Kay Tiong Khoo [Wed, 13 Jun 2012 15:53:04 +0000 (15:53 +0000)]
*typo: Cyles changed to Cycles
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158404
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 13 Jun 2012 13:30:24 +0000 (13:30 +0000)]
Use LTO_CODEGEN_PIC_MODEL_DYNAMIC for PIE. This requirest a git version of
gold to work. Since the enum value LDPO_PIE has just been added to plugin-api.h,
use a numeric constant for now so that we don't require an unreleased
version of gold to build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158402
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Wed, 13 Jun 2012 12:15:56 +0000 (12:15 +0000)]
It is possible for several constants which aren't individually absorbing to
combine to the absorbing element. Thanks to nbjoerg on IRC for pointing this
out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158399
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Wed, 13 Jun 2012 09:42:13 +0000 (09:42 +0000)]
When linearizing a multiplication, return at once if we see a factor of zero,
since then the entire expression must equal zero (similarly for other operations
with an absorbing element). With this in place a bunch of reassociate code for
handling constants is dead since it is all taken care of when linearizing. No
intended functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158398
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 13 Jun 2012 07:18:53 +0000 (07:18 +0000)]
Fix intrinsics for XOP frczss/sd instructions. These instructions only take one source register and zero the upper bits of the destination rather than preserving them.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158396
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Wed, 13 Jun 2012 05:55:09 +0000 (05:55 +0000)]
Add another missing 64-bit itinerary definition for the PPC A2 core.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158393
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Wed, 13 Jun 2012 05:43:29 +0000 (05:43 +0000)]
SimplifyCFG: fold unconditional branch to its predecessor if profitable.
This patch extends FoldBranchToCommonDest to fold unconditional branches.
For unconditional branches, we fold them if it is easy to update the phi nodes
in the common successors.
rdar://
10554090
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158392
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Wed, 13 Jun 2012 05:15:49 +0000 (05:15 +0000)]
Eliminate struct TableGenBackend.
TableGen backends are simply written as functions now.
Patch by Sean Silva!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158389
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Wed, 13 Jun 2012 02:42:47 +0000 (02:42 +0000)]
Clean up trailing blanks in Mips16InstrFormats.td
Patch by Reed Kotler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158382
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Wed, 13 Jun 2012 02:41:14 +0000 (02:41 +0000)]
disable use of directive .set nomicromips
until this directive is pushed in gas to open source fsf
Patch by Reed Kotler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158381
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 13 Jun 2012 02:39:03 +0000 (02:39 +0000)]
sched: fix latency of memory dependence chain edges for consistency.
For store->load dependencies that may alias, we should always use
TrueMemOrderLatency, which may eventually become a subtarget hook. In
effect, we should guarantee at least TrueMemOrderLatency on at least
one DAG path from a store to a may-alias load.
This should fix the standard mode as well as -enable-aa-sched-mi".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158380
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Wed, 13 Jun 2012 02:39:00 +0000 (02:39 +0000)]
sched: Avoid trivially redundant DAG edges. Take the one with higher latency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158379
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Wed, 13 Jun 2012 02:37:54 +0000 (02:37 +0000)]
1. fix places where immed is used in place of imm to be consistent with
non mips16
2. fix some comments to change OPcode->EXTEND for extended instructions
Patch by Reed Kotler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158378
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 12 Jun 2012 20:32:29 +0000 (20:32 +0000)]
Add some missing 64-bit itinerary definitions for the PPC A2 core.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158373
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 12 Jun 2012 20:26:43 +0000 (20:26 +0000)]
Use DenseMap as SmallMap workaround rather than std::map, at Chandler's request.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158371
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 12 Jun 2012 20:16:51 +0000 (20:16 +0000)]
Use std::map rather than SmallMap because SmallMap assumes that the value has
POD type, causing memory corruption when mapping to APInts with bitwidth > 64.
Merge another crash testcase into crash.ll while there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158369
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Tue, 12 Jun 2012 19:25:13 +0000 (19:25 +0000)]
[arm-fast-isel] Add support for -arm-long-calls.
