https://github.com/halide/Halide
Tip revision: d7cf9bcb11c9b7f0c2101e57b1439491313152fa authored by Steven Johnson on 09 October 2020, 16:25:38 UTC
Merge branch 'master' into abadams/nested_vectorization_tweaks
Merge branch 'master' into abadams/nested_vectorization_tweaks
Tip revision: d7cf9bc
arm.ll
; Absolute value ops
declare <4 x float> @llvm.fabs.v4f32(<4 x float>) nounwind readnone
declare <2 x float> @llvm.fabs.v2f32(<2 x float>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vabs.v4i32(<4 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vabs.v2i32(<2 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vabs.v4i16(<4 x i16>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vabs.v8i16(<8 x i16>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vabs.v8i8(<8 x i8>) nounwind readnone
declare <16 x i8> @llvm.arm.neon.vabs.v16i8(<16 x i8>) nounwind readnone
define weak_odr <4 x float> @abs_f32x4(<4 x float> %x) nounwind alwaysinline {
%tmp = call <4 x float> @llvm.fabs.v4f32(<4 x float> %x)
ret <4 x float> %tmp
}
define weak_odr <2 x float> @abs_f32x2(<2 x float> %x) nounwind alwaysinline {
%tmp = call <2 x float> @llvm.fabs.v2f32(<2 x float> %x)
ret <2 x float> %tmp
}
define weak_odr <4 x i32> @abs_i32x4(<4 x i32> %x) nounwind alwaysinline {
%tmp = call <4 x i32> @llvm.arm.neon.vabs.v4i32(<4 x i32> %x)
ret <4 x i32> %tmp
}
define weak_odr <2 x i32> @abs_i32x2(<2 x i32> %x) nounwind alwaysinline {
%tmp = call <2 x i32> @llvm.arm.neon.vabs.v2i32(<2 x i32> %x)
ret <2 x i32> %tmp
}
define weak_odr <4 x i16> @abs_i16x4(<4 x i16> %x) nounwind alwaysinline {
%tmp = call <4 x i16> @llvm.arm.neon.vabs.v4i16(<4 x i16> %x)
ret <4 x i16> %tmp
}
define weak_odr <8 x i16> @abs_i16x8(<8 x i16> %x) nounwind alwaysinline {
%tmp = call <8 x i16> @llvm.arm.neon.vabs.v8i16(<8 x i16> %x)
ret <8 x i16> %tmp
}
define weak_odr <8 x i8> @abs_i8x8(<8 x i8> %x) nounwind alwaysinline {
%tmp = call <8 x i8> @llvm.arm.neon.vabs.v8i8(<8 x i8> %x)
ret <8 x i8> %tmp
}
define weak_odr <16 x i8> @abs_i8x16(<16 x i8> %x) nounwind alwaysinline {
%tmp = call <16 x i8> @llvm.arm.neon.vabs.v16i8(<16 x i8> %x)
ret <16 x i8> %tmp
}
declare <8 x i8> @llvm.arm.neon.vabds.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vabdu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vabds.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vabdu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vabds.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vabdu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <16 x i8> @llvm.arm.neon.vabds.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
declare <16 x i8> @llvm.arm.neon.vabdu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vabds.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vabdu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vabds.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vabdu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
; Absolute difference ops
define weak_odr <4 x i32> @absd_i32x4(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline {
%tmp = call <4 x i32> @llvm.arm.neon.vabds.v4i32(<4 x i32> %a, <4 x i32> %b)
ret <4 x i32> %tmp
}
define weak_odr <2 x i32> @absd_i32x2(<2 x i32> %a, <2 x i32> %b) nounwind alwaysinline {
%tmp = call <2 x i32> @llvm.arm.neon.vabds.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %tmp
}
define weak_odr <4 x i16> @absd_i16x4(<4 x i16> %a, <4 x i16> %b) nounwind alwaysinline {
%tmp = call <4 x i16> @llvm.arm.neon.vabds.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %tmp
}
define weak_odr <8 x i16> @absd_i16x8(<8 x i16> %a, <8 x i16> %b) nounwind alwaysinline {
%tmp = call <8 x i16> @llvm.arm.neon.vabds.v8i16(<8 x i16> %a, <8 x i16> %b)
ret <8 x i16> %tmp
}
