6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-11-05 13:09:10 +00:00
Code Issues Projects Releases Wiki Activity
0c57babfc6
llvm-6502/test/CodeGen/PowerPC/bdzlr.ll

67 lines
4.0 KiB
LLVM
Raw Normal View History

Add CR-bit tracking to the PowerPC backend for i1 values This change enables tracking i1 values in the PowerPC backend using the condition register bits. These bits can be treated on PowerPC as separate registers; individual bit operations (and, or, xor, etc.) are supported. Tracking booleans in CR bits has several advantages: - Reduction in register pressure (because we no longer need GPRs to store boolean values). - Logical operations on booleans can be handled more efficiently; we used to have to move all results from comparisons into GPRs, perform promoted logical operations in GPRs, and then move the result back into condition register bits to be used by conditional branches. This can be very inefficient, because the throughput of these CR <-> GPR moves have high latency and low throughput (especially when other associated instructions are accounted for). - On the POWER7 and similar cores, we can increase total throughput by using the CR bits. CR bit operations have a dedicated functional unit. Most of this is more-or-less mechanical: Adjustments were needed in the calling-convention code, support was added for spilling/restoring individual condition-register bits, and conditional branch instruction definitions taking specific CR bits were added (plus patterns and code for generating bit-level operations). This is enabled by default when running at -O2 and higher. For -O0 and -O1, where the ability to debug is more important, this feature is disabled by default. Individual CR bits do not have assigned DWARF register numbers, and storing values in CR bits makes them invisible to the debugger. It is critical, however, that we don't move i1 values that have been promoted to larger values (such as those passed as function arguments) into bit registers only to quickly turn around and move the values back into GPRs (such as happens when values are returned by functions). A pair of target-specific DAG combines are added to remove the trunc/extends in: trunc(binary-ops(binary-ops(zext(x), zext(y)), ...) and: zext(binary-ops(binary-ops(trunc(x), trunc(y)), ...) In short, we only want to use CR bits where some of the i1 values come from comparisons or are used by conditional branches or selects. To put it another way, if we can do the entire i1 computation in GPRs, then we probably should (on the POWER7, the GPR-operation throughput is higher, and for all cores, the CR <-> GPR moves are expensive). POWER7 test-suite performance results (from 10 runs in each configuration): SingleSource/Benchmarks/Misc/mandel-2: 35% speedup MultiSource/Benchmarks/Prolangs-C++/city/city: 21% speedup MultiSource/Benchmarks/MiBench/automotive-susan: 23% speedup SingleSource/Benchmarks/CoyoteBench/huffbench: 13% speedup SingleSource/Benchmarks/Misc-C++/Large/sphereflake: 13% speedup SingleSource/Benchmarks/Misc-C++/mandel-text: 10% speedup SingleSource/Benchmarks/Misc-C++-EH/spirit: 10% slowdown MultiSource/Applications/lemon/lemon: 8% slowdown git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@202451 91177308-0d34-0410-b5e6-96231b3b80d8
2014-02-28 00:27:01 +00:00
; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 -mattr=-crbits | FileCheck %s
; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 | FileCheck %s -check-prefix=CHECK-CRB
Allow PPC B and BLR to be if-converted into some predicated forms This enables us to form predicated branches (which are the same conditional branches we had before) and also a larger set of predicated returns (including instructions like bdnzlr which is a conditional return and loop-counter decrement all in one). At the moment, if conversion does not capture all possible opportunities. A simple example is provided in early-ret2.ll, where if conversion forms one predicated return, and then the PPCEarlyReturn pass picks up the other one. So, at least for now, we'll keep both mechanisms. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179134 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-09 22:58:37 +00:00
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
%struct.lua_TValue.17.692 = type { %union.Value.16.691, i32 }
%union.Value.16.691 = type { %union.GCObject.15.690* }
%union.GCObject.15.690 = type { %struct.lua_State.14.689 }
%struct.lua_State.14.689 = type { %union.GCObject.15.690*, i8, i8, i8, %struct.lua_TValue.17.692*, %struct.lua_TValue.17.692*, %struct.global_State.10.685*, %struct.CallInfo.11.686*, i32*, %struct.lua_TValue.17.692*, %struct.lua_TValue.17.692*, %struct.CallInfo.11.686*, %struct.CallInfo.11.686*, i32, i32, i16, i16, i8, i8, i32, i32, void (%struct.lua_State.14.689*, %struct.lua_Debug.12.687*)*, %struct.lua_TValue.17.692, %struct.lua_TValue.17.692, %union.GCObject.15.690*, %union.GCObject.15.690*, %struct.lua_longjmp.13.688*, i64 }
