llvm/test/CodeGen/PowerPC/optcmp.ll
Hal Finkel 36e1825e68 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

153 lines
3.6 KiB
LLVM

; RUN: llc < %s -mtriple=powerpc64-unknown-linux-gnu -mcpu=a2 -mattr=-crbits -disable-ppc-cmp-opt=0 | FileCheck %s
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"
define signext i32 @foo(i32 signext %a, i32 signext %b, i32* nocapture %c) #0 {
entry:
%sub = sub nsw i32 %a, %b
store i32 %sub, i32* %c, align 4
%cmp = icmp sgt i32 %a, %b
%cond = select i1 %cmp, i32 %a, i32 %b
ret i32 %cond
; CHECK: @foo
; CHECK-NOT: subf.
}
define signext i32 @foo2(i32 signext %a, i32 signext %b, i32* nocapture %c) #0 {
entry:
%shl = shl i32 %a, %b
store i32 %shl, i32* %c, align 4
%cmp = icmp sgt i32 %shl, 0
%conv = zext i1 %cmp to i32
ret i32 %conv
; CHECK: @foo2
; CHECK-NOT: slw.
}
define i64 @fool(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8
%cmp = icmp sgt i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @fool
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 3, 4, 1
; CHECK: std [[REG]], 0(5)
}
define i64 @foolb(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8
%cmp = icmp sle i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foolb
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 4, 3, 1
; CHECK: std [[REG]], 0(5)
}
define i64 @foolc(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %b, %a
store i64 %sub, i64* %c, align 8
%cmp = icmp sgt i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foolc
; CHECK: subf. [[REG:[0-9]+]], 3, 4
; CHECK: isel 3, 3, 4, 0
; CHECK: std [[REG]], 0(5)
}
define i64 @foold(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %b, %a
store i64 %sub, i64* %c, align 8
%cmp = icmp eq i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foold
; CHECK: subf. [[REG:[0-9]+]], 3, 4
; CHECK: isel 3, 3, 4, 2
; CHECK: std [[REG]], 0(5)
}
define i64 @foold2(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
store i64 %sub, i64* %c, align 8
%cmp = icmp eq i64 %a, %b
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @foold2
; CHECK: subf. [[REG:[0-9]+]], 4, 3
; CHECK: isel 3, 3, 4, 2
; CHECK: std [[REG]], 0(5)
}
define i64 @foo2l(i64 %a, i64 %b, i64* nocapture %c) #0 {
entry:
%shl = shl i64 %a, %b
store i64 %shl, i64* %c, align 8
%cmp = icmp sgt i64 %shl, 0
%conv1 = zext i1 %cmp to i64
ret i64 %conv1
; CHECK: @foo2l
; CHECK: sld. 4, 3, 4
; CHECK: std 4, 0(5)
}
define double @food(double %a, double %b, double* nocapture %c) #0 {
entry:
%sub = fsub double %a, %b
store double %sub, double* %c, align 8
%cmp = fcmp ogt double %a, %b
%cond = select i1 %cmp, double %a, double %b
ret double %cond
; CHECK: @food
; CHECK-NOT: fsub. 0, 1, 2
; CHECK: stfd 0, 0(5)
}
define float @foof(float %a, float %b, float* nocapture %c) #0 {
entry:
%sub = fsub float %a, %b
store float %sub, float* %c, align 4
%cmp = fcmp ogt float %a, %b
%cond = select i1 %cmp, float %a, float %b
ret float %cond
; CHECK: @foof
; CHECK-NOT: fsubs. 0, 1, 2
; CHECK: stfs 0, 0(5)
}
declare i64 @llvm.ctpop.i64(i64);
define signext i64 @fooct(i64 signext %a, i64 signext %b, i64* nocapture %c) #0 {
entry:
%sub = sub nsw i64 %a, %b
%subc = call i64 @llvm.ctpop.i64(i64 %sub)
store i64 %subc, i64* %c, align 4
%cmp = icmp sgt i64 %subc, 0
%cond = select i1 %cmp, i64 %a, i64 %b
ret i64 %cond
; CHECK: @fooct
; CHECK-NOT: popcntd.
}