Ticket UUID: | 1530262 | |||
Title: | Minor optimisation to BinaryProc | |||
Type: | Patch | Version: | None | |
Submitter: | afredd | Created on: | 2006-07-28 09:54:33 | |
Subsystem: | 10. Objects | Assigned To: | dgp | |
Priority: | 5 Medium | Severity: | ||
Status: | Closed | Last Modified: | 2006-08-10 01:13:07 | |
Resolution: | Accepted | Closed By: | dgp | |
Closed on: | 2006-08-09 18:13:07 | |||
Description: |
Replacing the for loop in BinaryProc (which is used to copy characters from one buffer to another) with a memcpy provides a minor speed up. Justication: BinaryProc is often used on channelBuffer's ie. srcLen is often 4096, and memcpy is much quicker than a for loop in such cases (i measure it at about 10x). It seems like a memcpy() is faster that a for loop whenever more than ~16 characters are copied, which is probably the most common case in calls to BinaryProc. | |||
User Comments: |
dgp added on 2006-08-10 01:13:07:
Logged In: YES user_id=80530 looks fine to me; tests ok. memcpy() already used throughout Tcl sources. committed to both active branches. afredd added on 2006-07-28 16:54:34: File Added - 186554: tclEncoding.c-patch |
Attachments:
- tclEncoding.c-patch [download] added by afredd on 2006-07-28 16:54:34. [details]