%struct.global_State.10.685 = type { %struct.stringtable.0.675, i8* (i8*, i8*, i64, i64)*, i8*, i8, i8, i32, %union.GCObject.15.690*, %union.GCObject.15.690**, %union.GCObject.15.690*, %union.GCObject.15.690*, %union.GCObject.15.690*, %union.GCObject.15.690*, %struct.Mbuffer.1.676, i64, i64, i64, i64, i32, i32, i32 (%struct.lua_State.14.689*)*, %struct.lua_TValue.17.692, %struct.lua_State.14.689*, %struct.UpVal.3.678, [9 x %struct.Table.7.682*], [17 x %union.TString.9.684*] }
%struct.stringtable.0.675 = type { %union.GCObject.15.690**, i32, i32 }
%struct.Mbuffer.1.676 = type { i8*, i64, i64 }
%struct.UpVal.3.678 = type { %union.GCObject.15.690*, i8, i8, %struct.lua_TValue.17.692*, %union.anon.2.677 }
%union.anon.2.677 = type { %struct.lua_TValue.17.692 }
%struct.Table.7.682 = type { %union.GCObject.15.690*, i8, i8, i8, i8, %struct.Table.7.682*, %struct.lua_TValue.17.692*, %struct.Node.6.681*, %struct.Node.6.681*, %union.GCObject.15.690*, i32 }
%struct.Node.6.681 = type { %struct.lua_TValue.17.692, %union.TKey.5.680 }
%union.TKey.5.680 = type { %struct.anon.0.4.679 }
%struct.anon.0.4.679 = type { %union.Value.16.691, i32, %struct.Node.6.681* }
%union.TString.9.684 = type { %struct.anon.1.8.683 }
%struct.anon.1.8.683 = type { %union.GCObject.15.690*, i8, i8, i8, i32, i64 }
%struct.CallInfo.11.686 = type { %struct.lua_TValue.17.692*, %struct.lua_TValue.17.692*, %struct.lua_TValue.17.692*, i32*, i32, i32 }
%struct.lua_Debug.12.687 = type { i32, i8*, i8*, i8*, i8*, i32, i32, i32, i32, [60 x i8], i32 }
%struct.lua_longjmp.13.688 = type opaque
define void @lua_xmove(i32 signext %n) #0 {
entry:
br i1 undef, label %for.end, label %if.end
if.end: ; preds = %entry
br i1 undef, label %for.body.lr.ph, label %for.end
for.body.lr.ph: ; preds = %if.end
br label %for.body
for.body: ; preds = %for.body.for.body_crit_edge, %for.body.lr.ph
%0 = phi %struct.lua_TValue.17.692* [ undef, %for.body.lr.ph ], [ %.pre, %for.body.for.body_crit_edge ]
%indvars.iv = phi i64 [ 0, %for.body.lr.ph ], [ %indvars.iv.next, %for.body.for.body_crit_edge ]
%tt = getelementptr inbounds %struct.lua_TValue.17.692* %0, i64 %indvars.iv, i32 1
TBAA: remove !tbaa from testing cases when they are not needed. This will make it easier to turn on struct-path aware TBAA since the metadata format will change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188944 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-21 22:20:53 +00:00
%1 = load i32* %tt, align 4
store i32 %1, i32* undef, align 4
Allow PPC B and BLR to be if-converted into some predicated forms This enables us to form predicated branches (which are the same conditional branches we had before) and also a larger set of predicated returns (including instructions like bdnzlr which is a conditional return and loop-counter decrement all in one). At the moment, if conversion does not capture all possible opportunities. A simple example is provided in early-ret2.ll, where if conversion forms one predicated return, and then the PPCEarlyReturn pass picks up the other one. So, at least for now, we'll keep both mechanisms. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179134 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-09 22:58:37 +00:00
%indvars.iv.next = add i64 %indvars.iv, 1
%lftr.wideiv = trunc i64 %indvars.iv.next to i32
%exitcond = icmp eq i32 %lftr.wideiv, %n
br i1 %exitcond, label %for.end, label %for.body.for.body_crit_edge
for.body.for.body_crit_edge: ; preds = %for.body
TBAA: remove !tbaa from testing cases when they are not needed. This will make it easier to turn on struct-path aware TBAA since the metadata format will change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@188944 91177308-0d34-0410-b5e6-96231b3b80d8
2013-08-21 22:20:53 +00:00
%.pre = load %struct.lua_TValue.17.692** undef, align 8
Allow PPC B and BLR to be if-converted into some predicated forms This enables us to form predicated branches (which are the same conditional branches we had before) and also a larger set of predicated returns (including instructions like bdnzlr which is a conditional return and loop-counter decrement all in one). At the moment, if conversion does not capture all possible opportunities. A simple example is provided in early-ret2.ll, where if conversion forms one predicated return, and then the PPCEarlyReturn pass picks up the other one. So, at least for now, we'll keep both mechanisms. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179134 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-09 22:58:37 +00:00
br label %for.body
for.end: ; preds = %for.body, %if.end, %entry
ret void
; CHECK: @lua_xmove
; CHECK: bnelr
; CHECK: bnelr
; CHECK: bdzlr
