diff options
| author | Andrew Trick <atrick@apple.com> | 2011-04-13 00:38:32 +0000 |
|---|---|---|
| committer | Andrew Trick <atrick@apple.com> | 2011-04-13 00:38:32 +0000 |
| commit | 87896d9368e08d93493427ce7bf8272d1e5cca35 (patch) | |
| tree | d3183747b2917bf4e1254c7a331ff86fd3352a2f /test/CodeGen/ARM/memcpy-inline.ll | |
| parent | f93f7b2446bec3febc30b7136e18704664bd98cc (diff) | |
| download | llvm-87896d9368e08d93493427ce7bf8272d1e5cca35.tar.gz llvm-87896d9368e08d93493427ce7bf8272d1e5cca35.tar.bz2 llvm-87896d9368e08d93493427ce7bf8272d1e5cca35.tar.xz | |
Recommit r129383. PreRA scheduler heuristic fixes: VRegCycle, TokenFactor latency.
Additional fixes:
Do something reasonable for subtargets with generic
itineraries by handle node latency the same as for an empty
itinerary. Now nodes default to unit latency unless an itinerary
explicitly specifies a zero cycle stage or it is a TokenFactor chain.
Original fixes:
UnitsSharePred was a source of randomness in the scheduler: node
priority depended on the queue data structure. I rewrote the recent
VRegCycle heuristics to completely replace the old heuristic without
any randomness. To make the ndoe latency adjustments work, I also
needed to do something a little more reasonable with TokenFactor. I
gave it zero latency to its consumers and always schedule it as low as
possible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129421 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/ARM/memcpy-inline.ll')
| -rw-r--r-- | test/CodeGen/ARM/memcpy-inline.ll | 8 |
1 files changed, 3 insertions, 5 deletions
diff --git a/test/CodeGen/ARM/memcpy-inline.ll b/test/CodeGen/ARM/memcpy-inline.ll index e8a2a3b7d5..5bae037caf 100644 --- a/test/CodeGen/ARM/memcpy-inline.ll +++ b/test/CodeGen/ARM/memcpy-inline.ll @@ -1,10 +1,8 @@ -; RUN: llc < %s -mtriple=arm-apple-darwin -regalloc=linearscan -disable-post-ra | FileCheck %s -; RUN: llc < %s -mtriple=arm-apple-darwin -regalloc=basic -disable-post-ra | FileCheck %s +; RUN: llc < %s -mtriple=thumbv7-apple-darwin -regalloc=linearscan -disable-post-ra | FileCheck %s ; The ARM magic hinting works best with linear scan. -; CHECK: ldmia -; CHECK: stmia -; CHECK: ldrh +; CHECK: ldrd +; CHECK: strd ; CHECK: ldrb %struct.x = type { i8, i8, i8, i8, i8, i8, i8, i8, i8, i8, i8 } |