define weak_odr <8 x i8> @absd_i8x8(<8 x i8> %a, <8 x i8> %b) nounwind alwaysinline {
%tmp = call <8 x i8> @llvm.arm.neon.vabds.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %tmp
}
define weak_odr <16 x i8> @absd_i8x16(<16 x i8> %a, <16 x i8> %b) nounwind alwaysinline {
%tmp = call <16 x i8> @llvm.arm.neon.vabds.v16i8(<16 x i8> %a, <16 x i8> %b)
ret <16 x i8> %tmp
}
define weak_odr <4 x i32> @absd_u32x4(<4 x i32> %a, <4 x i32> %b) nounwind alwaysinline {
%tmp = call <4 x i32> @llvm.arm.neon.vabdu.v4i32(<4 x i32> %a, <4 x i32> %b)
ret <4 x i32> %tmp
}
define weak_odr <2 x i32> @absd_u32x2(<2 x i32> %a, <2 x i32> %b) nounwind alwaysinline {
%tmp = call <2 x i32> @llvm.arm.neon.vabdu.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %tmp
}
define weak_odr <4 x i16> @absd_u16x4(<4 x i16> %a, <4 x i16> %b) nounwind alwaysinline {
%tmp = call <4 x i16> @llvm.arm.neon.vabdu.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %tmp
}
define weak_odr <8 x i16> @absd_u16x8(<8 x i16> %a, <8 x i16> %b) nounwind alwaysinline {
%tmp = call <8 x i16> @llvm.arm.neon.vabdu.v8i16(<8 x i16> %a, <8 x i16> %b)
ret <8 x i16> %tmp
}
define weak_odr <8 x i8> @absd_u8x8(<8 x i8> %a, <8 x i8> %b) nounwind alwaysinline {
%tmp = call <8 x i8> @llvm.arm.neon.vabdu.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %tmp
}
define weak_odr <16 x i8> @absd_u8x16(<16 x i8> %a, <16 x i8> %b) nounwind alwaysinline {
%tmp = call <16 x i8> @llvm.arm.neon.vabdu.v16i8(<16 x i8> %a, <16 x i8> %b)
ret <16 x i8> %tmp
}
; Widening absolute difference ops. llvm peephole recognizes vabdl and
; vabal as calls to vabd followed by widening. Regardless of the
; signedness of the arg, these always zero-extend, because an absolute
; difference is always positive and may overflow a signed int.
define weak_odr <8 x i16> @vabdl_i8x8(<8 x i8> %a, <8 x i8> %b) nounwind alwaysinline {
%1 = call <8 x i8> @llvm.arm.neon.vabds.v8i8(<8 x i8> %a, <8 x i8> %b)
%2 = zext <8 x i8> %1 to <8 x i16>
ret <8 x i16> %2
}
define weak_odr <8 x i16> @vabdl_u8x8(<8 x i8> %a, <8 x i8> %b) nounwind alwaysinline {
%1 = call <8 x i8> @llvm.arm.neon.vabdu.v8i8(<8 x i8> %a, <8 x i8> %b)
%2 = zext <8 x i8> %1 to <8 x i16>
ret <8 x i16> %2
}
define weak_odr <4 x i32> @vabdl_i16x4(<4 x i16> %a, <4 x i16> %b) nounwind alwaysinline {
%1 = call <4 x i16> @llvm.arm.neon.vabds.v4i16(<4 x i16> %a, <4 x i16> %b)
%2 = zext <4 x i16> %1 to <4 x i32>
ret <4 x i32> %2
}
define weak_odr <4 x i32> @vabdl_u16x4(<4 x i16> %a, <4 x i16> %b) nounwind alwaysinline {
%1 = call <4 x i16> @llvm.arm.neon.vabdu.v4i16(<4 x i16> %a, <4 x i16> %b)
%2 = zext <4 x i16> %1 to <4 x i32>
ret <4 x i32> %2
}
define weak_odr <2 x i64> @vabdl_i32x2(<2 x i32> %a, <2 x i32> %b) nounwind alwaysinline {
%1 = call <2 x i32> @llvm.arm.neon.vabds.v2i32(<2 x i32> %a, <2 x i32> %b)
%2 = zext <2 x i32> %1 to <2 x i64>
ret <2 x i64> %2
}
define weak_odr <2 x i64> @vabdl_u32x2(<2 x i32> %a, <2 x i32> %b) nounwind alwaysinline {
%1 = call <2 x i32> @llvm.arm.neon.vabdu.v2i32(<2 x i32> %a, <2 x i32> %b)
%2 = zext <2 x i32> %1 to <2 x i64>
ret <2 x i64> %2
}
declare <4 x float> @llvm.sqrt.v4f32(<4 x float>);
declare <2 x double> @llvm.sqrt.v2f64(<2 x double>);
define weak_odr <4 x float> @sqrt_f32x4(<4 x float> %x) nounwind alwaysinline {
%tmp = call <4 x float> @llvm.sqrt.v4f32(<4 x float> %x)
ret <4 x float> %tmp
}
define weak_odr <2 x double> @sqrt_f64x2(<2 x double> %x) nounwind alwaysinline {
%tmp = call <2 x double> @llvm.sqrt.v2f64(<2 x double> %x)
ret <2 x double> %tmp
}
declare <4 x float> @llvm.arm.neon.vrecpe.v4f32(<4 x float> %x) nounwind readnone;
declare <2 x float> @llvm.arm.neon.vrecpe.v2f32(<2 x float> %x) nounwind readnone;
declare <4 x float> @llvm.arm.neon.vrsqrte.v4f32(<4 x float> %x) nounwind readnone;
