 Remove all of the annoying statistics now that I'm finished (for the near
term) working on bytecode size stuff.

llvm-svn: 11046
2004-02-01 01:50:31 +00:00
Chris Lattner
c70b3f63cc SlotCalculator.h moved
llvm-svn: 10931
2004-01-20 19:50:34 +00:00
Chris Lattner
a395324383 Bugfixes for dealing with partially compactified functions
llvm-svn: 10920
2004-01-20 00:54:06 +00:00
Chris Lattner
5a66bb7507 Save another 30K from 176.gcc by encoding the compaction table a bit more
intelligently.

llvm-svn: 10918
2004-01-18 22:35:34 +00:00
Chris Lattner
979ca2fccd Remove -debug output
llvm-svn: 10917
2004-01-18 22:26:53 +00:00
Chris Lattner
bc02f4ce97 Add support for writing bytecode files with compactiontables 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.

llvm-svn: 10915
2004-01-18 21:08:52 +00:00
Chris Lattner
677af4a1cb 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.

llvm-svn: 10908
2004-01-17 23:25:43 +00:00
Chris Lattner
4c57267ece If these blocks are empty, there is no reason to even emit the bytecode blocks.
This saves about 15K in 176.gcc, coupled with another patch that I'm working on.

llvm-svn: 10889
2004-01-15 21:06:57 +00:00
Chris Lattner
f4572852dc ConstantPointerRef's are no longer emitted. This saves 20028 bytes in the
bytecode files when compiling 176.gcc, but more importantly will make it
easier to eliminate CPR's in the future (no new .bc revision will be
required to support them)

llvm-svn: 10884
2004-01-15 18:46:56 +00:00
Chris Lattner
394fc96f39 Fix PR73: bytecode format inconsistent
llvm-svn: 10876
2004-01-15 17:55:09 +00:00
Chris Lattner
1af644d9d6 Remove the optimization that depends on ENDIAN_LITTLE: it's not worth it.
llvm-svn: 10870
2004-01-15 06:11:30 +00:00
Chris Lattner
6229fbf328 The new bytecode format supports emitting strings a special case. This is
intended to save size (and does on small programs), but on big programs it
actually increases the size of the program slightly.  The deal is that many
functions end up using the characters that the string contained, and the
characters are no longer in the global constant table, so they have to be
emitted in function specific constant pools.

This pessimization will be fixed in subsequent patches.

llvm-svn: 10864
2004-01-14 23:36:54 +00:00
Chris Lattner
eeb4ffd8ea It is obvious that this has never been used for outputing more than a single
byte, it's totally endian incorrect!

llvm-svn: 10857
2004-01-14 18:50:28 +00:00
Chris Lattner
29f9590e82 Be const correct
llvm-svn: 10856
2004-01-14 18:41:38 +00:00
Chris Lattner
9a38c78fe6 Ok, I can't handle it. This is a temporary checkin of a ton of statistics that
i'm using in my work to reduce the bytecode file sizes.  These will eventually
be removed.

llvm-svn: 10849
2004-01-14 16:54:21 +00:00
Chris Lattner
1f018c2d01 Toggle sense of flag
llvm-svn: 10838
2004-01-14 02:50:16 +00:00
Chris Lattner
a2bfab849f Do not bother to emit a BytecodeBlock for an empty symbol table. This commonly
occurs when the symbol table for a module has been stripped, making all of the
function local symbols go away.

This saves 6728 bytes in the stripped bytecode file of 254.gap (which obviously
has 841 functions), which isn't a ton, but helps and was easy.

llvm-svn: 10750
2004-01-10 19:56:59 +00:00
Chris Lattner
0c53031c5e Err, we don't need Config/*.h files for things that are standard C++
llvm-svn: 10742
2004-01-10 19:10:01 +00:00
Chris Lattner
b0a5964356 minor comment tweaks
llvm-svn: 10741
2004-01-10 19:07:06 +00:00
Chris Lattner
82167909a8 Refactor writer stuff out of include/llvm/Bytecode/Primitives.h. This is
internal implementation details for the writer, not public interfaces!

llvm-svn: 10738
2004-01-10 18:56:59 +00:00
Chris Lattner
dfe0346e96 Finegrainify namespacification
llvm-svn: 10737
2004-01-10 18:49:43 +00:00
Chris Lattner
d9cf9b30eb Emit & read more compressed bytecode by not emitting a bytecodeblock for
each basic block in function.  Instead, just emit a stream of instructions,
chopping up basic blocks based on when we find terminator instructions.  This
saves a fairly substantial chunk of bytecode space.  In stripped, sample
cases, for example, we get this reduction in size:

197.parser: 163036 -> 137180:  18.8% reduction
254.gap   : 844936 -> 689392:  22.6%
255.vortex: 621724 -> 528444:  17.7%

...

Not bad for something this simple.  :)  Note that this doesn't require a new
bytecode version number at all, though version 1.1 should not need to support
the old format.

llvm-svn: 10280
2003-12-01 07:05:31 +00:00
Chris Lattner
ea25538936 Remove usage of ConstantPointer
llvm-svn: 10051
2003-11-17 17:28:29 +00:00
Brian Gaeke
960707c335 Put all LLVM code into the llvm namespace, as per bug 109.
llvm-svn: 9903
2003-11-11 22:41:34 +00:00
Brian Gaeke
45eb2eeb74 Explicitly cast type, so we aren't passing output_vbr a size_t.
llvm-svn: 9590
2003-10-29 20:09:01 +00:00
John Criswell
29265fe981 Added LLVM copyright header.
llvm-svn: 9321
2003-10-21 15:17:13 +00:00
John Criswell
4436c49787 Added LLVM copyright notice to Makefiles.
llvm-svn: 9312
2003-10-20 22:26:57 +00:00
John Criswell
482202a601 Added LLVM project notice to the top of every C++ source file.
Header files will be on the way.

llvm-svn: 9298
2003-10-20 19:43:21 +00:00
Chris Lattner
dc83293a38 add support for new linkage types
llvm-svn: 9228
2003-10-18 06:30:21 +00:00
Chris Lattner
9d57347490 Updated to emit the final 1.0 bytecode format. This supports weak linkage,
more efficient encoding of varargs calls, and the new varargs intrinsics.

llvm-svn: 9221
2003-10-18 05:54:48 +00:00
Alkis Evlogimenos
8faf8d93d5 Eliminate some extraneous code in SlotCalculator::insertVal().
Rename SlotCalculator::getValSlot() to SlotCalculator::getSlot(),
       SlotCalculator::insertValue() to SlotCalculator::getOrCreateSlot(),
       SlotCalculator::insertVal() to SlotCalculator::insertValue(), and
       SlotCalculator::doInsertVal() to SlotCalculator::doInsertValue().

llvm-svn: 9190
2003-10-17 02:02:40 +00:00