Manually remove successors in if conversion when CopyAndPredicateBlock is used In the simple and triangle if-conversion cases, when CopyAndPredicateBlock is used because the to-be-predicated block has other predecessors, we need to explicitly remove the old copied block from the successors list. Normally if conversion relies on TII->AnalyzeBranch combined with BB->CorrectExtraCFGEdges to cleanup the successors list, but if the predicated block contained an un-analyzable branch (such as a now-predicated return), then this will fail. These extra successors were causing a problem on PPC because it was causing later passes (such as PPCEarlyReturm) to leave dead return-only basic blocks in the code. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179227 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-10 22:05:25 +00:00
; CHECK-NOT: blr
Add CR-bit tracking to the PowerPC backend for i1 values This change enables tracking i1 values in the PowerPC backend using the condition register bits. These bits can be treated on PowerPC as separate registers; individual bit operations (and, or, xor, etc.) are supported. Tracking booleans in CR bits has several advantages: - Reduction in register pressure (because we no longer need GPRs to store boolean values). - Logical operations on booleans can be handled more efficiently; we used to have to move all results from comparisons into GPRs, perform promoted logical operations in GPRs, and then move the result back into condition register bits to be used by conditional branches. This can be very inefficient, because the throughput of these CR <-> GPR moves have high latency and low throughput (especially when other associated instructions are accounted for). - On the POWER7 and similar cores, we can increase total throughput by using the CR bits. CR bit operations have a dedicated functional unit. Most of this is more-or-less mechanical: Adjustments were needed in the calling-convention code, support was added for spilling/restoring individual condition-register bits, and conditional branch instruction definitions taking specific CR bits were added (plus patterns and code for generating bit-level operations). This is enabled by default when running at -O2 and higher. For -O0 and -O1, where the ability to debug is more important, this feature is disabled by default. Individual CR bits do not have assigned DWARF register numbers, and storing values in CR bits makes them invisible to the debugger. It is critical, however, that we don't move i1 values that have been promoted to larger values (such as those passed as function arguments) into bit registers only to quickly turn around and move the values back into GPRs (such as happens when values are returned by functions). A pair of target-specific DAG combines are added to remove the trunc/extends in: trunc(binary-ops(binary-ops(zext(x), zext(y)), ...) and: zext(binary-ops(binary-ops(trunc(x), trunc(y)), ...) In short, we only want to use CR bits where some of the i1 values come from comparisons or are used by conditional branches or selects. To put it another way, if we can do the entire i1 computation in GPRs, then we probably should (on the POWER7, the GPR-operation throughput is higher, and for all cores, the CR <-> GPR moves are expensive). POWER7 test-suite performance results (from 10 runs in each configuration): SingleSource/Benchmarks/Misc/mandel-2: 35% speedup MultiSource/Benchmarks/Prolangs-C++/city/city: 21% speedup MultiSource/Benchmarks/MiBench/automotive-susan: 23% speedup SingleSource/Benchmarks/CoyoteBench/huffbench: 13% speedup SingleSource/Benchmarks/Misc-C++/Large/sphereflake: 13% speedup SingleSource/Benchmarks/Misc-C++/mandel-text: 10% speedup SingleSource/Benchmarks/Misc-C++-EH/spirit: 10% slowdown MultiSource/Applications/lemon/lemon: 8% slowdown git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@202451 91177308-0d34-0410-b5e6-96231b3b80d8
2014-02-28 00:27:01 +00:00
; CHECK-CRB: @lua_xmove
; CHECK-CRB: bclr 12,
; CHECK-CRB: bclr 12,
; CHECK-CRB: bdzlr
; CHECK-CRB-NOT: blr
Allow PPC B and BLR to be if-converted into some predicated forms This enables us to form predicated branches (which are the same conditional branches we had before) and also a larger set of predicated returns (including instructions like bdnzlr which is a conditional return and loop-counter decrement all in one). At the moment, if conversion does not capture all possible opportunities. A simple example is provided in early-ret2.ll, where if conversion forms one predicated return, and then the PPCEarlyReturn pass picks up the other one. So, at least for now, we'll keep both mechanisms. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179134 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-09 22:58:37 +00:00
}
attributes #0 = { nounwind }
Reference in New Issue Copy Permalink