declare <2 x float> @llvm.arm.neon.vrsqrte.v2f32(<2 x float> %x) nounwind readnone;
declare <4 x float> @llvm.arm.neon.vrecps.v4f32(<4 x float> %x, <4 x float> %y) nounwind readnone;
declare <2 x float> @llvm.arm.neon.vrecps.v2f32(<2 x float> %x, <2 x float> %y) nounwind readnone;
declare <4 x float> @llvm.arm.neon.vrsqrts.v4f32(<4 x float> %x, <4 x float> %y) nounwind readnone;
declare <2 x float> @llvm.arm.neon.vrsqrts.v2f32(<2 x float> %x, <2 x float> %y) nounwind readnone;
define weak_odr float @fast_inverse_f32(float %x) nounwind alwaysinline {
%vec = insertelement <2 x float> undef, float %x, i32 0
%approx = tail call <2 x float> @fast_inverse_f32x2(<2 x float> %vec)
%result = extractelement <2 x float> %approx, i32 0
ret float %result
}
define weak_odr <2 x float> @fast_inverse_f32x2(<2 x float> %x) nounwind alwaysinline {
%approx = tail call <2 x float> @llvm.arm.neon.vrecpe.v2f32(<2 x float> %x)
%correction = tail call <2 x float> @llvm.arm.neon.vrecps.v2f32(<2 x float> %approx, <2 x float> %x)
%result = fmul <2 x float> %approx, %correction
ret <2 x float> %result
}
define weak_odr <4 x float> @fast_inverse_f32x4(<4 x float> %x) nounwind alwaysinline {
%approx = tail call <4 x float> @llvm.arm.neon.vrecpe.v4f32(<4 x float> %x)
%correction = tail call <4 x float> @llvm.arm.neon.vrecps.v4f32(<4 x float> %approx, <4 x float> %x)
%result = fmul <4 x float> %approx, %correction
ret <4 x float> %result
}
define weak_odr float @fast_inverse_sqrt_f32(float %x) nounwind alwaysinline {
%vec = insertelement <2 x float> undef, float %x, i32 0
%approx = tail call <2 x float> @fast_inverse_sqrt_f32x2(<2 x float> %vec)
%result = extractelement <2 x float> %approx, i32 0
ret float %result
}
define weak_odr <2 x float> @fast_inverse_sqrt_f32x2(<2 x float> %x) nounwind alwaysinline {
%approx = tail call <2 x float> @llvm.arm.neon.vrsqrte.v2f32(<2 x float> %x)
%approx2 = fmul <2 x float> %approx, %approx
%correction = tail call <2 x float> @llvm.arm.neon.vrsqrts.v2f32(<2 x float> %approx2, <2 x float> %x)
%result = fmul <2 x float> %approx, %correction
ret <2 x float> %result
}
define weak_odr <4 x float> @fast_inverse_sqrt_f32x4(<4 x float> %x) nounwind alwaysinline {
%approx = tail call <4 x float> @llvm.arm.neon.vrsqrte.v4f32(<4 x float> %x)
%approx2 = fmul <4 x float> %approx, %approx
%correction = tail call <4 x float> @llvm.arm.neon.vrsqrts.v4f32(<4 x float> %approx2, <4 x float> %x)
%result = fmul <4 x float> %approx, %correction
ret <4 x float> %result
}
define weak_odr <8 x i8> @strided_load_i8x8(i8 * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<8 x i8>, i8 *} asm sideeffect "
vld1.8 $0[0], [$1], $3
vld1.8 $0[1], [$1], $3
vld1.8 $0[2], [$1], $3
vld1.8 $0[3], [$1], $3
vld1.8 $0[4], [$1], $3
vld1.8 $0[5], [$1], $3
vld1.8 $0[6], [$1], $3
vld1.8 $0[7], [$1], $3
", "=w,=r,1,r"(i8 *%ptr, i32 %stride) nounwind
%val = extractvalue {<8 x i8>, i8 *} %tmp, 0
ret <8 x i8> %val
}
define weak_odr void @strided_store_i8x8(i8 * %ptr, i32 %stride, <8 x i8> %val) nounwind alwaysinline {
tail call i8 * asm sideeffect "
vst1.8 $3[0], [$0], $2
vst1.8 $3[1], [$0], $2
vst1.8 $3[2], [$0], $2
vst1.8 $3[3], [$0], $2
vst1.8 $3[4], [$0], $2
vst1.8 $3[5], [$0], $2
vst1.8 $3[6], [$0], $2
vst1.8 $3[7], [$0], $2
", "=r,0,r,w,~{mem}"(i8 *%ptr, i32 %stride, <8 x i8> %val) nounwind
ret void
}
define weak_odr <16 x i8> @strided_load_i8x16(i8 * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<16 x i8>, i8 *} asm sideeffect "
vld1.8 ${0:e}[0], [$1], $3
vld1.8 ${0:e}[1], [$1], $3
vld1.8 ${0:e}[2], [$1], $3
vld1.8 ${0:e}[3], [$1], $3
vld1.8 ${0:e}[4], [$1], $3
vld1.8 ${0:e}[5], [$1], $3
vld1.8 ${0:e}[6], [$1], $3
vld1.8 ${0:e}[7], [$1], $3
