mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-01 00:12:50 +01:00
a106725fc5
patch brings numerous advantages to LLVM. One way to look at it is through diffstat: 109 files changed, 3005 insertions(+), 5906 deletions(-) Removing almost 3K lines of code is a good thing. Other advantages include: 1. Value::getType() is a simple load that can be CSE'd, not a mutating union-find operation. 2. Types a uniqued and never move once created, defining away PATypeHolder. 3. Structs can be "named" now, and their name is part of the identity that uniques them. This means that the compiler doesn't merge them structurally which makes the IR much less confusing. 4. Now that there is no way to get a cycle in a type graph without a named struct type, "upreferences" go away. 5. Type refinement is completely gone, which should make LTO much MUCH faster in some common cases with C++ code. 6. Types are now generally immutable, so we can use "Type *" instead "const Type *" everywhere. Downsides of this patch are that it removes some functions from the C API, so people using those will have to upgrade to (not yet added) new API. "LLVM 3.0" is the right time to do this. There are still some cleanups pending after this, this patch is large enough as-is. llvm-svn: 134829
67 lines
2.5 KiB
LLVM
67 lines
2.5 KiB
LLVM
; RUN: llc < %s -march=x86-64 -O0
|
|
; PR8211
|
|
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-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
|
|
target triple = "x86_64-unknown-linux-gnu"
|
|
|
|
module asm "\09.ident\09\22GCC: (GNU) 4.5.2 20100914 (prerelease) LLVM: 114628\22"
|
|
|
|
%"int[]" = type [4 x i32]
|
|
%0 = type { %"int[]" }
|
|
%float = type float
|
|
%"float[]" = type [4 x float]
|
|
%int = type i32
|
|
%"long unsigned int" = type i64
|
|
|
|
define void @swizzle(i8* %a, %0* %b, %0* %c) nounwind {
|
|
entry:
|
|
%a_addr = alloca i8*
|
|
%b_addr = alloca %0*
|
|
%c_addr = alloca %0*
|
|
%"alloca point" = bitcast i32 0 to i32
|
|
store i8* %a, i8** %a_addr
|
|
store %0* %b, %0** %b_addr
|
|
store %0* %c, %0** %c_addr
|
|
%0 = load i8** %a_addr, align 64
|
|
%1 = load %0** %b_addr, align 64
|
|
%2 = load %0** %c_addr, align 64
|
|
%"ssa point" = bitcast i32 0 to i32
|
|
br label %"2"
|
|
|
|
"2": ; preds = %entry
|
|
%3 = bitcast i8* %0 to <2 x i32>*
|
|
%4 = getelementptr inbounds %0* %1, i32 0, i32 0
|
|
%5 = bitcast %"int[]"* %4 to <4 x float>*
|
|
%6 = load <4 x float>* %5, align 16
|
|
%7 = bitcast <2 x i32>* %3 to <2 x float>*
|
|
%8 = bitcast <2 x float>* %7 to double*
|
|
%9 = load double* %8
|
|
%10 = insertelement <2 x double> undef, double %9, i32 0
|
|
%11 = insertelement <2 x double> %10, double undef, i32 1
|
|
%12 = bitcast <2 x double> %11 to <4 x float>
|
|
%13 = shufflevector <4 x float> %6, <4 x float> %12, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
|
|
%14 = getelementptr inbounds %0* %1, i32 0, i32 0
|
|
%15 = bitcast %"int[]"* %14 to <4 x float>*
|
|
store <4 x float> %13, <4 x float>* %15, align 16
|
|
%16 = bitcast i8* %0 to <2 x i32>*
|
|
%17 = bitcast <2 x i32>* %16 to i8*
|
|
%18 = getelementptr i8* %17, i64 8
|
|
%19 = bitcast i8* %18 to <2 x i32>*
|
|
%20 = getelementptr inbounds %0* %2, i32 0, i32 0
|
|
%21 = bitcast %"int[]"* %20 to <4 x float>*
|
|
%22 = load <4 x float>* %21, align 16
|
|
%23 = bitcast <2 x i32>* %19 to <2 x float>*
|
|
%24 = bitcast <2 x float>* %23 to double*
|
|
%25 = load double* %24
|
|
%26 = insertelement <2 x double> undef, double %25, i32 0
|
|
%27 = insertelement <2 x double> %26, double undef, i32 1
|
|
%28 = bitcast <2 x double> %27 to <4 x float>
|
|
%29 = shufflevector <4 x float> %22, <4 x float> %28, <4 x i32> <i32 4, i32 5, i32 2, i32 3>
|
|
%30 = getelementptr inbounds %0* %2, i32 0, i32 0
|
|
%31 = bitcast %"int[]"* %30 to <4 x float>*
|
|
store <4 x float> %29, <4 x float>* %31, align 16
|
|
br label %return
|
|
|
|
return: ; preds = %"2"
|
|
ret void
|
|
}
|