Changes in directory llvm/lib/VMCore:
ConstantFold.cpp updated: 1.146 -> 1.147
---
Log message:
Fix constant folding of fp->int bitcast for vectors.
---
Diffs of the changes: (+2 -2)
ConstantFold.cpp |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Index: llvm/lib/VMCore/Con
Changes in directory llvm/include/llvm/ADT:
APInt.h updated: 1.38 -> 1.39
---
Log message:
Regularize the interface for conversion functions to/from float/double.
---
Diffs of the changes: (+9 -3)
APInt.h | 12 +---
1 files changed, 9 insertions(+), 3 deletions(-)
Index: llvm/i
Changes in directory llvm/include/llvm/ExecutionEngine:
ExecutionEngine.h updated: 1.41 -> 1.42
---
Log message:
Require the destination GlobalValue for LoadValueFromMemory to be passed
in as a parameter instead of returned.
---
Diffs of the changes: (+2 -1)
ExecutionEngine.h |3 ++-
1
Changes in directory llvm/lib/ExecutionEngine:
ExecutionEngine.cpp updated: 1.108 -> 1.109
---
Log message:
Avoid memory leakage by having caller construct the APInt for the
destination value of LoadValueFromMemory.
---
Diffs of the changes: (+4 -5)
ExecutionEngine.cpp |9 -
1
Changes in directory llvm/lib/ExecutionEngine/Interpreter:
Execution.cpp updated: 1.171 -> 1.172
---
Log message:
Complete the APIntification of the interpreter. All asserts for > 64 bits
have been removed and dealt with. The interpreter should now be able to
execute any LLVM program using any
> ...(3) having gcc add one. For C, (3) involves adding one to a
> constant, which does not involve any heavy operations like memory
> allocation, just a two word integer plus operation.
This is wrong: adding one of course requires allocating a tree node in
general (i.e. unless a previously const
Chris Lattner wrote:
>
>> Because R31 is not used as a callee saved register when
>> llvm::NoFramePointerElim is set to 1.
>
> NoFramePointerElim doesn't change the ABI, only the codegen model.
Yes it doesn't change the ABI, but R31's offset in Linux/ELF is in the
callee saved area. In Macho it
Changes in directory llvm/lib/Target/CBackend:
CBackend.cpp updated: 1.328 -> 1.329
---
Log message:
Make sure that when we store a value it is masked to its correct bit
width. This helps CBE work with non-standard integer bit widths.
---
Diffs of the changes: (+15 -1)
CBackend.cpp | 16
Changes in directory llvm/include/llvm/ExecutionEngine:
ExecutionEngine.h updated: 1.42 -> 1.43
---
Log message:
Make the creation functions take an error string. This will help the
ExecutionEngine report errors on creation (like bytecode stuff) and also
help to get rid of exception handling.
Changes in directory llvm/lib/ExecutionEngine:
ExecutionEngine.cpp updated: 1.109 -> 1.110
---
Log message:
Deal with error handling better.
---
Diffs of the changes: (+5 -3)
ExecutionEngine.cpp |8 +---
1 files changed, 5 insertions(+), 3 deletions(-)
Index: llvm/lib/ExecutionEn
Changes in directory llvm/lib/ExecutionEngine/JIT:
JIT.h updated: 1.30 -> 1.31
TargetSelect.cpp updated: 1.13 -> 1.14
---
Log message:
Deal with error handling better.
---
Diffs of the changes: (+10 -3)
JIT.h|2 +-
TargetSelect.cpp | 11 +--
2 files changed, 10 in
Changes in directory llvm/lib/ExecutionEngine/Interpreter:
Interpreter.cpp updated: 1.36 -> 1.37
Interpreter.h updated: 1.84 -> 1.85
---
Log message:
Deal with error handling better.
---
Diffs of the changes: (+16 -8)
Interpreter.cpp | 22 +++---
Interpreter.h |2 +-
Changes in directory llvm/tools/lli:
lli.cpp updated: 1.67 -> 1.68
---
Log message:
1. Handle errors around the ModuleProvider. This is necessary since it is
reading bytecode.
2. The interpreter can delete the ModuleProvider and replace it with
another so don't depend on it being around
Changes in directory llvm/lib/ExecutionEngine/Interpreter:
Interpreter.cpp updated: 1.37 -> 1.38
---
Log message:
Remove unnecessary #include.
---
Diffs of the changes: (+0 -1)
Interpreter.cpp |1 -
1 files changed, 1 deletion(-)
Index: llvm/lib/ExecutionEngine/Interpreter/Interprete
Changes in directory llvm/lib/CodeGen/SelectionDAG:
LegalizeDAG.cpp updated: 1.480 -> 1.481
---
Log message:
Add an expand action for ISD label which just deletes the label.
This "fixes" PR1238: http://llvm.org/PR1238 .
---
Diffs of the changes: (+3 -0)
LegalizeDAG.cpp |3 +++
1 files
On Mar 2, 2007, at 11:37 PM, Reid Spencer wrote:
> Changes in directory llvm/include/llvm/ExecutionEngine:
> GenericValue.h updated: 1.9 -> 1.10
> ---
> Log message:
>
> Add APIntVal as a possible GenericeValue.