vld1.8 ${0:f}[0], [$1], $3
vld1.8 ${0:f}[1], [$1], $3
vld1.8 ${0:f}[2], [$1], $3
vld1.8 ${0:f}[3], [$1], $3
vld1.8 ${0:f}[4], [$1], $3
vld1.8 ${0:f}[5], [$1], $3
vld1.8 ${0:f}[6], [$1], $3
vld1.8 ${0:f}[7], [$1], $3
", "=w,=r,1,r"(i8 *%ptr, i32 %stride) nounwind
%val = extractvalue {<16 x i8>, i8 *} %tmp, 0
ret <16 x i8> %val
}
define weak_odr void @strided_store_i8x16(i8 * %ptr, i32 %stride, <16 x i8> %val) nounwind alwaysinline {
tail call i8 * asm sideeffect "
vst1.8 ${3:e}[0], [$0], $2
vst1.8 ${3:e}[1], [$0], $2
vst1.8 ${3:e}[2], [$0], $2
vst1.8 ${3:e}[3], [$0], $2
vst1.8 ${3:e}[4], [$0], $2
vst1.8 ${3:e}[5], [$0], $2
vst1.8 ${3:e}[6], [$0], $2
vst1.8 ${3:e}[7], [$0], $2
vst1.8 ${3:f}[0], [$0], $2
vst1.8 ${3:f}[1], [$0], $2
vst1.8 ${3:f}[2], [$0], $2
vst1.8 ${3:f}[3], [$0], $2
vst1.8 ${3:f}[4], [$0], $2
vst1.8 ${3:f}[5], [$0], $2
vst1.8 ${3:f}[6], [$0], $2
vst1.8 ${3:f}[7], [$0], $2
", "=r,0,r,w,~{mem}"(i8 *%ptr, i32 %stride, <16 x i8> %val) nounwind
ret void
}
define weak_odr <4 x i16> @strided_load_i16x4(i16 * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<4 x i16>, i16 *} asm sideeffect "
vld1.16 $0[0], [$1], $3
vld1.16 $0[1], [$1], $3
vld1.16 $0[2], [$1], $3
vld1.16 $0[3], [$1], $3
", "=w,=r,1,r"(i16 *%ptr, i32 %stride) nounwind
%val = extractvalue {<4 x i16>, i16 *} %tmp, 0
ret <4 x i16> %val
}
define weak_odr void @strided_store_i16x4(i16 * %ptr, i32 %stride, <4 x i16> %val) nounwind alwaysinline {
tail call i16 * asm sideeffect "
vst1.16 $3[0], [$0], $2
vst1.16 $3[1], [$0], $2
vst1.16 $3[2], [$0], $2
vst1.16 $3[3], [$0], $2
", "=r,0,r,w,~{mem}"(i16 *%ptr, i32 %stride, <4 x i16> %val) nounwind
ret void
}
define weak_odr <8 x i16> @strided_load_i16x8(i16 * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<8 x i16>, i16 *} asm sideeffect "
vld1.16 ${0:e}[0], [$1], $3
vld1.16 ${0:e}[1], [$1], $3
vld1.16 ${0:e}[2], [$1], $3
vld1.16 ${0:e}[3], [$1], $3
vld1.16 ${0:f}[0], [$1], $3
vld1.16 ${0:f}[1], [$1], $3
vld1.16 ${0:f}[2], [$1], $3
vld1.16 ${0:f}[3], [$1], $3
", "=w,=r,1,r"(i16 *%ptr, i32 %stride) nounwind
%val = extractvalue {<8 x i16>, i16 *} %tmp, 0
ret <8 x i16> %val
}
define weak_odr void @strided_store_i16x8(i16 * %ptr, i32 %stride, <8 x i16> %val) nounwind alwaysinline {
tail call i16 * asm sideeffect "
vst1.16 ${3:e}[0], [$0], $2
vst1.16 ${3:e}[1], [$0], $2
vst1.16 ${3:e}[2], [$0], $2
vst1.16 ${3:e}[3], [$0], $2
vst1.16 ${3:f}[0], [$0], $2
vst1.16 ${3:f}[1], [$0], $2
vst1.16 ${3:f}[2], [$0], $2
vst1.16 ${3:f}[3], [$0], $2
", "=r,0,r,w,~{mem}"(i16 *%ptr, i32 %stride, <8 x i16> %val) nounwind
ret void
}
define weak_odr <4 x i32> @strided_load_i32x4(i32 * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<4 x i32>, i32 *} asm sideeffect "
vld1.32 ${0:e}[0], [$1], $3
vld1.32 ${0:e}[1], [$1], $3
vld1.32 ${0:f}[0], [$1], $3
vld1.32 ${0:f}[1], [$1], $3
", "=w,=r,1,r"(i32 *%ptr, i32 %stride) nounwind
%val = extractvalue {<4 x i32>, i32 *} %tmp, 0
ret <4 x i32> %val
}
define weak_odr void @strided_store_i32x4(i32 * %ptr, i32 %stride, <4 x i32> %val) nounwind alwaysinline {
tail call i32 * asm sideeffect "
vst1.32 ${3:e}[0], [$0], $2
vst1.32 ${3:e}[1], [$0], $2
vst1.32 ${3:f}[0], [$0], $2
vst1.32 ${3:f}[1], [$0], $2
", "=r,0,r,w,~{mem}"(i32 *%ptr, i32 %stride, <4 x i32> %val) nounwind
ret void
}
define weak_odr <4 x float> @strided_load_f32x4(float * %ptr, i32 %stride) nounwind alwaysinline {
%tmp = tail call {<4 x float>, float *} asm sideeffect "
vld1.32 ${0:e}[0], [$1], $3
vld1.32 ${0:e}[1], [$1], $3
vld1.32 ${0:f}[0], [$1], $3
vld1.32 ${0:f}[1], [$1], $3
", "=w,=r,1,r"(float *%ptr, i32 %stride) nounwind
%val = extractvalue {<4 x float>, float *} %tmp, 0
ret <4 x float> %val
}
define weak_odr void @strided_store_f32x4(float * %ptr, i32 %stride, <4 x float> %val) nounwind alwaysinline {
