Add NEON and ASIMD intrinsics for SPECK-128 (GH #538)

Performance increased by about 115% on a 980 MHz BananaPi dev-board. Throughput went from about 46.2 cpb to about 21.5 cpb.

1 files changed, 6 insertions, 2 deletions

diff --git a/speck.h b/speck.h
index 037dab92..2f0c5e08 100644
--- a/speck.h
+++ b/speck.h
@@ -16,6 +16,10 @@
 #include "seckey.h"

 #include "secblock.h"

 

+#if CRYPTOPP_BOOL_X64 || CRYPTOPP_BOOL_X32 || CRYPTOPP_BOOL_X86 || CRYPTOPP_BOOL_ARM32 || CRYPTOPP_BOOL_ARM64

+# define CRYPTOPP_SPECK_ADVANCED_PROCESS_BLOCKS 1

+#endif

+

 NAMESPACE_BEGIN(CryptoPP)

 

 //! \class SPECK_Info

@@ -142,7 +146,7 @@ public:
     {

     protected:

         void ProcessAndXorBlock(const byte *inBlock, const byte *xorBlock, byte *outBlock) const;

-#if CRYPTOPP_BOOL_X86 || CRYPTOPP_BOOL_X32 || CRYPTOPP_BOOL_X64

+#if CRYPTOPP_SPECK_ADVANCED_PROCESS_BLOCKS

         size_t AdvancedProcessBlocks(const byte *inBlocks, const byte *xorBlocks, byte *outBlocks, size_t length, word32 flags) const;

 #endif

     };

@@ -155,7 +159,7 @@ public:
     {

     protected:

         void ProcessAndXorBlock(const byte *inBlock, const byte *xorBlock, byte *outBlock) const;

-#if CRYPTOPP_BOOL_X86 || CRYPTOPP_BOOL_X32 || CRYPTOPP_BOOL_X64

+#if CRYPTOPP_SPECK_ADVANCED_PROCESS_BLOCKS

         size_t AdvancedProcessBlocks(const byte *inBlocks, const byte *xorBlocks, byte *outBlocks, size_t length, word32 flags) const;

 #endif

     };