[WASM SIMD] Implement horizontal add for float and integer types.

src/wasm/wasm-opcodes.h

Index: src/wasm/wasm-opcodes.h
diff --git a/src/wasm/wasm-opcodes.h b/src/wasm/wasm-opcodes.h
index f307e2bd740f95354049a0ca055acc6ef6f4843b..cf6d612cf40370191e1a5ced60c031a9e4d4e625 100644
--- a/src/wasm/wasm-opcodes.h
+++ b/src/wasm/wasm-opcodes.h
@@ -290,6 +290,7 @@ constexpr WasmCodePosition kNoCodePosition = -1;
   V(F32x4RecipApprox, 0xe506, s_s)       \
   V(F32x4RecipSqrtApprox, 0xe507, s_s)   \
   V(F32x4Add, 0xe508, s_ss)              \
+  V(F32x4AddHoriz, 0xe5b9, s_ss)         \

[georgia.kouveli, 2017/04/20 14:53:06]

Perhaps pairwise add would be a better name than horizontal add? (I would think
horizontal add refers to adding up all the elements in a single vector).

[bbudge, 2017/04/21 20:18:58]

Intel and MIPS naming is horizontal, though they are really only pairwise like
ARM.

MIPS - hadd, hsub
Intel - PHADDS  (packed horizontal add)

It's challenging to design a portable operation for this, since the MIPS
operation widens the sum (to avoid overflow) and there is no FP variant. Intel
and ARM have the normal behavior here, which is to return the result modulo 2 ^
number of bits per lane. Intel has no type for byte-sized lanes, and it seems
less useful than the others, so I'm removing that operation from this CL.

   V(F32x4Sub, 0xe509, s_ss)              \
   V(F32x4Mul, 0xe50a, s_ss)              \
   V(F32x4Div, 0xe50b, s_ss)              \
@@ -310,6 +311,7 @@ constexpr WasmCodePosition kNoCodePosition = -1;
   V(I32x4Splat, 0xe51b, s_i)             \
   V(I32x4Neg, 0xe51e, s_s)               \
   V(I32x4Add, 0xe51f, s_ss)              \
+  V(I32x4AddHoriz, 0xe5ba, s_ss)         \
   V(I32x4Sub, 0xe520, s_ss)              \
   V(I32x4Mul, 0xe521, s_ss)              \
   V(I32x4MinS, 0xe522, s_ss)             \
@@ -336,6 +338,7 @@ constexpr WasmCodePosition kNoCodePosition = -1;
   V(I16x8Neg, 0xe53b, s_s)               \
   V(I16x8Add, 0xe53c, s_ss)              \
   V(I16x8AddSaturateS, 0xe53d, s_ss)     \
+  V(I16x8AddHoriz, 0xe5bb, s_ss)         \
   V(I16x8Sub, 0xe53e, s_ss)              \
   V(I16x8SubSaturateS, 0xe53f, s_ss)     \
   V(I16x8Mul, 0xe540, s_ss)              \
@@ -365,6 +368,7 @@ constexpr WasmCodePosition kNoCodePosition = -1;
   V(I8x16Neg, 0xe55a, s_s)               \
   V(I8x16Add, 0xe55b, s_ss)              \
   V(I8x16AddSaturateS, 0xe55c, s_ss)     \
+  V(I8x16AddHoriz, 0xe5bc, s_ss)         \
   V(I8x16Sub, 0xe55d, s_ss)              \
   V(I8x16SubSaturateS, 0xe55e, s_ss)     \
   V(I8x16Mul, 0xe55f, s_ss)              \