tail call float * asm sideeffect "
vst1.32 ${3:e}[0], [$0], $2
vst1.32 ${3:e}[1], [$0], $2
vst1.32 ${3:f}[0], [$0], $2
vst1.32 ${3:f}[1], [$0], $2
", "=r,0,r,w,~{mem}"(float *%ptr, i32 %stride, <4 x float> %val) nounwind
ret void
}
; The way llvm represents intrinsics for horizontal addition are
; somewhat ad-hoc, and can be incompatible with the way we slice up
; intrinsics to meet the native vector width. We define wrappers for
; everything here instead.
declare <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float>, <2 x float>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
define weak_odr <8 x i8> @pairwise_Add_int8x8_int8x16(<16 x i8> %x) nounwind alwaysinline {
%a = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%b = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%result = tail call <8 x i8> @llvm.arm.neon.vpadd.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %result
}
define weak_odr <4 x i16> @pairwise_Add_int16x4_int16x8(<8 x i16> %x) nounwind alwaysinline {
%a = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%b = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%result = tail call <4 x i16> @llvm.arm.neon.vpadd.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %result
}
define weak_odr <2 x i32> @pairwise_Add_int32x2_int32x4(<4 x i32> %x) nounwind alwaysinline {
%a = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x i32> @llvm.arm.neon.vpadd.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %result
}
define weak_odr <2 x float> @pairwise_Add_float32x2_float32x4(<4 x float> %x) nounwind alwaysinline {
%a = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x float> @llvm.arm.neon.vpadd.v2f32(<2 x float> %a, <2 x float> %b)
ret <2 x float> %result
}
declare <1 x i64> @llvm.arm.neon.vpaddls.v1i64.v2i32(<2 x i32>) nounwind readnone
declare <1 x i64> @llvm.arm.neon.vpaddlu.v1i64.v2i32(<2 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpaddls.v2i32.v4i16(<4 x i16>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpaddlu.v2i32.v4i16(<4 x i16>) nounwind readnone
declare <2 x i64> @llvm.arm.neon.vpaddls.v2i64.v4i32(<4 x i32>) nounwind readnone
declare <2 x i64> @llvm.arm.neon.vpaddlu.v2i64.v4i32(<4 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpaddls.v4i16.v8i8(<8 x i8>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpaddlu.v4i16.v8i8(<8 x i8>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vpaddls.v4i32.v8i16(<8 x i16>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vpaddlu.v4i32.v8i16(<8 x i16>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vpaddls.v8i16.v16i8(<16 x i8>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vpaddlu.v8i16.v16i8(<16 x i8>) nounwind readnone
define weak_odr <8 x i16> @pairwise_Add_int16x8_int8x16(<16 x i8> %x) nounwind alwaysinline {
%result = tail call <8 x i16> @llvm.arm.neon.vpaddls.v8i16.v16i8(<16 x i8> %x)
ret <8 x i16> %result
}
define weak_odr <4 x i16> @pairwise_Add_int16x4_int8x8(<8 x i8> %x) nounwind alwaysinline {
%result = tail call <4 x i16> @llvm.arm.neon.vpaddls.v4i16.v8i8(<8 x i8> %x)
ret <4 x i16> %result
}
define weak_odr <4 x i32> @pairwise_Add_int32x4_int16x8(<8 x i16> %x) nounwind alwaysinline {
%result = tail call <4 x i32> @llvm.arm.neon.vpaddls.v4i32.v8i16(<8 x i16> %x)
ret <4 x i32> %result
}
define weak_odr <2 x i32> @pairwise_Add_int32x2_int16x4(<4 x i16> %x) nounwind alwaysinline {
%result = tail call <2 x i32> @llvm.arm.neon.vpaddls.v2i32.v4i16(<4 x i16> %x)
ret <2 x i32> %result
}
define weak_odr <2 x i64> @pairwise_Add_int64x2_int32x4(<4 x i32> %x) nounwind alwaysinline {
%result = tail call <2 x i64> @llvm.arm.neon.vpaddls.v2i64.v4i32(<4 x i32> %x)
ret <2 x i64> %result
}
