Sears Russell
f375176537
Performance section edits.
2004-10-22 06:07:44 +00:00
Eric Brewer
39520efd16
added pdf
2004-10-22 05:58:34 +00:00
Sears Russell
75b8e7e62c
Another manual merge.
2004-10-22 05:44:40 +00:00
Sears Russell
e9f41b8671
New intro / abstract Eric.
2004-10-22 05:07:22 +00:00
Sears Russell
1bd2fbf2dd
minor revision.
2004-10-22 05:03:16 +00:00
Sears Russell
bca199968b
*** empty log message ***
2004-10-22 04:57:25 +00:00
Sears Russell
3d2a36c417
Fixed typo
2004-10-20 23:10:37 +00:00
Sears Russell
939badfda0
New benchmark. :)
2004-10-20 23:02:16 +00:00
Sears Russell
aead6a5a83
New benchmark, fixed for PROFILE_LATCHES.
2004-10-20 21:55:00 +00:00
Sears Russell
e05286c9c5
set this back to old value.
2004-10-20 21:36:39 +00:00
Sears Russell
b59ba35ed4
parameter tweak..
2004-10-20 06:54:39 +00:00
Sears Russell
70b8aca134
Set some constants to values appropriate for my desktop.
2004-10-20 06:48:29 +00:00
Sears Russell
281781dfd0
Forgot this file...
2004-10-20 06:29:55 +00:00
Sears Russell
6a744690f8
Added special 'EVERYTHING.script' which runs the tests from fastest to slowest. (To be run overnight...)
2004-10-20 06:28:31 +00:00
Sears Russell
af9e7adee0
Starting work on the actual graph-generation portion of the benchmarks. :)
2004-10-20 06:21:21 +00:00
Sears Russell
903ba4b97b
Bugfix.
2004-10-19 22:58:56 +00:00
Sears Russell
ec99c0e258
Lots of optimizations; re-did on-disk layout for the linear hash table.
2004-10-19 21:16:37 +00:00
Sears Russell
6b265f28b0
Speedups, most notably in the logging subsystem.
2004-10-19 04:45:42 +00:00
Sears Russell
c850ce3a51
The logger now calls fwrite half as often as it used to. This could be improved significantly, but some latching details internal to the logger need to be worked out first.
2004-10-18 19:19:10 +00:00
Sears Russell
ec1276f26f
Preliminary version of logical logging linear hash. (No latching yet, and there are some bugs re-opening a closed hash.)
2004-10-18 18:24:54 +00:00
Sears Russell
04cfc7da8e
Optimized the 'naive' linear hash table, and moved to sourceforge. :)
2004-10-17 02:17:00 +00:00
Sears Russell
a078f25475
Implemented a linear hash from scratch. Concurrency hasn't been done, but it's probably correct in the single thread case.
2004-10-12 02:44:47 +00:00
Sears Russell
5064e3fac2
Implemented pages that store fixed sized records efficiently, and a java-style ArrayList data structure that efficiently supports (relatively) clustered, O(1) access time expandable arrays. (This will be used for the hash implementation...)
2004-10-06 06:08:09 +00:00
Sears Russell
9268b1d9cf
Added support for gcov, and ran gprof. Shaved 10% off the runtime of check_page by optimizing rw.c
2004-10-04 05:24:12 +00:00
Sears Russell
96e7af92a4
Added test cases for Tprepare(), implemented some redo-only operations, and started to clean up record allocation/deallocation.
...
Also, numerous bugfixes.
2004-10-02 07:29:34 +00:00
Sears Russell
811bc5c710
Implemented a freespace manager that should safely allocate space, even in the face of crashes, and can reclaim unused space (unless an application opens more than one simultaneous transaction that performs allocations)
...
Fixed some blob bugs (by adding extra fdatasync() calls).
Began factoring out the page management code so that it is an extenstion, and a less integral part of lladd.
2004-08-21 00:03:30 +00:00
Sears Russell
b4d7883f66
Fixed regressions, major update of page / slotted interfaces in anticipation of moving LLADD towards a generic transactional page system.