Patch by Jush Lu <jush.msn@gmail.com>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158368
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 12 Jun 2012 19:01:24 +0000 (19:01 +0000)]
Split out the PPC instruction class IntSimple from IntGeneral.
On the POWER7, adds and logical operations can also be handled
in the load/store pipelines. We'll call these IntSimple.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158366
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Tue, 12 Jun 2012 17:06:32 +0000 (17:06 +0000)]
Remove use of GNU extension to resolve Clang warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158364
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 12 Jun 2012 16:39:23 +0000 (16:39 +0000)]
Fixes for PPC host detection and features.
POWER4 is a 64-bit CPU (better matched to the 970).
The g3 is really the 750 (no altivec), the g4+ is the 74xx (not the 750).
Patch by Andreas Tobler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158363
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Tue, 12 Jun 2012 15:45:07 +0000 (15:45 +0000)]
Use correct syntax highliter in code blocks. Noticed by Sean Silva.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158359
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Tue, 12 Jun 2012 14:33:56 +0000 (14:33 +0000)]
Now that Reassociate's LinearizeExprTree can look through arbitrary expression
topologies, it is quite possible for a leaf node to have huge multiplicity, for
example: x0 = x*x, x1 = x0*x0, x2 = x1*x1, ... rapidly gives a value which is x
raised to a vast power (the multiplicity, or weight, of x). This patch fixes
the computation of weights by correctly computing them no matter how big they
are, rather than just overflowing and getting a wrong value. It turns out that
the weight for a value never needs more bits to represent than the value itself,
so it is enough to represent weights as APInts of the same bitwidth and do the
right overflow-avoiding dance steps when computing weights. As a side-effect it
reduces the number of multiplies needed in some cases of large powers. While
there, in view of external uses (eg by the vectorizer) I made LinearizeExprTree
static, pushing the rank computation out into users. This is progress towards
fixing PR13021.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158358
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 12 Jun 2012 04:21:36 +0000 (04:21 +0000)]
Add two newlines in ParseSubtargetFeatures's debug output after the CPU is printed.
There is otherwise not a newline between the CPU name and the start of the next
pass's output which makes both difficult to read.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158350
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 12 Jun 2012 03:03:13 +0000 (03:03 +0000)]
Reapply r158337, this time properly protect Darwin/PPC host CPU use with __ppc__.
Original commit message:
Move PPC host-CPU detection logic from PPCSubtarget into sys::getHostCPUName().
Both the new Linux functionality and the old Darwin functions have been moved.
This change also allows this information to be queried directly by clang and
other frontends (clang, for example, will now have real -mcpu=native support).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158349
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 12 Jun 2012 01:06:16 +0000 (01:06 +0000)]
Satisfy C++ aliasing rules, per suggestion by Chandler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158346
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Tue, 12 Jun 2012 00:58:40 +0000 (00:58 +0000)]
Revert r158337 "Move PPC host-CPU detection logic from PPCSubtarget into sys::getHostCPUName()."
This commit broke most of the PowerPC unit tests when running on
Intel/Apple.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158345
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Tue, 12 Jun 2012 00:48:47 +0000 (00:48 +0000)]
FileCheck docs: remove leftover HTML markup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158344
91177308-0d34-0410-b5e6-
96231b3b80d8
Argyrios Kyrtzidis [Tue, 12 Jun 2012 00:21:31 +0000 (00:21 +0000)]
For llvm::sys::ThreadLocalImpl instead of malloc'ing the platform-specific
thread local data, embed them in the class using a uint64_t and make sure
we get compiler errors if there's a platform where this is not big enough.
This makes ThreadLocal more safe for using it in conjunction with CrashRecoveryContext.
