R600: Reorganize lit tests and document how they should be organized

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179828 91177308-0d34-0410-b5e6-96231b3b80d8

test/CodeGen/R600/README

@@ -0,0 +1,21 @@
++==============================================================================+
+| How to organize the lit tests                                                |
++==============================================================================+
+
+- If you write a test for matching a single DAG opcode or intrinsic, it should
+  go in a file called {opcode_name,intrinsic_name}.ll (e.g. fadd.ll)
+
+- If you write a test that matches several DAG opcodes and checks for a single
+  ISA instruction, then that test should go in a file called {ISA_name}.ll (e.g.
+  bfi_int.ll
+
+- For all other tests, use your best judgement for organizing tests and naming
+  the files.
+
++==============================================================================+
+| Naming conventions                                                           |
++==============================================================================+
+
+- Use dash '-' and not underscore '_' to separate words in file names, unless
+  the file is named after a DAG opcode or ISA instruction that has an
+  underscore '_' in its name.

test/CodeGen/R600/fadd.ll

@@ -1,8 +1,9 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
 
+; CHECK: @fadd_f32
 ; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
 
-define void @test() {
+define void @fadd_f32() {
    %r0 = call float @llvm.R600.load.input(i32 0)
    %r1 = call float @llvm.R600.load.input(i32 1)
    %r2 = fadd float %r0, %r1
@@ -14,3 +15,17 @@ declare float @llvm.R600.load.input(i32) readnone
 
 declare void @llvm.AMDGPU.store.output(float, i32)
 
+; CHECK: @fadd_v4f32
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @fadd_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %b_ptr = getelementptr <4 x float> addrspace(1)* %in, i32 1
+  %a = load <4 x float> addrspace(1) * %in
+  %b = load <4 x float> addrspace(1) * %b_ptr
+  %result = fadd <4 x float> %a, %b
+  store <4 x float> %result, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/fadd.v4f32.ll

@@ -1,15 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @test(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
-  %b_ptr = getelementptr <4 x float> addrspace(1)* %in, i32 1
-  %a = load <4 x float> addrspace(1) * %in
-  %b = load <4 x float> addrspace(1) * %b_ptr
-  %result = fadd <4 x float> %a, %b
-  store <4 x float> %result, <4 x float> addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/fmul.ll

@@ -1,8 +1,9 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
 
+; CHECK: @fmul_f32
 ; CHECK: MUL_IEEE T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
 
-define void @test() {
+define void @fmul_f32() {
    %r0 = call float @llvm.R600.load.input(i32 0)
    %r1 = call float @llvm.R600.load.input(i32 1)
    %r2 = fmul float %r0, %r1
@@ -14,3 +15,17 @@ declare float @llvm.R600.load.input(i32) readnone
 
 declare void @llvm.AMDGPU.store.output(float, i32)
 
+; CHECK: @fmul_v4f32
+; CHECK: MUL_IEEE T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: MUL_IEEE T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: MUL_IEEE T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: MUL_IEEE T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @fmul_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %b_ptr = getelementptr <4 x float> addrspace(1)* %in, i32 1
+  %a = load <4 x float> addrspace(1) * %in
+  %b = load <4 x float> addrspace(1) * %b_ptr
+  %result = fmul <4 x float> %a, %b
+  store <4 x float> %result, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/fp_to_sint.ll

@@ -0,0 +1,14 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+
+; CHECK: @fp_to_sint_v4i32
+; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @fp_to_sint_v4i32(<4 x i32> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %value = load <4 x float> addrspace(1) * %in
+  %result = fptosi <4 x float> %value to <4 x i32>
+  store <4 x i32> %result, <4 x i32> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/fp_to_uint.ll

@@ -0,0 +1,14 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+
+; CHECK: @fp_to_uint_v4i32
+; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @fp_to_uint_v4i32(<4 x i32> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %value = load <4 x float> addrspace(1) * %in
+  %result = fptoui <4 x float> %value to <4 x i32>
+  store <4 x i32> %result, <4 x i32> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/fsub.ll