2004-08-17 01:46:17 +00:00
Sears Russell
82e3fdb53a
Added (untested) support for whole-page operations, lladdhash now works.
2004-08-03 02:04:56 +00:00
Sears Russell
20e78dd0db
Test cases for indirect blocks.
2004-07-31 00:28:45 +00:00
Sears Russell
e51759b517
Implemented indirect pages, and fixed some bugs here and there. (Still need to add transactional support for indirect pages...)
2004-07-31 00:27:55 +00:00
Sears Russell
a74c499dd7
Fixed data corruption bug in the freelist implementation, and corrected lsn updates.
2004-07-30 02:07:41 +00:00
Sears Russell
961b63af15
Rewrote / refactored page.h. Now, it uses a freelist to reclaim slotid's within a page, and is implemented as two seperate modules. The first handles lsn's, and (will soon) check the type of each page, and then delegate work to a particular implementation of the second module.
2004-07-30 01:28:39 +00:00
Sears Russell
c649ba7504
O_DIRECT is working now.
2004-07-27 22:04:59 +00:00
Sears Russell
3cbe112926
Moved some utility functions into their own file.
2004-07-27 21:33:00 +00:00
Sears Russell
78eb2cbf6a
O_DIRECT (sort of...)
2004-07-27 21:30:54 +00:00
Sears Russell
54ba9b0347
Refactored bufferMananger / pageCache, so that bufferMananger handles loading an unloading pages, and pageCache only implements the replacement policy.
...
Also, used memprof to detect and remove all memory leaks (at least all memory leaks that are encountered by check_transactional2), and fixed a number of misuses of the pblHash.
2004-07-27 01:04:35 +00:00
Sears Russell
e53117a1a0
Implemented (but didn't test) iterator over the records in a page.
2004-07-26 22:01:09 +00:00
Sears Russell
476cb7f1ca
Re-enabled non-blob tests.
2004-07-26 20:37:59 +00:00
Sears Russell
8cb59f0c23
Blobs are now re-entrant.
2004-07-26 20:37:04 +00:00
Sears Russell
422198418a
bufferMananger is (really!) re-entrant now, performed big refactoring to prevent deadlocks, check_transactional2 ran overnight without coring, w/ 25 threads, and 10000 allocate/write/reads per thread. (The whole test suite was run in a loop...)
2004-07-23 20:21:44 +00:00
Sears Russell
38ae73c146
Just committing some documentation updates before big refactoring.
2004-07-22 17:48:36 +00:00
Sears Russell
126ae31392
bufferMananger is *really* reentrant now! (I think) However, the same thread may not load a page more than once. So, LLADD has to be altered so that it never calls loadPage() more than once per operation (more efficient), and/or pin/unpin need to be implemented. An easy way to do pin/unpin is to have them pull things out of the cache replacement list, and then re-insert them.
2004-07-21 02:13:28 +00:00
Sears Russell
0ce77903fb
Continuing work on multi-threading. r/w access to buffer manager getting close, but still buggy.
2004-07-20 03:40:57 +00:00
Sears Russell
490dd86c09
pageCache.c is now re-entrant.
2004-07-20 00:15:17 +00:00
Sears Russell
12fc5665ab
pageCache is now re-entrant, in theory.
2004-07-15 00:42:36 +00:00
Sears Russell
9712e291e6
Moved page.h and some of the logging headers out of the public API.
2004-07-14 21:25:59 +00:00
Sears Russell
2b42451280
Cleaning up bufferManager / page for locking. Want to limit access to the Page struct.
2004-07-14 20:49:18 +00:00
Sears Russell
475b8141d3
page.c is re-entrant (mostly), and now reuses DeRalloced space properly. (For now, BufferManager still is not re-entrant, and also prevents space from being reused.)
2004-07-13 23:48:20 +00:00
Sears Russell
282d284035
Made pageWriteLSN static.
2004-07-09 22:05:33 +00:00
Sears Russell
d564b9ded6
More documentation fixes.
2004-07-06 21:41:33 +00:00