Related to crash in rdar://
11434201.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158342
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Mon, 11 Jun 2012 23:42:23 +0000 (23:42 +0000)]
misched: When querying RegisterPressureTracker, always save current and max pressure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158340
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Mon, 11 Jun 2012 23:42:20 +0000 (23:42 +0000)]
misched: regpressure getMaxPressureDelta, revert accidental checkin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158339
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 23:14:31 +0000 (23:14 +0000)]
Move PPC host-CPU detection logic from PPCSubtarget into sys::getHostCPUName().
Both the new Linux functionality and the old Darwin functions have been moved.
This change also allows this information to be queried directly by clang and
other frontends (clang, for example, will now have real -mcpu=native support).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158337
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 11 Jun 2012 21:14:28 +0000 (21:14 +0000)]
Fix test that depends on register allocation.
The test is really checking the prolog/epilog load/store multiple
formation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158328
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 19:57:04 +0000 (19:57 +0000)]
Enable MFOCRF generation on the PPC A2 core.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158324
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 19:57:01 +0000 (19:57 +0000)]
Rename the PPC target feature gpul to mfocrf.
The PPC target feature gpul (IsGigaProcessor) was only used for one thing:
To enable the generation of the MFOCRF instruction. Furthermore, this
instruction is available on other PPC cores outside of the G5 line. This
feature now corresponds to the HasMFOCRF flag.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158323
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 19:56:57 +0000 (19:56 +0000)]
Add A2 to the list of PPC CPUs recognized by Linux host CPU-type detection.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158322
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 11 Jun 2012 16:01:14 +0000 (16:01 +0000)]
Fix test case to work on ARM.
Patch by James Benton!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158316
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 15:43:15 +0000 (15:43 +0000)]
Emit the two-operand form of the PPC mfcr instruction as mfocrf.
This is necessary on Linux and supported on Darwin, see PR2604.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158315
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 15:43:13 +0000 (15:43 +0000)]
Add local CPU detection for Linux PPC.
This functionality mirrors that available on PPC/Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158314
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Mon, 11 Jun 2012 15:43:08 +0000 (15:43 +0000)]
Add POWER6 and POWER7 CPU types to the PPC backend.
No functional change; these will be used by upcoming scheduler enhancements.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158313
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 11 Jun 2012 15:37:55 +0000 (15:37 +0000)]
Write llvm-tblgen backends as functions instead of sub-classes.
The TableGenBackend base class doesn't do much, and will be removed
completely soon.
Patch by Sean Silva!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158311
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Mon, 11 Jun 2012 15:11:12 +0000 (15:11 +0000)]
Fix a problem with the reverse bundle iterators.
This showed up the first time rend() was called on a bundled instruction
in the Mips backend.
Also avoid dereferencing end() in bundle_iterator::operator++().
We still don't have a place to put unit tests for this stuff.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158310
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Mon, 11 Jun 2012 09:40:10 +0000 (09:40 +0000)]
Object file output from llc isn't experimental anymore.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158305
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Mon, 11 Jun 2012 08:07:26 +0000 (08:07 +0000)]
Re-enable the CMN instruction.
We turned off the CMN instruction because it had semantics which we weren't
getting correct. If we are comparing with an immediate, then it's okay to use
the CMN instruction.
<rdar://problem/
7569620>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158302
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Mon, 11 Jun 2012 08:01:25 +0000 (08:01 +0000)]
InstCombine: factor code better.
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158301
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 10 Jun 2012 20:35:00 +0000 (20:35 +0000)]
InstCombine: Turn (zext A) == (B & (1<<X)-1) into A == (trunc B), narrowing the compare.
This saves a cast, and zext is more expensive on platforms with subreg support
than trunc is. This occurs in the BSD implementation of memchr(3), see PR12750.
On the synthetic benchmark from that bug stupid_memchr and bsd_memchr have the
same performance now when not inlining either function.
stupid_memchr: 323.0us
bsd_memchr: 321.0us
memchr: 479.0us
where memchr is the llvm-gcc compiled bsd_memchr from osx lion's libc. When
inlining is enabled bsd_memchr still regresses down to llvm-gcc memchr time,
I haven't fully understood the issue yet, something is grossly mangling the
loop after inlining.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158297
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sun, 10 Jun 2012 19:32:29 +0000 (19:32 +0000)]
Enable ILP scheduling for all nodes by default on PPC.