define weak_odr i64 @pairwise_Add_int64_int32x2(<2 x i32> %x) nounwind alwaysinline {
%result = tail call <1 x i64> @llvm.arm.neon.vpaddls.v1i64.v2i32(<2 x i32> %x)
%scalar = extractelement <1 x i64> %result, i32 0
ret i64 %scalar
}
define weak_odr i64 @pairwise_Add_int64_int64x2(<2 x i64> %x) nounwind alwaysinline {
; There's no intrinsic for this on arm32, but we include an implementation for completeness.
%a = extractelement <2 x i64> %x, i32 0
%b = extractelement <2 x i64> %x, i32 1
%result = add i64 %a, %b
ret i64 %result
}
define weak_odr <8 x i16> @pairwise_Add_uint16x8_uint8x16(<16 x i8> %x) nounwind alwaysinline {
%result = tail call <8 x i16> @llvm.arm.neon.vpaddlu.v8i16.v16i8(<16 x i8> %x)
ret <8 x i16> %result
}
define weak_odr <4 x i16> @pairwise_Add_uint16x4_uint8x8(<8 x i8> %x) nounwind alwaysinline {
%result = tail call <4 x i16> @llvm.arm.neon.vpaddlu.v4i16.v8i8(<8 x i8> %x)
ret <4 x i16> %result
}
define weak_odr <4 x i32> @pairwise_Add_uint32x4_uint16x8(<8 x i16> %x) nounwind alwaysinline {
%result = tail call <4 x i32> @llvm.arm.neon.vpaddlu.v4i32.v8i16(<8 x i16> %x)
ret <4 x i32> %result
}
define weak_odr <2 x i32> @pairwise_Add_uint32x2_uint16x4(<4 x i16> %x) nounwind alwaysinline {
%result = tail call <2 x i32> @llvm.arm.neon.vpaddlu.v2i32.v4i16(<4 x i16> %x)
ret <2 x i32> %result
}
define weak_odr <2 x i64> @pairwise_Add_uint64x2_uint32x4(<4 x i32> %x) nounwind alwaysinline {
%result = tail call <2 x i64> @llvm.arm.neon.vpaddlu.v2i64.v4i32(<4 x i32> %x)
ret <2 x i64> %result
}
define weak_odr i64 @pairwise_Add_uint64_uint32x2(<2 x i32> %x) nounwind alwaysinline {
%result = tail call <1 x i64> @llvm.arm.neon.vpaddlu.v1i64.v2i32(<2 x i32> %x)
%scalar = extractelement <1 x i64> %result, i32 0
ret i64 %scalar
}
declare <4 x i16> @llvm.arm.neon.vpadals.v4i16.v8i8(<4 x i16>, <8 x i8>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpadals.v2i32.v4i16(<2 x i32>, <4 x i16>) nounwind readnone
declare <1 x i64> @llvm.arm.neon.vpadals.v1i64.v2i32(<1 x i64>, <2 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpadalu.v4i16.v8i8(<4 x i16>, <8 x i8>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpadalu.v2i32.v4i16(<2 x i32>, <4 x i16>) nounwind readnone
declare <1 x i64> @llvm.arm.neon.vpadalu.v1i64.v2i32(<1 x i64>, <2 x i32>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vpadals.v8i16.v16i8(<8 x i16>, <16 x i8>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vpadals.v4i32.v8i16(<4 x i32>, <8 x i16>) nounwind readnone
declare <2 x i64> @llvm.arm.neon.vpadals.v2i64.v4i32(<2 x i64>, <4 x i32>) nounwind readnone
declare <8 x i16> @llvm.arm.neon.vpadalu.v8i16.v16i8(<8 x i16>, <16 x i8>) nounwind readnone
declare <4 x i32> @llvm.arm.neon.vpadalu.v4i32.v8i16(<4 x i32>, <8 x i16>) nounwind readnone
declare <2 x i64> @llvm.arm.neon.vpadalu.v2i64.v4i32(<2 x i64>, <4 x i32>) nounwind readnone
define weak_odr <8 x i16> @pairwise_Add_int16x8_int8x16_accumulate(<8 x i16> %a, <16 x i8> %x) nounwind alwaysinline {
%result = tail call <8 x i16> @llvm.arm.neon.vpadals.v8i16.v16i8(<8 x i16> %a, <16 x i8> %x)
ret <8 x i16> %result
}
define weak_odr <4 x i16> @pairwise_Add_int16x4_int8x8_accumulate(<4 x i16> %a, <8 x i8> %x) nounwind alwaysinline {
%result = tail call <4 x i16> @llvm.arm.neon.vpadals.v4i16.v8i8(<4 x i16> %a, <8 x i8> %x)
ret <4 x i16> %result
}
define weak_odr <4 x i32> @pairwise_Add_int32x4_int16x8_accumulate(<4 x i32> %a, <8 x i16> %x) nounwind alwaysinline {
%result = tail call <4 x i32> @llvm.arm.neon.vpadals.v4i32.v8i16(<4 x i32> %a, <8 x i16> %x)
ret <4 x i32> %result
}
define weak_odr <2 x i32> @pairwise_Add_int32x2_int16x4_accumulate(<2 x i32> %a, <4 x i16> %x) nounwind alwaysinline {
%result = tail call <2 x i32> @llvm.arm.neon.vpadals.v2i32.v4i16(<2 x i32> %a, <4 x i16> %x)
