6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-11-10 01:10:48 +00:00
Code Issues Projects Releases Wiki Activity
0309c5d4bc
llvm-6502/test/CodeGen/R600/udivrem24.ll
Jan Vesely 3c57820bbb R600: Use optimized 24bit path in udivrem 
v2: drop enum keyword
    use correct extension mode
    don't bother computing the sign in unsinged case

Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@215462 91177308-0d34-0410-b5e6-96231b3b80d8
2014-08-12 17:31:20 +00:00

245 lines
7.1 KiB
LLVM
Raw Blame History

; RUN: llc -march=r600 -mcpu=SI < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=EG -check-prefix=FUNC %s
; FUNC-LABEL: @udiv24_i8
; SI: V_CVT_F32_UBYTE
; SI: V_CVT_F32_UBYTE
; SI: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @udiv24_i8(i8 addrspace(1)* %out, i8 addrspace(1)* %in) {
%den_ptr = getelementptr i8 addrspace(1)* %in, i8 1
%num = load i8 addrspace(1) * %in
%den = load i8 addrspace(1) * %den_ptr
%result = udiv i8 %num, %den
store i8 %result, i8 addrspace(1)* %out
ret void
}
; FUNC-LABEL: @udiv24_i16
; SI: V_CVT_F32_U32
; SI: V_CVT_F32_U32
; SI: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @udiv24_i16(i16 addrspace(1)* %out, i16 addrspace(1)* %in) {
%den_ptr = getelementptr i16 addrspace(1)* %in, i16 1
%num = load i16 addrspace(1) * %in, align 2
%den = load i16 addrspace(1) * %den_ptr, align 2
%result = udiv i16 %num, %den
store i16 %result, i16 addrspace(1)* %out, align 2
ret void
}
; FUNC-LABEL: @udiv24_i32
; SI: V_CVT_F32_U32
; SI-DAG: V_CVT_F32_U32
; SI-DAG: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @udiv24_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 8
%den.i24.0 = shl i32 %den, 8
%num.i24 = lshr i32 %num.i24.0, 8
%den.i24 = lshr i32 %den.i24.0, 8
%result = udiv i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @udiv25_i32
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @udiv25_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 7
%den.i24.0 = shl i32 %den, 7
%num.i24 = lshr i32 %num.i24.0, 7
%den.i24 = lshr i32 %den.i24.0, 7
%result = udiv i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @test_no_udiv24_i32_1
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @test_no_udiv24_i32_1(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 8
%den.i24.0 = shl i32 %den, 7
%num.i24 = lshr i32 %num.i24.0, 8
%den.i24 = lshr i32 %den.i24.0, 7
%result = udiv i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @test_no_udiv24_i32_2
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @test_no_udiv24_i32_2(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 7
%den.i24.0 = shl i32 %den, 8
%num.i24 = lshr i32 %num.i24.0, 7
%den.i24 = lshr i32 %den.i24.0, 8
%result = udiv i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @urem24_i8
; SI: V_CVT_F32_UBYTE
; SI: V_CVT_F32_UBYTE
; SI: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @urem24_i8(i8 addrspace(1)* %out, i8 addrspace(1)* %in) {
%den_ptr = getelementptr i8 addrspace(1)* %in, i8 1
%num = load i8 addrspace(1) * %in
%den = load i8 addrspace(1) * %den_ptr
%result = urem i8 %num, %den
store i8 %result, i8 addrspace(1)* %out
ret void
}
; FUNC-LABEL: @urem24_i16
; SI: V_CVT_F32_U32
; SI: V_CVT_F32_U32
; SI: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @urem24_i16(i16 addrspace(1)* %out, i16 addrspace(1)* %in) {
%den_ptr = getelementptr i16 addrspace(1)* %in, i16 1
%num = load i16 addrspace(1) * %in, align 2
%den = load i16 addrspace(1) * %den_ptr, align 2
%result = urem i16 %num, %den
store i16 %result, i16 addrspace(1)* %out, align 2
ret void
}
; FUNC-LABEL: @urem24_i32
; SI: V_CVT_F32_U32
; SI: V_CVT_F32_U32
; SI: V_RCP_F32
; SI: V_CVT_U32_F32
; EG: UINT_TO_FLT
; EG-DAG: UINT_TO_FLT
; EG-DAG: RECIP_IEEE
; EG: FLT_TO_UINT
define void @urem24_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 8
%den.i24.0 = shl i32 %den, 8
%num.i24 = lshr i32 %num.i24.0, 8
%den.i24 = lshr i32 %den.i24.0, 8
%result = urem i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @urem25_i32
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @urem25_i32(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 7
%den.i24.0 = shl i32 %den, 7
%num.i24 = lshr i32 %num.i24.0, 7
%den.i24 = lshr i32 %den.i24.0, 7
%result = urem i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @test_no_urem24_i32_1
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @test_no_urem24_i32_1(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 8
%den.i24.0 = shl i32 %den, 7
%num.i24 = lshr i32 %num.i24.0, 8
%den.i24 = lshr i32 %den.i24.0, 7
%result = urem i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
; FUNC-LABEL: @test_no_urem24_i32_2
; RCP_IFLAG is for URECIP in the full 32b alg
; SI: V_RCP_IFLAG
; SI-NOT: V_RCP_F32
; EG-NOT: UINT_TO_FLT
; EG-NOT: RECIP_IEEE
define void @test_no_urem24_i32_2(i32 addrspace(1)* %out, i32 addrspace(1)* %in) {
%den_ptr = getelementptr i32 addrspace(1)* %in, i32 1
%num = load i32 addrspace(1) * %in, align 4
%den = load i32 addrspace(1) * %den_ptr, align 4
%num.i24.0 = shl i32 %num, 7
%den.i24.0 = shl i32 %den, 8
%num.i24 = lshr i32 %num.i24.0, 7
%den.i24 = lshr i32 %den.i24.0, 8
%result = urem i32 %num.i24, %den.i24
store i32 %result, i32 addrspace(1)* %out, align 4
ret void
}
Reference in New Issue View Git Blame Copy Permalink