Over the entire test-suite, this has an insignificantly negative average
performance impact, but reduces some of the worst slowdowns from the
anti-dep. change (r158294).
Largest speedups:
SingleSource/Benchmarks/Stanford/Quicksort - 28%
SingleSource/Benchmarks/Stanford/Towers - 24%
SingleSource/Benchmarks/Shootout-C++/matrix - 23%
MultiSource/Benchmarks/SciMark2-C/scimark2 - 19%
MultiSource/Benchmarks/MiBench/automotive-bitcount/automotive-bitcount - 15%
(matrix and automotive-bitcount were both in the top-5 slowdown list from the
anti-dep. change)
Largest slowdowns:
MultiSource/Benchmarks/McCat/03-testtrie/testtrie - 28%
MultiSource/Benchmarks/mediabench/gsm/toast/toast - 26%
MultiSource/Benchmarks/MiBench/automotive-susan/automotive-susan - 21%
SingleSource/Benchmarks/CoyoteBench/lpbench - 20%
MultiSource/Applications/d/make_dparser - 16%
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158296
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 10 Jun 2012 18:42:51 +0000 (18:42 +0000)]
Add AutoUpgrade support for the SSE4 ptest intrinsics.
Patch by Michael Kuperstein.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158295
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sun, 10 Jun 2012 11:15:36 +0000 (11:15 +0000)]
Use critical anti-dep. breaking on all PPC targets, but also add other register classes.
Using 'all' instead of 'critical' would be better because it would make it easier to
satisfy the bundling constraints, but, as noted in the FIXME, that is currently not
possible with the crs.
This yields an average 1% speedup over the entire test suite (on Power 7). Largest speedups:
SingleSource/Benchmarks/Shootout-C++/moments - 40%
MultiSource/Benchmarks/McCat/03-testtrie/testtrie - 28%
SingleSource/Benchmarks/BenchmarkGame/nsieve-bits - 26%
SingleSource/Benchmarks/McGill/misr - 23%
MultiSource/Applications/JM/ldecod/ldecod - 22%
Largest slowdowns:
SingleSource/Benchmarks/Shootout-C++/matrix - -29%
SingleSource/Benchmarks/Shootout-C++/ary3 - -22%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - -18%
SingleSource/Benchmarks/Shootout-C++/ary - -17%
MultiSource/Benchmarks/MiBench/automotive-bitcount/automotive-bitcount - -15%
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158294
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sun, 10 Jun 2012 07:31:56 +0000 (07:31 +0000)]
Add intrinsics for immediate form of XOP vprot instructions. Use i128mem instead of f128mem for integer XOP instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158291
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 9 Jun 2012 22:10:19 +0000 (22:10 +0000)]
Improve ext/trunc patterns on PPC64.
The PPC64 backend had patterns for i32 <-> i64 extensions and truncations that
would leave self-moves in the final assembly. Replacing those patterns with ones
based on the SUBREG builtins yields better-looking code.
Thanks to Jakob and Owen for their suggestions in this matter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158283
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 9 Jun 2012 17:02:24 +0000 (17:02 +0000)]
Use XOP vpcom intrinsics in patterns instead of a target specific SDNode type. Remove the custom lowering code that selected the SDNode type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158279
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 9 Jun 2012 16:46:13 +0000 (16:46 +0000)]
Replace XOP vpcom intrinsics with fewer intrinsics that take the immediate as an argument.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158278
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 9 Jun 2012 15:33:28 +0000 (15:33 +0000)]
Hashing: Remove outdated comment. Support for reserved hash values was removed in r151865.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158276
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Sat, 9 Jun 2012 13:59:29 +0000 (13:59 +0000)]
Disabling a spurious deprecation warning about using PathV1 from within the PathV1 implementation file.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158274
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Sat, 9 Jun 2012 13:46:36 +0000 (13:46 +0000)]
Fixing a typo in the comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158273
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 9 Jun 2012 10:34:15 +0000 (10:34 +0000)]
Allocate the contents of DwarfDebug's StringMaps in a single big BumpPtrAllocator.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158265
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Sat, 9 Jun 2012 10:04:03 +0000 (10:04 +0000)]
Silence a gcc-4.6 warning: GCC fails to understand that secondReg and cmpOp2 are
correlated, and thinks that cmpOp2 may be used uninitialized.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158263
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Sat, 9 Jun 2012 03:14:50 +0000 (03:14 +0000)]
Enable tail merging on PPC.