@@ -1,8 +1,9 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
 
+; CHECK: @fsub_f32
 ; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], -T[0-9]+\.[XYZW]}}
 
-define void @test() {
+define void @fsub_f32() {
    %r0 = call float @llvm.R600.load.input(i32 0)
    %r1 = call float @llvm.R600.load.input(i32 1)
    %r2 = fsub float %r0, %r1
@@ -14,3 +15,17 @@ declare float @llvm.R600.load.input(i32) readnone
 
 declare void @llvm.AMDGPU.store.output(float, i32)
 
+; CHECK: @fsub_v4f32
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @fsub_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %b_ptr = getelementptr <4 x float> addrspace(1)* %in, i32 1
+  %a = load <4 x float> addrspace(1) * %in
+  %b = load <4 x float> addrspace(1) * %b_ptr
+  %result = fsub <4 x float> %a, %b
+  store <4 x float> %result, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/fsub.v4f32.ll

@@ -1,15 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-;CHECK: ADD T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @test(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
-  %b_ptr = getelementptr <4 x float> addrspace(1)* %in, i32 1
-  %a = load <4 x float> addrspace(1) * %in
-  %b = load <4 x float> addrspace(1) * %b_ptr
-  %result = fsub <4 x float> %a, %b
-  store <4 x float> %result, <4 x float> addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/load.constant_addrspace.f32.ll

@@ -1,9 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: VTX_READ_32 T{{[0-9]+\.X, T[0-9]+\.X}}
-
-define void @test(float addrspace(1)* %out, float addrspace(2)* %in) {
-  %1 = load float addrspace(2)* %in
-  store float %1, float addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/load.i8.ll

@@ -1,10 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: VTX_READ_8 T{{[0-9]+\.X, T[0-9]+\.X}}
-
-define void @test(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
-  %1 = load i8 addrspace(1)* %in
-  %2 = zext i8 %1 to i32
-  store i32 %2, i32 addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/load.ll

@@ -0,0 +1,20 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+
+; Load an i8 value from the global address space.
+; CHECK: VTX_READ_8 T{{[0-9]+\.X, T[0-9]+\.X}}
+
+define void @load_i8(i32 addrspace(1)* %out, i8 addrspace(1)* %in) {
+  %1 = load i8 addrspace(1)* %in
+  %2 = zext i8 %1 to i32
+  store i32 %2, i32 addrspace(1)* %out
+  ret void
+}
+
+; Load a f32 value from the constant address space.
+; CHECK: VTX_READ_32 T{{[0-9]+\.X, T[0-9]+\.X}}
+
+define void @load_const_addrspace_f32(float addrspace(1)* %out, float addrspace(2)* %in) {
+  %1 = load float addrspace(2)* %in
+  store float %1, float addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/sint_to_fp.ll

@@ -0,0 +1,14 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+
+; CHECK: @sint_to_fp_v4i32
+; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @sint_to_fp_v4i32(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
+  %value = load <4 x i32> addrspace(1) * %in
+  %result = sitofp <4 x i32> %value to <4 x float>
+  store <4 x float> %result, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/store.ll

@@ -1,11 +1,13 @@
 ; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck --check-prefix=EG-CHECK %s
 ; RUN: llc < %s -march=r600 -mcpu=verde | FileCheck --check-prefix=SI-CHECK %s
 
-; CHECK: @store_float
+; floating-point store
+; EG-CHECK: @store_f32
 ; EG-CHECK: RAT_WRITE_CACHELESS_32_eg T{{[0-9]+\.X, T[0-9]+\.X}}, 1
+; SI-CHECK: @store_f32
 ; SI-CHECK: BUFFER_STORE_DWORD
 
