RPCSX/llvm
1
0
Fork 0
mirror of https://github.com/RPCSX/llvm.git synced 2024-12-12 22:26:14 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Enable runtime unrolling with unroll pragma metadata

Browse Source 
Enable runtime unrolling for loops with unroll count metadata ("#pragma unroll N")
and a runtime trip count. Also, do not unroll loops with unroll full metadata if the
loop has a runtime loop count. Previously, such loops would be unrolled with a
very large threshold (pragma-unroll-threshold) if runtime unrolled happened to be
enabled resulting in a very large (and likely unwise) unroll factor.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@242047 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Mark Heffernan 2015-07-13 18:26:27 +00:00
parent 8f5464d3ab
commit 8a9e01d606
2 changed files with 17 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
lib/Transforms/Scalar/LoopUnrollPass.cpp
View File

@ -840,8 +840,10 @@ bool LoopUnroll::runOnLoop(Loop *L, LPPassManager &LPM) {
// Reduce count based on the type of unrolling and the threshold values.
unsigned OriginalCount = Count;
bool AllowRuntime = UserRuntime ? CurrentRuntime : UP.Runtime;
if (HasRuntimeUnrollDisablePragma(L)) {
bool AllowRuntime =
(PragmaCount > 0) || (UserRuntime ? CurrentRuntime : UP.Runtime);
// Don't unroll a runtime trip count loop with unroll full pragma.
if (HasRuntimeUnrollDisablePragma(L) || PragmaFullUnroll) {
AllowRuntime = false;
}
if (Unrolling == Partial) {

27
test/Transforms/LoopUnroll/unroll-pragmas.ll
View File

@ -86,9 +86,9 @@ for.end: ; preds = %for.body
; #pragma clang loop unroll(full)
; Loop should be fully unrolled.
;
; CHECK-LABEL: @loop64_with_enable(
; CHECK-LABEL: @loop64_with_full(
; CHECK-NOT: br i1
define void @loop64_with_enable(i32* nocapture %a) {
define void @loop64_with_full(i32* nocapture %a) {
entry:
br label %for.body
@ -139,14 +139,13 @@ for.end: ; preds = %for.body
!6 = !{!"llvm.loop.unroll.count", i32 4}
; #pragma clang loop unroll(full)
; Full unrolling is requested, but loop has a dynamic trip count so
; Full unrolling is requested, but loop has a runtime trip count so
; no unrolling should occur.
;
; CHECK-LABEL: @dynamic_loop_with_enable(
; CHECK-LABEL: @runtime_loop_with_full(
; CHECK: store i32
; CHECK-NOT: store i32
; CHECK: br i1
define void @dynamic_loop_with_enable(i32* nocapture %a, i32 %b) {
define void @runtime_loop_with_full(i32* nocapture %a, i32 %b) {
entry:
%cmp3 = icmp sgt i32 %b, 0
br i1 %cmp3, label %for.body, label %for.end, !llvm.loop !8
@ -168,22 +167,22 @@ for.end: ; preds = %for.body, %entry
!8 = !{!8, !4}
; #pragma clang loop unroll_count(4)
; Loop has a dynamic trip count. Unrolling should occur, but no
; conditional branches can be removed.
; Loop has a runtime trip count. Runtime unrolling should occur and loop
; should be duplicated (original and 4x unrolled).
;
; CHECK-LABEL: @dynamic_loop_with_count4(
; CHECK-LABEL: @runtime_loop_with_count4(
; CHECK: for.body.prol:
; CHECK: store
; CHECK-NOT: store
; CHECK: br i1
; CHECK: for.body
; CHECK: store
; CHECK: br i1
; CHECK: store
; CHECK: br i1
; CHECK: store
; CHECK: br i1
; CHECK: store
; CHECK-NOT: store
; CHECK: br i1
; CHECK-NOT: br i1
define void @dynamic_loop_with_count4(i32* nocapture %a, i32 %b) {
define void @runtime_loop_with_count4(i32* nocapture %a, i32 %b) {
entry:
%cmp3 = icmp sgt i32 %b, 0
br i1 %cmp3, label %for.body, label %for.end, !llvm.loop !9