6502/llvm-6502
1
0 Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2026-04-27 04:18:42 +00:00
Code Issues Projects Releases Wiki Activity
66,260 Commits 39 Branches 0 Tags
15b337c28da771b8426fafad58f2e2f81ca4c64e
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Owen Anderson 93ef3fd9c0 Tests for NEON encoding of vmax. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117327 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 23:45:34 +00:00
Owen Anderson a13067e366 Tests for NEON encoding of vmin. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117326 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 23:35:36 +00:00
Owen Anderson 5258b61966 Add correct encodings for NEON vabal. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117315 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 21:29:04 +00:00
Owen Anderson 410aebc670 Add correct NEON encodings for vaba. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117309 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 20:52:57 +00:00
Owen Anderson 28bae6106f Tests for NEON encoding of vabdl. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117303 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 20:36:28 +00:00
Owen Anderson b7e1d77ff5 Add tests for NEON encoding of vabd. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117302 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 20:29:27 +00:00
Owen Anderson 31e6ed890a Attempt to provide correct encodings for NEON vbit and vbif, even though we can't test them at the moment. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117294 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 20:17:22 +00:00
Owen Anderson 4110b4325d Provide correct NEON encodings for vbsl. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117293 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 20:13:13 +00:00
Owen Anderson 162875a9f3 Add correct instruction encodings for vbic, vorn, and vmvn. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117282 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 18:43:52 +00:00
Owen Anderson 8c71eff594 Provide correct NEON encodings for vand, veor, and vorr. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117279 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 18:28:30 +00:00
Owen Anderson c61ec2a2b0 Add tests for NEON encoding of vtst. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117277 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 18:10:34 +00:00
Owen Anderson d0c5b6170f Add NEON encoding tests for vcgt and vacgt. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117276 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 18:03:59 +00:00
Owen Anderson 10c15e5d58 Add tests for NEON encodings of vcge and vacge. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117274 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 17:49:32 +00:00
Owen Anderson 4fe20bbd66 Add a warning about our inability to test the encoding of vceq with immediate zero. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117273 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 17:33:02 +00:00
Owen Anderson a2041f1807 Add tests for NEON encoding of vceq. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117270 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 17:20:26 +00:00
Owen Anderson 8b7ce020c3 Add tests for NEON encoding of vsubhn and vrsubhn. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117269 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-25 17:12:46 +00:00
Owen Anderson c052a8c772 Add tests for NEON encoding of vqsub. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117214 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-23 18:02:16 +00:00
Owen Anderson 61f34bc4bc Add tests for NEON encoding of vhsub. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117189 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:58:22 +00:00
Owen Anderson 884f228692 Add tests for NEON encoding of vsubw. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117186 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:46:07 +00:00
Owen Anderson 2b6b97c815 Add tests for NEON encoding of vsubl. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117183 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:36:36 +00:00
Owen Anderson 313252022d Add tests for NEON encoding of vsub. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117177 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:21:04 +00:00
Owen Anderson c9db331433 Add tests for NEON encoding of vqdmlsl. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117173 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:08:47 +00:00
Owen Anderson 353f8668b8 Add tests for NEON encoding of vmlsl. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117171 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 23:02:27 +00:00
Owen Anderson 458509476b Add tests for the correct encoding of NEON vmls. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117145 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 20:31:24 +00:00
Owen Anderson 9b26497273 Add correct NEON encodings for vqdmlal. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117134 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 19:35:48 +00:00
Owen Anderson 92205842ca Provide correct encodings for NEON vmlal. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117131 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 19:05:25 +00:00
Owen Anderson 18341e9e31 Provide correct NEON encodings for vmla. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117126 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 18:54:37 +00:00
Owen Anderson 81faa805ce Add testscases for encoding of NEON vdqmull. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117115 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-22 17:57:37 +00:00
Owen Anderson 4ceccc4e57 Add tests for NEON encoding of vmull. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117077 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 22:19:53 +00:00
Owen Anderson de5370fcbb Add tests for NEON encoding of vqdmulh and vqrdmulh. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117074 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 22:05:33 +00:00
Owen Anderson 636ad14c8a Add tests for NEON vmul encoding. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117069 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 21:51:58 +00:00
Owen Anderson 7eca0e17ba Rename this test to better reflect its contents. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117067 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 21:40:15 +00:00
Owen Anderson 35ea7a4022 Add tests for NEON encodings of vaddhn and vraddhn. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117064 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 20:56:57 +00:00
Owen Anderson 90d4cf9314 Move the encoding logic for Q registers into getMachineOpValue(). 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117060 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 20:49:13 +00:00
Owen Anderson 4bcb949e18 Add tests for NEON encodings of vqadd, which was already correctly encoded. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117059 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 20:42:04 +00:00
Owen Anderson d451f888b8 ARM encodes Q registers as 2xregno (i.e. the number of the D register that corresponds to the lower 
half of the Q register), rather than with just regno.  This allows us to unify the encodings for
a lot of different NEON instrucitons that differ only in whether they have Q or D register operands.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117056 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 20:21:49 +00:00
Owen Anderson 1e93466c3a Add correct NEON encodings for vhadd and vrhadd. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117047 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 18:55:04 +00:00
Owen Anderson 9d50559bae Add correct encodings for NEON vaddw.s* and vaddw.u*. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117040 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 18:20:25 +00:00
Owen Anderson e0e6dc3f4e Provide correct NEON encodings for vaddl.u* and vaddl.s*. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@117039 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 18:09:17 +00:00
Owen Anderson d9707e3d85 Revert r116983, which is breaking all the buildbots. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116987 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 03:11:16 +00:00
Owen Anderson d7795540d0 Implement correct encodings for NEON vadd, both integer and floating point. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116981 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-21 00:48:00 +00:00
Owen Anderson 75f6df283f Attempt to fix valgrind complaining about (benign) leaks in pass registration by having PassRegistry 
delete PassInfo objects that were created using new-style initialization.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116959 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-20 22:22:30 +00:00
Owen Anderson 205942a4a5 Passes do not need to recursively initialize passes that they preserve, if 
they do not also require them.  This allows us to reduce inter-pass linkage
dependencies.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116854 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-19 20:08:44 +00:00
Owen Anderson 879b65b396 Remove extraneous slash. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116851 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-19 19:54:48 +00:00
Owen Anderson 62d4ced64c Factor out the call-once implementation into its own macro. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116832 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-19 18:02:06 +00:00
Owen Anderson 081c34b725 Get rid of static constructors for pass registration. Instead, every pass exposes an initializeMyPassFunction(), which 
must be called in the pass's constructor.  This function uses static dependency declarations to recursively initialize
the pass's dependencies.