Tail merging had been disabled on PPC because it would disturb bundling decisions
made during pre-RA scheduling on the 970 cores. Now, however, all bundling decisions
are made during post-RA scheduling, and tail merging is generally beneficial (the
average test-suite speedup is insignificantly positive).
Largest test-suite speedups:
MultiSource/Benchmarks/mediabench/gsm/toast/toast - 30%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - 23%
SingleSource/Benchmarks/Shootout-C++/ary - 21%
SingleSource/Benchmarks/Stanford/Queens - 17%
Largest slowdowns:
MultiSource/Benchmarks/MiBench/security-sha/security-sha - 24%
MultiSource/Benchmarks/McCat/03-testtrie/testtrie - 22%
MultiSource/Applications/JM/ldecod/ldecod - 14%
MultiSource/Benchmarks/mediabench/g721/g721encode/encode - 9%
This is improved by using full (instead of just critical) anti-dependency breaking,
but doing so still causes miscompiles and so cannot yet be enabled by default.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158259
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Sat, 9 Jun 2012 02:16:58 +0000 (02:16 +0000)]
Register pressure: added getPressureAfterInstr.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158256
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 9 Jun 2012 02:13:10 +0000 (02:13 +0000)]
Sketch a LiveRegMatrix analysis pass.
The LiveRegMatrix represents the live range of assigned virtual
registers in a Live interval union per register unit. This is not
fundamentally different from the interference tracking in RegAllocBase
that both RABasic and RAGreedy use.
The important differences are:
- LiveRegMatrix tracks interference per register unit instead of per
physical register. This makes interference checks cheaper and
assignments slightly more expensive. For example, the ARM D7 reigster
has 24 aliases, so we would check 24 physregs before assigning to one.
With unit-based interference, we check 2 units before assigning to 2
units.
- LiveRegMatrix caches regmask interference checks. That is currently
duplicated functionality in RABasic and RAGreedy.
- LiveRegMatrix is a pass which makes it possible to insert
target-dependent passes between register allocation and rewriting.
Such passes could tweak the register assignments with interference
checking support from LiveRegMatrix.
Eventually, RABasic and RAGreedy will be switched to LiveRegMatrix.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158255
91177308-0d34-0410-b5e6-
96231b3b80d8
Jack Carter [Sat, 9 Jun 2012 00:27:55 +0000 (00:27 +0000)]
Test commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158250
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 9 Jun 2012 00:14:47 +0000 (00:14 +0000)]
Also compute MBB live-in lists in the new rewriter pass.
This deduplicates some code from the optimizing register allocators, and
it means that it is now possible to change the register allocators'
solutions simply by editing the VirtRegMap between the register
allocator pass and the rewriter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158249
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Sat, 9 Jun 2012 00:01:45 +0000 (00:01 +0000)]
Convert comments to proper Doxygen comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158248
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 8 Jun 2012 23:56:26 +0000 (23:56 +0000)]
Removing strange "using" declarations form TargetInstrInfo.
I can't imagine why these were added. Trial and error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158247
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 8 Jun 2012 23:44:45 +0000 (23:44 +0000)]
Reintroduce VirtRegRewriter.
OK, not really. We don't want to reintroduce the old rewriter hacks.
