Speed up FullStepRow index comparison by leveraging big-endian byte layout
This commit is contained in:
@@ -80,7 +80,7 @@ public:
|
||||
FullStepRow(const FullStepRow<W>& a, const FullStepRow<W>& b, size_t len, size_t lenIndices, int trim);
|
||||
FullStepRow& operator=(const FullStepRow<WIDTH>& a);
|
||||
|
||||
inline bool IndicesBefore(const FullStepRow<WIDTH>& a, size_t len) const { return ArrayToEhIndex(hash+len) < ArrayToEhIndex(a.hash+len); }
|
||||
inline bool IndicesBefore(const FullStepRow<WIDTH>& a, size_t len, size_t lenIndices) const { return memcmp(hash+len, a.hash+len, lenIndices) < 0; }
|
||||
std::vector<eh_index> GetIndices(size_t len, size_t lenIndices) const;
|
||||
|
||||
template<size_t W>
|
||||
|
||||
Reference in New Issue
Block a user