Use packed 32-bit MMX operations for updating the PSQT accumulator

This improves the speed of NNUE by a bit on old hardware that code path
is intended for, like a Pentium III 1.13 GHz:

10 repeats of "./stockfish bench 16 1 13 default depth NNUE":

Before:
54 642 504 897 cycles (± 0.12%)
62 301 937 829 instructions (± 0.03%)

After:
54 320 821 928 cycles (± 0.13%)
62 084 742 699 instructions (± 0.02%)

Speed of go depth 20 from startpos:

Before: 53103 nps
After: 53856 nps

closes https://github.com/official-stockfish/Stockfish/pull/3476

No functional change.

diff --git a/src/nnue/nnue_feature_transformer.h b/src/nnue/nnue_feature_transformer.h
index 2c0a0c6d3134b61270a9a16a3a1ae199176d9d07..bfa2e25a3e041ea1ef1704f1a8b5d8b48d3e0290 100644 (file)
--- a/src/nnue/nnue_feature_transformer.h
+++ b/src/nnue/nnue_feature_transformer.h
@@ -84,18 +84,18 @@ namespace Stockfish::Eval::NNUE {
 
   #elif USE_MMX
   typedef __m64 vec_t;
-  typedef std::int32_t psqt_vec_t;
+  typedef __m64 psqt_vec_t;
   #define vec_load(a) (*(a))
   #define vec_store(a,b) *(a)=(b)
   #define vec_add_16(a,b) _mm_add_pi16(a,b)
   #define vec_sub_16(a,b) _mm_sub_pi16(a,b)
   #define vec_load_psqt(a) (*(a))
   #define vec_store_psqt(a,b) *(a)=(b)
-  #define vec_add_psqt_32(a,b) a+b
-  #define vec_sub_psqt_32(a,b) a-b
-  #define vec_zero_psqt() 0
+  #define vec_add_psqt_32(a,b) _mm_add_pi32(a,b)
+  #define vec_sub_psqt_32(a,b) _mm_sub_pi32(a,b)
+  #define vec_zero_psqt() _mm_setzero_si64()
   static constexpr IndexType NumRegs = 8;
-  static constexpr IndexType NumPsqtRegs = 8;
+  static constexpr IndexType NumPsqtRegs = 4;
 
   #elif USE_NEON
   typedef int16x8_t vec_t;