ret <2 x i32> %result
}
define weak_odr <2 x i64> @pairwise_Add_int64x2_int32x4_accumulate(<2 x i64> %a, <4 x i32> %x) nounwind alwaysinline {
%result = tail call <2 x i64> @llvm.arm.neon.vpadals.v2i64.v4i32(<2 x i64> %a, <4 x i32> %x)
ret <2 x i64> %result
}
define weak_odr i64 @pairwise_Add_int64_int32x2_accumulate(i64 %a, <2 x i32> %x) nounwind alwaysinline {
%vec = insertelement <1 x i64> undef, i64 %a, i32 0
%result = tail call <1 x i64> @llvm.arm.neon.vpadals.v1i64.v2i32(<1 x i64> %vec, <2 x i32> %x)
%scalar = extractelement <1 x i64> %result, i32 0
ret i64 %scalar
}
define weak_odr <8 x i16> @pairwise_Add_uint16x8_uint8x16_accumulate(<8 x i16> %a, <16 x i8> %x) nounwind alwaysinline {
%result = tail call <8 x i16> @llvm.arm.neon.vpadalu.v8i16.v16i8(<8 x i16> %a, <16 x i8> %x)
ret <8 x i16> %result
}
define weak_odr <4 x i16> @pairwise_Add_uint16x4_uint8x8_accumulate(<4 x i16> %a, <8 x i8> %x) nounwind alwaysinline {
%result = tail call <4 x i16> @llvm.arm.neon.vpadalu.v4i16.v8i8(<4 x i16> %a, <8 x i8> %x)
ret <4 x i16> %result
}
define weak_odr <4 x i32> @pairwise_Add_uint32x4_uint16x8_accumulate(<4 x i32> %a, <8 x i16> %x) nounwind alwaysinline {
%result = tail call <4 x i32> @llvm.arm.neon.vpadalu.v4i32.v8i16(<4 x i32> %a, <8 x i16> %x)
ret <4 x i32> %result
}
define weak_odr <2 x i32> @pairwise_Add_uint32x2_uint16x4_accumulate(<2 x i32> %a, <4 x i16> %x) nounwind alwaysinline {
%result = tail call <2 x i32> @llvm.arm.neon.vpadalu.v2i32.v4i16(<2 x i32> %a, <4 x i16> %x)
ret <2 x i32> %result
}
define weak_odr <2 x i64> @pairwise_Add_uint64x2_uint32x4_accumulate(<2 x i64> %a, <4 x i32> %x) nounwind alwaysinline {
%result = tail call <2 x i64> @llvm.arm.neon.vpadalu.v2i64.v4i32(<2 x i64> %a, <4 x i32> %x)
ret <2 x i64> %result
}
define weak_odr i64 @pairwise_Add_uint64_uint32x2_accumulate(i64 %a, <2 x i32> %x) nounwind alwaysinline {
%vec = insertelement <1 x i64> undef, i64 %a, i32 0
%result = tail call <1 x i64> @llvm.arm.neon.vpadalu.v1i64.v2i32(<1 x i64> %vec, <2 x i32> %x)
%scalar = extractelement <1 x i64> %result, i32 0
ret i64 %scalar
}
declare <2 x float> @llvm.arm.neon.vpmaxs.v2f32(<2 x float>, <2 x float>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpmaxs.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpmaxu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpmaxs.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpmaxu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vpmaxs.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vpmaxu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
define weak_odr <8 x i8> @pairwise_Max_int8x8_int8x16(<16 x i8> %x) nounwind alwaysinline {
%a = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%b = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%result = tail call <8 x i8> @llvm.arm.neon.vpmaxs.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %result
}
define weak_odr <4 x i16> @pairwise_Max_int16x4_int16x8(<8 x i16> %x) nounwind alwaysinline {
%a = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%b = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%result = tail call <4 x i16> @llvm.arm.neon.vpmaxs.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %result
}
define weak_odr <2 x i32> @pairwise_Max_int32x2_int32x4(<4 x i32> %x) nounwind alwaysinline {
%a = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x i32> @llvm.arm.neon.vpmaxs.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %result
}
define weak_odr <2 x float> @pairwise_Max_float32x2_float32x4(<4 x float> %x) nounwind alwaysinline {
%a = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x float> @llvm.arm.neon.vpmaxs.v2f32(<2 x float> %a, <2 x float> %b)
ret <2 x float> %result
}