Shouldn't APInt eliminate all of the other, smaller, integers?
-Chris
> ---
> Diffs
Changes in directory llvm/lib/Transforms/Scalar:
IndVarSimplify.cpp updated: 1.111 -> 1.112
---
Log message:
Make RewriteLoopExitValues far less nested by using continue in the loop
---
Diffs of the changes: (+94 -91)
IndVarSimplify.cpp | 185 ++
Revision: 124528
Author: clattner
Date: 2007-03-03 15:08:02 -0800 (Sat, 03 Mar 2007)
Log Message:
---
Fix an assertion I recently added, POINTER_TYPE and REFERENCE_TYPE
are the same at the treetollvm level. This fixes an ICE on 447.dealII.
Modified Paths:
--
apple-l
Changes in directory llvm/test/CodeGen/Generic:
fpowi-promote.ll added (r1.1)
---
Log message:
New testcase for PR1239: http://llvm.org/PR1239
---
Diffs of the changes: (+12 -0)
fpowi-promote.ll | 12
1 files changed, 12 insertions(+)
Index: llvm/test/CodeGen/Generic/fpow
Changes in directory llvm/lib/CodeGen/SelectionDAG:
LegalizeDAG.cpp updated: 1.481 -> 1.482
---
Log message:
Fix CodeGen/Generic/fpowi-promote.ll and PR1239: http://llvm.org/PR1239
---
Diffs of the changes: (+12 -0)
LegalizeDAG.cpp | 12
1 files changed, 12 insertions(+)
Hello, Everyone.
Please find updated LowerSwitch patch (synced with latest APInt-related
changes).
--
With best regards, Anton Korobeynikov.
Faculty of Mathematics & Mechanics, Saint Petersburg State University.
diff -r 1f58690658d2 lib/Transforms/Utils/LowerSwitch.cpp
--- a/lib/Transforms/Uti
Changes in directory llvm/docs:
Makefile updated: 1.12 -> 1.13
---
Log message:
Get the version number for the LLVM from the autoconf/configure.ac file
when building the documentation for the web site.
---
Diffs of the changes: (+9 -2)
Makefile | 11 +--
1 files changed, 9 insert
Changes in directory llvm/test/Transforms/IndVarsSimplify:
loop_evaluate_2.ll added (r1.1)
loop_evaluate_3.ll added (r1.1)
loop_evaluate_4.ll added (r1.1)
---
Log message:
New testcases for PR1179: http://llvm.org/PR1179 /PR1232:
http://llvm.org/PR1232 .
---
Diffs of the changes: (+67 -0)
On Sat, 2007-03-03 at 13:07 -0800, Chris Lattner wrote:
> On Mar 2, 2007, at 11:37 PM, Reid Spencer wrote:
> > Changes in directory llvm/include/llvm/ExecutionEngine:
> > GenericValue.h updated: 1.9 -> 1.10
> > ---
> > Log message:
> >
> > Add APIntVal as a possible GenericeValue.
>
> Shouldn't AP
On Sat, 2007-03-03 at 16:22 -0800, Chris Lattner wrote:
> On Mar 3, 2007, at 4:16 PM, Reid Spencer wrote:
> > On Sat, 2007-03-03 at 13:07 -0800, Chris Lattner wrote:
> >> On Mar 2, 2007, at 11:37 PM, Reid Spencer wrote:
> >>> Changes in directory llvm/include/llvm/ExecutionEngine:
> >>> GenericValu
Changes in directory llvm/lib/Transforms/Scalar:
IndVarSimplify.cpp updated: 1.112 -> 1.113
---
Log message:
Implement PR1179: http://llvm.org/PR1179 /PR1232: http://llvm.org/PR1232 and
test/Transforms/IndVarsSimplify/loop_evaluate_[234].ll
This makes -indvars require and use LCSSA, updating
On Mar 3, 2007, at 4:16 PM, Reid Spencer wrote:
> On Sat, 2007-03-03 at 13:07 -0800, Chris Lattner wrote:
>> On Mar 2, 2007, at 11:37 PM, Reid Spencer wrote:
>>> Changes in directory llvm/include/llvm/ExecutionEngine:
>>> GenericValue.h updated: 1.9 -> 1.10
>>> ---
>>> Log message:
>>>
>>> Add APIn
Changes in directory llvm/lib/Analysis:
ScalarEvolution.cpp updated: 1.104 -> 1.105
---
Log message:
Guard further against APInt operations with operands of unequal bit width.
---
Diffs of the changes: (+11 -4)
ScalarEvolution.cpp | 15 +++
1 files changed, 11 insertions(+),
Changes in directory llvm/lib/Transforms/Scalar:
IndVarSimplify.cpp updated: 1.113 -> 1.114
---
Log message:
make better use of LCSSA information in RewriteLoopExitValues. Before, we
would scan the entire loop body, then scan all users of instructions in the
loop, looking for users outside the
>> Ok, could we make GV not be a union? Or a struct with an APInt and
>> union of other things?
