Merge pull request #442 from HowJMay/pshufw

feat: Implement _m_pshufw

sse2neon.h

@@ -2186,6 +2186,11 @@ FORCE_INLINE void _mm_prefetch(const void *p, int i)
 // https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=m_psadbw
 #define _m_psadbw(a, b) _mm_sad_pu8(a, b)
 
+// Shuffle 16-bit integers in a using the control in imm8, and store the results
+// in dst.
+// https://software.intel.com/sites/landingpage/IntrinsicsGuide/#text=_m_pshufw
+#define _m_pshufw(a, imm) _mm_shuffle_pi16(a, imm)
+
 // Compute the approximate reciprocal of packed single-precision (32-bit)
 // floating-point elements in a, and store the results in dst. The maximum
 // relative error for this approximation is less than 1.5*2^-12.

tests/impl.cpp

@@ -241,6 +241,7 @@ static float ranf(float low, float high)
 // Enable the tests which are using the macro of another tests
 result_t test_mm_slli_si128(const SSE2NEONTestImpl &impl, uint32_t i);
 result_t test_mm_srli_si128(const SSE2NEONTestImpl &impl, uint32_t i);
+result_t test_mm_shuffle_pi16(const SSE2NEONTestImpl &impl, uint32_t i);
 
 // This function is not called from `runSingleTest`, but for other intrinsic
 // tests that might need to call `_mm_set_epi32`.
@@ -2317,7 +2318,7 @@ result_t test_m_psadbw(const SSE2NEONTestImpl &impl, uint32_t i)
 
 result_t test_m_pshufw(const SSE2NEONTestImpl &impl, uint32_t i)
 {
-    return TEST_UNIMPL;
+    return test_mm_shuffle_pi16(impl, i);
 }
 
 result_t test_mm_rcp_ps(const SSE2NEONTestImpl &impl, uint32_t i)