Clients that only create passes through the createFooPass() APIs will require no changes.  Clients that want to use the
CommandLine options for passes will need to manually call the appropriate initialization functions in PassInitialization.h
before parsing commandline arguments.

I have tested this with all standard configurations of clang and llvm-gcc on Darwin.  It is possible that there are problems
with the static dependencies that will only be visible with non-standard options.  If you encounter any crash in pass
registration/creation, please send the testcase to me directly.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116820 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-19 17:21:58 +00:00
Owen Anderson 6549121c66 Generalize MemCpyOpt's handling of call slot forwarding to function properly when the call slot 
forwarding is implemented with a load/store pair rather than a memcpy.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116637 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-15 22:52:12 +00:00
Owen Anderson c934a67ef6 Try again at implementing thread-safe lazy pass initialization, without depending on static local initialization 
being threadsafe AND ensuring that initialization is complete by the time the initializeFooPass method returns.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116492 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-14 17:59:03 +00:00
Owen Anderson 20a70c8472 Revert r116489. It included some changes I didn't intend to commit, and broke the buildbots. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116491 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-14 17:36:50 +00:00
Owen Anderson 05b4af7078 Apparently MSVC doesn't support thread-safe static local initialization. Roll our own solution instead. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116489 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-14 17:26:06 +00:00
Owen Anderson ae0a7bc683 CallGraphSCC passes implicity require CallGraph analysis. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116443 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-13 22:00:45 +00:00
Owen Anderson 139788fe6f Conversely, Analysis-implementations do NOT need to initialize the AnalysisGroup. It will only matter when 
someone tries to require that AG, in which case it is the requester's responsibility to initialize it.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116442 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-13 21:55:07 +00:00
Owen Anderson 325e264355 Analysis groups need to initialize their default implementations. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116441 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-13 21:49:58 +00:00
Owen Anderson 972beb5431 Take advantage of C++'s thread-safe static local initialization to simplify thread-safe pass initialization. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116426 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-13 20:24:34 +00:00
Owen Anderson 2ab36d3502 Begin adding static dependence information to passes, which will allow us to 
perform initialization without static constructors AND without explicit initialization
by the client.  For the moment, passes are required to initialize both their
(potential) dependencies and any passes they preserve.  I hope to be able to relax
the latter requirement in the future.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@116334 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-12 19:48:12 +00:00
Owen Anderson ce665bd2e2 Now with fewer extraneous semicolons! 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115996 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 22:25:06 +00:00
Owen Anderson aa4897fa13 Add initialization routines for Instrumentation. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115971 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 20:17:24 +00:00
Owen Anderson 154c5c46fc Improve comment. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115966 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 20:05:18 +00:00
Owen Anderson 74cfb0ce1c Add initialization routines to InstCombine. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115965 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 20:04:55 +00:00
Owen Anderson b8a1ccfc4b Add initialization routines for VMCore. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115963 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 19:51:21 +00:00
Owen Anderson 9966306aa7 Add initialization routines for Target. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115957 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:50:11 +00:00
Owen Anderson 78b152470d Add initialization routines for CodeGen. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115949 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:41:20 +00:00
Owen Anderson 1c8820d8d8 Add an implementation of the initialization routine for IPA. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115947 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:31:27 +00:00
Owen Anderson 861f4c38de Add initialization routines for Analysis and IPA. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115946 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:31:00 +00:00
Owen Anderson 2c81296bc2 Add an initialization routine for libLLVMipo.a 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115933 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:09:59 +00:00
Owen Anderson fbd121e198 Add a header that I forgot to commit. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115932 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 18:05:11 +00:00
Owen Anderson 1a3d233621 Next step on the getting-rid-of-static-ctors train: begin adding per-library 
initialization functions that initialize the set of passes implemented in
that library.  Add C bindings for these functions as well.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115927 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 17:55:47 +00:00
Owen Anderson 745c872bb5 Fix a warning when building with clang++. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115924 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 17:04:18 +00:00
Owen Anderson 7bf8a4b21f Add the header that I accidentally forgot from r115900. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115901 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 04:17:38 +00:00