-define void @store_float(float addrspace(1)* %out, float %in) {
+define void @store_f32(float addrspace(1)* %out, float %in) {
   store float %in, float addrspace(1)* %out
   ret void
 }

test/CodeGen/R600/store.r600.ll

@@ -0,0 +1,22 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck --check-prefix=EG-CHECK %s
+
+; XXX: Merge this test into store.ll once it is supported on SI
+
+; v4i32 store
+; EG-CHECK: @store_v4i32
+; EG-CHECK: RAT_WRITE_CACHELESS_128 T{{[0-9]+\.XYZW, T[0-9]+\.X}}, 1
+
+define void @store_v4i32(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
+  %1 = load <4 x i32> addrspace(1) * %in
+  store <4 x i32> %1, <4 x i32> addrspace(1)* %out
+  ret void
+}
+
+; v4f32 store
+; EG-CHECK: @store_v4f32
+; EG-CHECK: RAT_WRITE_CACHELESS_128 T{{[0-9]+\.XYZW, T[0-9]+\.X}}, 1
+define void @store_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
+  %1 = load <4 x float> addrspace(1) * %in
+  store <4 x float> %1, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/store.v4f32.ll

@@ -1,9 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: RAT_WRITE_CACHELESS_128 T{{[0-9]+\.XYZW, T[0-9]+\.X}}, 1
-
-define void @test(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
-  %1 = load <4 x float> addrspace(1) * %in
-  store <4 x float> %1, <4 x float> addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/store.v4i32.ll

@@ -1,9 +0,0 @@
-;RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-;CHECK: RAT_WRITE_CACHELESS_128 T{{[0-9]+\.XYZW, T[0-9]+\.X}}, 1
-
-define void @test(<4 x i32> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
-  %1 = load <4 x i32> addrspace(1) * %in
-  store <4 x i32> %1, <4 x i32> addrspace(1)* %out
-  ret void
-}

test/CodeGen/R600/uint_to_fp.ll

@@ -0,0 +1,14 @@
+; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
+
+; CHECK: @uint_to_fp_v4i32
+; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
+
+define void @uint_to_fp_v4i32(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
+  %value = load <4 x i32> addrspace(1) * %in
+  %result = uitofp <4 x i32> %value to <4 x float>
+  store <4 x float> %result, <4 x float> addrspace(1)* %out
+  ret void
+}

test/CodeGen/R600/vec4-expand.ll

@@ -1,53 +0,0 @@
-; RUN: llc < %s -march=r600 -mcpu=redwood | FileCheck %s
-
-; CHECK: @fp_to_sint
-; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_INT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @fp_to_sint(<4 x i32> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
-  %value = load <4 x float> addrspace(1) * %in
-  %result = fptosi <4 x float> %value to <4 x i32>
-  store <4 x i32> %result, <4 x i32> addrspace(1)* %out
-  ret void
-}
-
-; CHECK: @fp_to_uint
-; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: FLT_TO_UINT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @fp_to_uint(<4 x i32> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
-  %value = load <4 x float> addrspace(1) * %in
-  %result = fptoui <4 x float> %value to <4 x i32>
-  store <4 x i32> %result, <4 x i32> addrspace(1)* %out
-  ret void
-}
-
-; CHECK: @sint_to_fp
-; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: INT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @sint_to_fp(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
-  %value = load <4 x i32> addrspace(1) * %in
-  %result = sitofp <4 x i32> %value to <4 x float>
-  store <4 x float> %result, <4 x float> addrspace(1)* %out
-  ret void
-}
-
-; CHECK: @uint_to_fp
-; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-; CHECK: UINT_TO_FLT T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW]}}
-
-define void @uint_to_fp(<4 x float> addrspace(1)* %out, <4 x i32> addrspace(1)* %in) {
-  %value = load <4 x i32> addrspace(1) * %in
-  %result = uitofp <4 x i32> %value to <4 x float>
-  store <4 x float> %result, <4 x float> addrspace(1)* %out
-  ret void
-}