define weak_odr <8 x i8> @pairwise_Max_uint8x8_uint8x16(<16 x i8> %x) nounwind alwaysinline {
%a = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%b = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%result = tail call <8 x i8> @llvm.arm.neon.vpmaxu.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %result
}
define weak_odr <4 x i16> @pairwise_Max_uint16x4_uint16x8(<8 x i16> %x) nounwind alwaysinline {
%a = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%b = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%result = tail call <4 x i16> @llvm.arm.neon.vpmaxu.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %result
}
define weak_odr <2 x i32> @pairwise_Max_uint32x2_uint32x4(<4 x i32> %x) nounwind alwaysinline {
%a = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x i32> @llvm.arm.neon.vpmaxu.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %result
}
declare <2 x float> @llvm.arm.neon.vpmins.v2f32(<2 x float>, <2 x float>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpmins.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <2 x i32> @llvm.arm.neon.vpminu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpmins.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <4 x i16> @llvm.arm.neon.vpminu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vpmins.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
declare <8 x i8> @llvm.arm.neon.vpminu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
define weak_odr <8 x i8> @pairwise_Min_int8x8_int8x16(<16 x i8> %x) nounwind alwaysinline {
%a = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%b = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%result = tail call <8 x i8> @llvm.arm.neon.vpmins.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %result
}
define weak_odr <4 x i16> @pairwise_Min_int16x4_int16x8(<8 x i16> %x) nounwind alwaysinline {
%a = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%b = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%result = tail call <4 x i16> @llvm.arm.neon.vpmins.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %result
}
define weak_odr <2 x i32> @pairwise_Min_int32x2_int32x4(<4 x i32> %x) nounwind alwaysinline {
%a = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x i32> @llvm.arm.neon.vpmins.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %result
}
define weak_odr <2 x float> @pairwise_Min_float32x2_float32x4(<4 x float> %x) nounwind alwaysinline {
%a = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x float> %x, <4 x float> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x float> @llvm.arm.neon.vpmins.v2f32(<2 x float> %a, <2 x float> %b)
ret <2 x float> %result
}
define weak_odr <8 x i8> @pairwise_Min_uint8x8_uint8x16(<16 x i8> %x) nounwind alwaysinline {
%a = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%b = shufflevector <16 x i8> %x, <16 x i8> undef, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
%result = tail call <8 x i8> @llvm.arm.neon.vpminu.v8i8(<8 x i8> %a, <8 x i8> %b)
ret <8 x i8> %result
}
define weak_odr <4 x i16> @pairwise_Min_uint16x4_uint16x8(<8 x i16> %x) nounwind alwaysinline {
%a = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
%b = shufflevector <8 x i16> %x, <8 x i16> undef, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
%result = tail call <4 x i16> @llvm.arm.neon.vpminu.v4i16(<4 x i16> %a, <4 x i16> %b)
ret <4 x i16> %result
}
define weak_odr <2 x i32> @pairwise_Min_uint32x2_uint32x4(<4 x i32> %x) nounwind alwaysinline {
%a = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
%b = shufflevector <4 x i32> %x, <4 x i32> undef, <2 x i32> <i32 2, i32 3>
%result = tail call <2 x i32> @llvm.arm.neon.vpminu.v2i32(<2 x i32> %a, <2 x i32> %b)
ret <2 x i32> %result
}