Owen Anderson 71802344fc Move the pass initialization helper functions into the llvm namespace, and add 
a header declaring them all.  This is also where we will declare per-library pass-set
initializer functions down the road.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115900 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 04:13:08 +00:00
Owen Anderson cce7f7cd03 Since the Hello pass is built as a loadable dynamic library, don't try to convert it to new-style registration yet. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115881 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-07 00:31:16 +00:00
Owen Anderson 9875903799 Appease the clang self-host buildbot by providing a correct instantiation. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115857 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-06 22:23:20 +00:00
Owen Anderson e9ef41a47d Hide analysis group registration behind a macro, just like pass registration. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115835 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-06 21:02:27 +00:00
Owen Anderson 5f0d700278 Pass initialization functions should take a PassRegistry as a parameter 
rather than being fixed to the global registry.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115824 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-06 20:07:03 +00:00
Owen Anderson e2a268fd03 Another step towards getting rid of static ctors for pass registration: have INITIALIZE_PASS AND INITIALIZE_AG_PASS 
expand to an initializeMyPass() function (in additional to the extant static ctors).  Eventually, these will be called
from a big InitializeAllPasses() function, and the PassInfo's they create (which would be leaked if this code were used
at the moment) will be handed off to a PassRegistry for ownership.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115703 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-05 22:58:16 +00:00
Owen Anderson 3a9e7690ba Use a more efficient lowering of uint64_t --> float that can take advantage of hardware signed integer conversion without 
having to do a double cast (uint64_t --> double --> float).  This is based on the algorithm from compiler_rt's __floatundisf
for X86-64.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115634 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-05 17:24:05 +00:00
Owen Anderson e3cc84a43d Thread the determination of branch prediction hit rates back through the if-conversion heuristic APIs. For now, 
stick with a constant estimate of 90% (branch predictors are good!), but we might find that we want to provide
more nuanced estimates in the future.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115364 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-01 22:45:50 +00:00
Owen Anderson 00d4f48168 Make the spelling of the flags for old-style if-conversion heuristics consistent between ARM and Thumb2. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115341 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-01 20:33:47 +00:00
Owen Anderson aa9f0a57d0 Provide an option to restore old-style if-conversion heuristics for Thumb2. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115339 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-01 20:28:06 +00:00
Owen Anderson 7267e14327 Now that the profitable bits of EnableFullLoadPRE have been enabled by default, rip out the remainder. 
Anyone interested in more general PRE would be better served by implementing it separately, to get real
anticipation calculation, etc.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115337 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-10-01 20:02:55 +00:00
Owen Anderson b3c04ec956 Temporarily add a flag to make it easier to compare the new-style ARM if 
conversion heuristics to the old-style ones.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115239 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-30 23:48:38 +00:00
Owen Anderson 722cc1f414 We do want to allow LoadPRE to perform LICM-like transformations: we already consider PHI nodes to be negligible for 
code size (making this transform code size neutral), and it allows us to hoist values out of loops, which is always
a good thing.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115205 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-30 20:53:04 +00:00
Owen Anderson fe45b104b3 Revert r115099 (adding early jump threading). It's not clear if the benefits are worth the compile time cost. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115106 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-29 23:31:09 +00:00
Owen Anderson 235859d66c Early CFG simplification can fold conditionals down to selects, which is often a good thing, but it can also 
hide jump threading opportunities by turning control flow into data flow.  Run an early JumpThreading pass
(adds approximately an additional 1% to optimization time on SPEC), allowing it to get a shot at these cases
first.  Fixes <rdar://problem/8447345>.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115099 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-29 22:57:02 +00:00
Owen Anderson b0725316cd UnreachableBlockElim could incorrectly return false when it had not modified the CFG, but HAD modified some PHI nodes. Fixes PR8174. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115083 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-29 20:57:19 +00:00
Owen Anderson 36c4debc94 Fix PR8247: JumpThreading can cause a block to become unreachable while still having predecessor, if it is part of a self-loop. 
Because of this, we cannot use the Simplify* APIs, as they can assert-fail on unreachable code.  Since it's not easy to determine
if a given threading will cause a block to become unreachable, simply defer simplifying simplification to later InstCombine and/or
DCE passes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115082 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-29 20:34:41 +00:00
Owen Anderson 69daca59ff Revert r114919, which caused some serious regressions on ARM. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@115053 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-29 18:05:19 +00:00
Owen Anderson 654d5440a4 Add a subtarget hook for reporting the misprediction penalty. Use this to provide more precise 
cost modeling for if-conversion.  Now if only we had a way to estimate the misprediction probability.