This patch extracts virtual register rewriting as a separate pass that
runs after the register allocator. This is possible now that
CodeGen/Passes.cpp can configure the full optimizing register allocator
pipeline.
The rewriter pass uses register assignments in VirtRegMap to rewrite
virtual registers to physical registers, and it inserts kill flags based
on live intervals.
These finalization steps are the same for the optimizing register
allocators: RABasic, RAGreedy, and PBQP.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158244
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 8 Jun 2012 23:15:12 +0000 (23:15 +0000)]
Don't run RAFast in the optimizing regalloc pipeline.
The fast register allocator is not supposed to work in the optimizing
pipeline. It doesn't make sense to compute live intervals, run full copy
coalescing, and then run RAFast.
Fast register allocation in the optimizing pipeline is better done by
RABasic.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158242
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Fri, 8 Jun 2012 22:30:05 +0000 (22:30 +0000)]
canonicalize:
-%a + 42
into
42 - %a
previously we were emitting:
-(%a + 42)
This fixes the infinite loop in PR12338. The generated code is still not perfect, though.
Will work on that next
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158237
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 8 Jun 2012 21:53:50 +0000 (21:53 +0000)]
Start implementing pre-ra if-converter: using speculation and selects to eliminate branches.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158234
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 8 Jun 2012 21:52:38 +0000 (21:52 +0000)]
TargetInstrInfo hooks implemented in codegen should be declared pure virtual.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158233
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 8 Jun 2012 20:15:33 +0000 (20:15 +0000)]
Reapply commit 158073 with a fix (the testcase was already committed). The
problem was that by moving instructions around inside the function, the pass
could accidentally move the iterator being used to advance over the function
too. Fix this by only processing the instruction equal to the iterator, and
leaving processing of instructions that might not be equal to the iterator
to later (later = after traversing the basic block; it could also wait until
after traversing the entire function, but this might make the sets quite big).
Original commit message:
Grab-bag of reassociate tweaks. Unify handling of dead instructions and
instructions to reoptimize. Exploit this to more systematically eliminate
dead instructions (this isn't very useful in practice but is convenient for
analysing some testcase I am working on). No need for WeakVH any more: use
an AssertingVH instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158226
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 20:02:09 +0000 (20:02 +0000)]
Remove the TODO statement in the PPC README re: CTR loops
As Chris points out, this can now be removed!
TODO: check if the associated section on viterbi's inner loop can also be removed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158224
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 19:19:53 +0000 (19:19 +0000)]
Enable PPC CTR loop formation by default.
Thanks to Jakob's help, this now causes no new test suite failures!
Over the entire test suite, this gives an average 1% speedup. The largest speedups are:
SingleSource/Benchmarks/Misc/pi - 108%
SingleSource/Benchmarks/CoyoteBench/lpbench - 54%
MultiSource/Benchmarks/Prolangs-C/unix-smail/unix-smail - 50%
SingleSource/Benchmarks/Shootout/ary3 - 32%
SingleSource/Benchmarks/Shootout-C++/matrix - 30%
The largest slowdowns are:
MultiSource/Benchmarks/mediabench/gsm/toast/toast - -30%
MultiSource/Benchmarks/Prolangs-C/bison/mybison - -25%
MultiSource/Benchmarks/BitBench/uuencode/uuencode - -22%
MultiSource/Applications/d/make_dparser - -14%
SingleSource/Benchmarks/Shootout-C++/ary - -13%
In light of these slowdowns, additional profiling work is obviously needed!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158223
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 19:02:08 +0000 (19:02 +0000)]
Mark the PPC CTRRC and CTRRC8 register classes as non-allocatable.
Marking these classes as non-alocatable allows CTR loop generation to
work correctly with the block placement passes, etc. These register
classes are currently used only by some unused TCRETURN patterns.
In future cleanup, these will be removed.