>
> I thought about that, and started to implement it. That involves a 50%
> increase in memory consumption. Its already a big enough hog.
It would increase the memory use of the *interpreter*, but not
On Sat, 2007-03-03 at 19:45 -0800, Chris Lattner wrote:
> >> Ok, could we make GV not be a union? Or a struct with an APInt and
> >> union of other things?
> >
> > I thought about that, and started to implement it. That involves a 50%
> > increase in memory consumption. Its already a big enough ho
Changes in directory llvm/include/llvm/ADT:
SmallPtrSet.h updated: 1.8 -> 1.9
---
Log message:
add iterator range version of ctor.
---
Diffs of the changes: (+7 -0)
SmallPtrSet.h |7 +++
1 files changed, 7 insertions(+)
Index: llvm/include/llvm/ADT/SmallPtrSet.h
diff -u llvm/incl
Changes in directory llvm/lib/Analysis:
LoopInfo.cpp updated: 1.82 -> 1.83
---
Log message:
Speed up Loop::isLCSSAForm by using a hash table instead of a sorted vector.
---
Diffs of the changes: (+7 -9)
LoopInfo.cpp | 16 +++-
1 files changed, 7 insertions(+), 9 deletions(-)
On Mar 3, 2007, at 7:50 PM, Reid Spencer wrote:
> On Sat, 2007-03-03 at 19:45 -0800, Chris Lattner wrote:
Ok, could we make GV not be a union? Or a struct with an APInt and
union of other things?
>>>
>>> I thought about that, and started to implement it. That involves
>>> a 50%
>>> inc
Changes in directory llvm/lib/Transforms/Scalar:
SimplifyCFG.cpp updated: 1.19 -> 1.20
---
Log message:
switch MarkAliveBlocks over to using SmallPtrSet instead of std::set, speeding
up simplifycfg by 20%
---
Diffs of the changes: (+5 -5)
SimplifyCFG.cpp | 10 +-
1 files changed
On Sat, 2007-03-03 at 20:11 -0800, Chris Lattner wrote:
> On Mar 3, 2007, at 7:50 PM, Reid Spencer wrote:
> > On Sat, 2007-03-03 at 19:45 -0800, Chris Lattner wrote:
> Ok, could we make GV not be a union? Or a struct with an APInt and
> union of other things?
> >>>
> >>> I thought about
Changes in directory llvm/lib/Transforms/Scalar:
InstructionCombining.cpp updated: 1.642 -> 1.643
---
Log message:
Speed up -instcombine by 20% by avoiding a particularly expensive passmgr call.
---
Diffs of the changes: (+4 -1)
InstructionCombining.cpp |5 -
1 files changed, 4 ins
Changes in directory llvm/lib/CodeGen:
MachOWriter.cpp updated: 1.31 -> 1.32
---
Log message:
Remove unneeded header file.
---
Diffs of the changes: (+0 -1)
MachOWriter.cpp |1 -
1 files changed, 1 deletion(-)
Index: llvm/lib/CodeGen/MachOWriter.cpp
diff -u llvm/lib/CodeGen/MachOWrit
Changes in directory llvm/lib/Transforms/Scalar:
SCCP.cpp updated: 1.163 -> 1.164
---
Log message:
minor cleanups
---
Diffs of the changes: (+8 -7)
SCCP.cpp | 15 ---
1 files changed, 8 insertions(+), 7 deletions(-)
Index: llvm/lib/Transforms/Scalar/SCCP.cpp
diff -u llvm/li
It would increase the memory use of the *interpreter*, but nothing
else. The interpreter is already slow.
>>>
>>> The JIT makes use of the fields you're intending to delete as
>>> well. SO
>>> does ExecutionEngine (on each load and store). Sure you want to
>>> sacrifice JIT speed too?
Changes in directory llvm/lib/Target/X86:
X86InstrSSE.td updated: 1.177 -> 1.178
---
Log message:
add missing braces
---
Diffs of the changes: (+25 -25)
X86InstrSSE.td | 50 +-
1 files changed, 25 insertions(+), 25 deletions(-)
Index: llv
Changes in directory llvm/lib/System:
Disassembler.cpp updated: 1.1 -> 1.2
---
Log message:
fix typo
---
Diffs of the changes: (+2 -2)
Disassembler.cpp |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
Index: llvm/lib/System/Disassembler.cpp
diff -u llvm/lib/System/Disassembl
This patch eliminates Annotations by using an "analysis" pass as a bucket to hold onto the MachineFunction generated by instruction selection. PassManager then releases the memory at the appropriate time. I find this approach is rather fragile (the caveats are noted in the header file), but it's mu
On Mar 3, 2007, at 4:11 PM, Anton Korobeynikov wrote:
Hello, Everyone.
Please find updated LowerSwitch patch (synced with latest APInt-
related
changes).
Very interesting! Can you give an example with before/after code?
Please add this as a testcase as well. Some code comments:
Meta-c
44 matches
Mail list logo