Adjsut CodeGen/ARM/ifcvt10.ll.  The pipeline on Cortex-A8 is long enough that it is still profitable
to predicate an ldm, but the shorter pipeline on Cortex-A9 makes it unprofitable.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114995 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-28 21:57:50 +00:00
Owen Anderson 7571ee7885 Give the if-converter access to MachineLoopInfo, and use it to generate plausible branch prediction 
estimates.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114981 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-28 20:42:15 +00:00
Owen Anderson b20b85168c Part one of switching to using a more sane heuristic for determining if-conversion profitability. 
Rather than having arbitrary cutoffs, actually try to cost model the conversion.

For now, the constants are tuned to more or less match our existing behavior, but these will be
changed to reflect realistic values as this work proceeds.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114973 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-28 18:32:13 +00:00
Owen Anderson 8d9e92ce2d Weight loop unrolling counts by nesting depth. Unrolling deeply nested loops tends to cause high 
register pressure and thus excess spills, which we don't currently recover from well.  This should
be re-evaluated in the future if our ability to generate good spills/splits improves.

Partial fix for <rdar://problem/7635585>.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114919 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-27 22:58:54 +00:00
Owen Anderson b0ba0f4170 LoadPRE was not properly checking that the load it was PRE'ing post-dominated the block it was being hoisted to. 
Splitting critical edges at the merge point only addressed part of the issue; it is also possible for non-post-domination
to occur when the path from the load to the merge has branches in it.  Unfortunately, full anticipation analysis is
time-consuming, so for now approximate it.  This is strictly more conservative than real anticipation, so we will miss
some cases that real PRE would allow, but we also no longer insert loads into paths where they didn't exist before. :-)

This is a very slight net positive on SPEC for me (0.5% on average).  Most of the benchmarks are largely unaffected, but
when it pays off it pays off decently: 181.mcf improves by 4.5% on my machine.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114785 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-25 05:26:18 +00:00
Owen Anderson f523e476c2 Revert r114703 and r114702, removing the isConditionalMove flag from instructions. After further 
reflection, this isn't going to achieve the purpose I intended it for.  Back to the drawing board!


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114710 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-23 23:45:25 +00:00
Owen Anderson 71e416ac3a Add isConditionalMove bits to X86 and ARM instructions. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114703 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-23 22:57:01 +00:00
Owen Anderson 5716180b1a Add an TargetInstrDesc bit to indicate that a given instruction is a conditional move. 
Not intended functionality change, as nothing uses this yet.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114702 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-23 22:44:10 +00:00
Owen Anderson ed1088afb5 A select between a constant and zero, when fed by a bit test, can be efficiently 
lowered using a series of shifts.
Fixes <rdar://problem/8285015>.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114599 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-22 22:58:22 +00:00
Owen Anderson 8614167572 Enable target-specific mul-lowering on ARM, even at -Os. Remove a test that this makes 
irrelevant, but add a new test for the new, improved functionality.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114494 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-21 22:51:46 +00:00
Owen Anderson bc146b0a4d Reimplement r114460 in target-independent DAGCombine rather than target-dependent, by using 
the predicate to discover the number of sign bits.  Enhance X86's target lowering to provide
a useful response to this query.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114473 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-21 20:42:50 +00:00
Owen Anderson c004eec71b When adding the carry bit to another value on X86, exploit the fact that the carry-materialization 
(sbbl x, x) sets the registers to 0 or ~0.  Combined with two's complement arithmetic, we can fold
the intermediate AND and the ADD into a single SUB.

This fixes <rdar://problem/8449754>.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114460 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-21 18:41:19 +00:00
Owen Anderson 765fd9c62d CombinerAA is now reordering these stores. 
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@114354 91177308-0d34-0410-b5e6-96231b3b80d8
 2010-09-20 20:56:29 +00:00