Thanks again to Jakob for suggesting this fix to the CTR loop problem!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158221
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 8 Jun 2012 18:58:26 +0000 (18:58 +0000)]
Enable optimization for integer ABS on X86 if Subtarget has CMOV.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158220
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Fri, 8 Jun 2012 18:42:37 +0000 (18:42 +0000)]
Test case for r158160
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158218
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 8 Jun 2012 18:25:47 +0000 (18:25 +0000)]
Sched itinerary fix: Avoid static initializers.
This fixes an accidental dependence on static initialization order that I introduced yesterday.
Thank you Lang!!!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158215
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Fri, 8 Jun 2012 18:04:52 +0000 (18:04 +0000)]
Fix a crash in APInt::lshr when shiftAmt > BitWidth.
Patch by James Benton <jbenton@vmware.com>.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158213
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Fri, 8 Jun 2012 17:23:27 +0000 (17:23 +0000)]
Fix Target->Codegen dependence.
Bulk move of TargetInstrInfo implementation into
TargetInstrInfoImpl. This is dirty because the code isn't part of
TargetInstrInfoImpl class, nor should it be, because the methods are
not target hooks. However, it's the current mechanism for keeping
libTarget useful outside the backend. You'll get a not-so-nice link
error if you invoke a TargetInstrInfo method that depends on CodeGen.
The TargetInstrInfoImpl class should probably be removed since it
doesn't really solve this problem.
To really fix this, we probably need separate interfaces for the
CodeGen/nonCodeGen sides of TargetInstrInfo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158212
91177308-0d34-0410-b5e6-
96231b3b80d8
Nuno Lopes [Fri, 8 Jun 2012 16:31:42 +0000 (16:31 +0000)]
BoundsChecking: add support for ConstantPointerNull. fixes a bunch of instrumentation failures in loops with reallocs
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158210
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Fri, 8 Jun 2012 16:28:06 +0000 (16:28 +0000)]
test/CodeGen/Generic/APIntLoadStore.ll: Mark as XFAIL:ppc since r157911.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158209
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 15:38:25 +0000 (15:38 +0000)]
Disable the PPC CTR-Loops pass by default.
The pass itself works well, but the something in the Machine* infrastructure
does not understand terminators which define registers. Without the ability
to use the block-placement pass, etc. this causes performance regressions (and
so is turned off by default). Turning off the analysis turns off the problems
with the Machine* infrastructure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158206
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 15:38:23 +0000 (15:38 +0000)]
Fix a bug in the new PPC CTR-Loops pass.
The code which tests for an induction operation cannot assume that any
ADDI instruction will have a register operand because the operand could
also be a frame index; for example:
%vreg16<def> = ADDI8 <fi#0>, 0; G8RC:%vreg16
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158205
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Fri, 8 Jun 2012 15:38:21 +0000 (15:38 +0000)]
Add the PPCCTRLoops pass: a PPC machine-code-level optimization pass to form CTR-based loop branching code.
This pass is derived from the Hexagon HardwareLoops pass. The only significant enhancement over the Hexagon
pass is that PPCCTRLoops will also attempt to delete the replaced add and compare operations if they are
no longer otherwise used. Also, invalid preheader DebugLoc is not used.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158204
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 8 Jun 2012 13:37:30 +0000 (13:37 +0000)]
Revert commit 158073 while waiting for a fix. The issue is that reassociate
can move instructions within the instruction list. If the instruction just
happens to be the one the basic block iterator is pointing to, and it is
moved to a different basic block, then we get into an infinite loop due to
the iterator running off the end of the basic block (for some reason this
doesn't fire any assertions). Original commit message:
Grab-bag of reassociate tweaks. Unify handling of dead instructions and
instructions to reoptimize. Exploit this to more systematically eliminate
dead instructions (this isn't very useful in practice but is convenient for
analysing some testcase I am working on). No need for WeakVH any more: use
an AssertingVH instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158199
91177308-0d34-0410-b5e6-
96231b3b80d8
Tobias Grosser [Fri, 8 Jun 2012 09:41:23 +0000 (09:41 +0000)]
cmake: Pass the -m32 flag to modules if LLVM_BUILD_32_BITS is enabled
This was previously only done for executables and shared libraries, but not
for modules. As modules are essentially shared libraries (that need to be
dlopened explicitly), threating them the same as shared libraries seems
reasonable. This fixes the LLVM_BUILD_32_BITS build of Polly.
