Details
-
Task
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.2.1-1, 10.2.1-2, 10.2.2-1
Description
patch as per https://github.com/MariaDB/server/pull/170 adds optimized Power8 implemention for CRC32-IEEE to complement existing CRC32C.
While I haven't included an implementation of an optimized CRC32-(IEEE) for intel, there is the following links that would enable this:
- http://www.intel.com/content/dam/www/public/us/en/documents/white-papers/fast-crc-computation-generic-polynomials-pclmulqdq-paper.pdf
- http://www.intel.com/content/dam/www/public/us/en/documents/white-papers/fast-crc-computation-paper.pdf
- http://web.archive.org/web/20140101100538/http://mail.madler.net/pipermail/zlib-devel_madler.net/2013-November/003096.html
- https://golang.org/src/hash/crc32/
Note the documentation mentions patents on CRC32 folding (which I haven't done).
Attachments
Issue Links
- relates to
-
-
- Closed
-
Activity
| Field | Original Value | New Value |
|---|---|---|
| Fix Version/s | 10.1 [ 16100 ] | |
| Fix Version/s | 10.1.14 [ 21804 ] |
| Labels | contribution patch | contribution foundation patch |
| Assignee | Sergey Vojtovich [ svoj ] |
| Sprint | 10.1.14 [ 51 ] |
| Rank | Ranked lower |
| Fix Version/s | 10.2 [ 14601 ] | |
| Fix Version/s | 10.1 [ 16100 ] |
| Sprint | 10.1.14 [ 51 ] | 10.2.1-1 [ 56 ] |
| Rank | Ranked higher |
| Sprint | 10.2.1-1 [ 56 ] | 10.2.1-1, 10.2.1-2 [ 56, 63 ] |
| Sprint | 10.2.1-1, 10.2.1-2 [ 56, 63 ] | 10.2.1-1, 10.2.1-2, 10.2.2-1 [ 56, 63, 77 ] |
| Rank | Ranked lower |
| Assignee | Sergey Vojtovich [ svoj ] | Sergei Golubchik [ serg ] |
| Status | Open [ 1 ] | In Review [ 10002 ] |
svoj thanks, sorry I didn't get back to this sooner. If you need a hand let me know.
danblack, no problem. It'd be nice if you could review this patch and let me know if I missed something.
| Assignee | Sergei Golubchik [ serg ] | Sergey Vojtovich [ svoj ] |
| Status | In Review [ 10002 ] | Stalled [ 10000 ] |
| Assignee | Sergey Vojtovich [ svoj ] | Sergei Golubchik [ serg ] |
| Status | Stalled [ 10000 ] | In Review [ 10002 ] |
| Assignee | Sergei Golubchik [ serg ] | Sergey Vojtovich [ svoj ] |
| Status | In Review [ 10002 ] | Stalled [ 10000 ] |
| Fix Version/s | 10.2.2 [ 22013 ] | |
| Fix Version/s | 10.2 [ 14601 ] | |
| Resolution | Fixed [ 1 ] | |
| Status | Stalled [ 10000 ] | Closed [ 6 ] |
danblack, we're now getting failures like this:
http://buildbot.askmonty.org/buildbot/builders/p8-xenial-bintar/builds/88/steps/test/logs/stdio
I noticed that crc32-vpmsum sources differ in your PR. Did you also see these failures? Could you elaborate the difference?
The crux of it is innodb uses CRC32C and the rest of MySQL (my_checksum) uses CRC32-IEEE.
I'll patch in both implementations into the crc32-vpmsum library and then get my_checksum to point to the vpmsum crc32-ieee implementation
https://github.com/MariaDB/server/pull/211 - test case for SQL CRC32 function
https://github.com/MariaDB/server/pull/210 - Add both versions of the CRC32 and adjust my_checksum to use the optimized version.
| Link |
This issue relates to |
| Workflow | MariaDB v3 [ 74762 ] | MariaDB v4 [ 132833 ] |
serg, please review patch for this task (b6d7d5878eecc7f0cb76f8f99a9989686ff9715c).