Contributed by: Ondra Hosek <ondra.hosek@gmail.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158195
91177308-0d34-0410-b5e6-
96231b3b80d8
Owen Anderson [Fri, 8 Jun 2012 00:25:03 +0000 (00:25 +0000)]
Teach the AsmMatcherEmitter to allow InstAlias' where the suboperands of a complex operand are called out explicitly in the asm string.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158183
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 7 Jun 2012 23:33:56 +0000 (23:33 +0000)]
[CMake] Promote extension warnings to errors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158176
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Thu, 7 Jun 2012 22:39:10 +0000 (22:39 +0000)]
X86: optimize generated code for integer ABS
This patch will generate the following for integer ABS:
movl %edi, %eax
negl %eax
cmovll %edi, %eax
INSTEAD OF
movl %edi, %ecx
sarl $31, %ecx
leal (%rdi,%rcx), %eax
xorl %ecx, %eax
There exists a target-independent DAG combine for integer ABS, which converts
integer ABS to sar+add+xor. For X86, we match this pattern back to neg+cmov.
This is implemented in PerformXorCombine.
rdar://
10695237
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158175
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 7 Jun 2012 21:34:31 +0000 (21:34 +0000)]
[CMake] Order MSVC warnings numerically.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158171
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael J. Spencer [Thu, 7 Jun 2012 21:34:15 +0000 (21:34 +0000)]
[CMake] Adjust MSVC warnings.
Remove /Wall from LLVM_ENABLE_WARNINGS (it's useless) and promote 4239
to a level 1 warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158170
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 7 Jun 2012 20:53:48 +0000 (20:53 +0000)]
Do not optimize the used bits of the x86 vselect condition operand, when the condition operand is a vector of 1-bit predicates.
This may happen on MIC devices.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158168
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 7 Jun 2012 20:28:57 +0000 (20:28 +0000)]
Fix a bug in FoldSelectOpOp. Bitcast ops may change the number of vector elements, which may disagree with the select condition type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158166
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 7 Jun 2012 19:42:04 +0000 (19:42 +0000)]
Continue factoring computeOperandLatency. Use it for ARM hasHighOperandLatency.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158164
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 7 Jun 2012 19:42:00 +0000 (19:42 +0000)]
ARM getOperandLatency rewrite.
Match expectations of the new latency API. Cleanup and make the logic consistent.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158163
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 7 Jun 2012 19:41:58 +0000 (19:41 +0000)]
ARM getOperandLatency should return -1 for unknown, consistent with API
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158162
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Trick [Thu, 7 Jun 2012 19:41:55 +0000 (19:41 +0000)]
Fix ARM getInstrLatency logic to work with the current API.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158161
91177308-0d34-0410-b5e6-
96231b3b80d8
Manman Ren [Thu, 7 Jun 2012 19:27:33 +0000 (19:27 +0000)]
PR13046: we can't replace usage of SUB with CMP in the lowering phase.
It will cause assertion failure later on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158160
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 7 Jun 2012 18:39:19 +0000 (18:39 +0000)]
Use a base register instead of an index register with the local dynamic model.
Fixes pr13048.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158158
91177308-0d34-0410-b5e6-
96231b3b80d8
Pete Cooper [Thu, 7 Jun 2012 17:41:39 +0000 (17:41 +0000)]
Move terminator machine verification to check MachineBasicBlock::instr_iterator instead of MBB::iterator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158154
91177308-0d34-0410-b5e6-
96231b3b80d8
Pete Cooper [Thu, 7 Jun 2012 04:43:52 +0000 (04:43 +0000)]
Add internal read flags to MachineInstrBuilder and hook them into the MachineOperand flag of the same name
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@158137
91177308-0d34-0410-b5e6-
96231b3b80d8