summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Updates for dynamic rate limitingnew-concurrencyDouglas Rumbaugh7 days6-24/+201
|
* Code reformattingDouglas Rumbaugh2025-09-2539-1664/+1626
|
* Insertion Stall UpdatesDouglas Rumbaugh2025-09-233-27/+80
| | | | | | | Revised insertion stalling mechanism to work for structures with > 1us required stall time per record, and got dynamic calculation of the stall time working.
* Implemented dynamic insertion rate adjustmentDouglas Rumbaugh2025-09-172-0/+43
| | | | | | | | | | | | | | The insertion rate will now be updated on the fly to block inserts based on the sum of the predicted runtime and record counts for currently active reconstructions. The scheduling system now also calls into the runtime prediction system and handles this data appropriately. All that remains is to implement the prediction models themselves--then we should be good to test.
* Beginnings of per-level cost modelingDouglas Rumbaugh2025-09-174-4/+28
| | | | | | | | | | | | The total time required for each shard construction is now measured, and hooks have been added to InternalLevel to support the creation and use of per-level cost models. The appropriate calls to these functions are now made during reconstruction, but the scheduling process doesn't yet call them, nor are the models actually implemented.
* Merge branch 'new-concurrency' of ↵Douglas Rumbaugh2025-09-175-9/+106
|\ | | | | | | github.com:dbrumbaugh/dynamic-extension-working into new-concurrency
| * Per record cost estimation progressDouglas B. Rumbaugh2025-08-144-8/+102
| |
| * file_util.h: Swapped from random_shuffle to shuffleDouglas B. Rumbaugh2025-08-121-1/+4
| |
* | Last updates for dissertationDouglas Rumbaugh2025-09-1712-31/+567
| |
* | updatesDouglas Rumbaugh2025-06-0210-137/+54
|/
* update to knn sweepDouglas Rumbaugh2025-06-021-2/+1
|
* Merge branch 'new-concurrency' of ↵Douglas Rumbaugh2025-06-024-1/+155
|\ | | | | | | github.com:dbrumbaugh/dynamic-extension-working into new-concurrency
| * Merge branch 'new-concurrency' of ↵Douglas Rumbaugh2025-05-294-1/+160
| |\ | | | | | | | | | github.com:dbrumbaugh/dynamic-extension-working into new-concurrency
| * | Stall rate benchmarkingDouglas Rumbaugh2025-05-294-1/+155
| | |
* | | KNN sweepDouglas Rumbaugh2025-06-023-1/+152
| |/ |/|
* | Added selectivity sweep benchmarkDouglas Rumbaugh2025-05-254-1/+160
|/
* Updates for tail latency stall testingDouglas Rumbaugh2025-05-257-64/+308
|
* Stats tweaks and KNN benchmarkDouglas Rumbaugh2025-04-307-9/+248
|
* Improved statistics calculationDouglas Rumbaugh2025-04-253-83/+116
|
* Scheduler statistics tracking updateDouglas Rumbaugh2025-04-256-38/+130
| | | | | | | | | The current scheme is really inefficient in terms of retreival of the results, but keeps the critical path mostly clear. It's probably worth it to do a more organized tracking of the data as it comes in, to avoid an n^2 statistics generation step at the end.
* switched k to a doubleDouglas Rumbaugh2025-04-175-16/+18
|
* DynamicExtension.h: Updated function docstringsDouglas Rumbaugh2025-04-101-23/+26
|
* DynamicExtension.h: Added the insertion rate limiterDouglas Rumbaugh2025-04-101-52/+98
| | | | | | The exact determination of *how* to use this mechanism will be contigent on some of the math I'm working on at the moment.
* Missed cmakelists in last commitDouglas Rumbaugh2025-04-091-0/+11
|
* New benchmarksDouglas Rumbaugh2025-04-092-0/+337
|
* Disabled early abort for point lookupsDouglas Rumbaugh2025-04-093-9/+36
| | | | | | | I'm having some trouble getting "bad" query performance, so I'm going to try using PLs w/o early abort as a worst-case scenario. This should get the best per-shard performance scaling
* BSM bugfixesDouglas Rumbaugh2025-04-083-16/+12
|
* Implemented the legacy policies and fixed a few bugsDouglas Rumbaugh2025-04-0710-163/+195
|
* working commit (temporary progress, doesn't build)Douglas Rumbaugh2025-04-0512-118/+148
|
* LockManager.h: adjusted unlock procedureDouglas Rumbaugh2025-03-031-1/+2
| | | | | There was a race condition. Now there isn't... or at least there's one less.
* Fixed a few concurrency bugsDouglas Rumbaugh2025-03-0321-253/+370
|
* Query Preemption: still has one or two bugs, but mostly worksDouglas Rumbaugh2025-02-248-17/+246
|
* Added uniform data generator as file optionDouglas Rumbaugh2025-02-201-215/+240
|
* Average version of mixed-workload benchDouglas Rumbaugh2025-02-173-8/+192
|
* Thread count sweepDouglas Rumbaugh2025-02-171-7/+14
|
* Fixed a few benchmark bugsDouglas Rumbaugh2025-02-172-1/+10
|
* Multi-threaded mixed workload benchmarkDouglas Rumbaugh2025-02-173-10/+157
|
* Fixed legacy tiering policyDouglas Rumbaugh2025-02-174-13/+19
|
* BTree insertion latency benchmarksDouglas Rumbaugh2025-02-172-0/+82
|
* Retooled dist benchmark to actually print dist infoDouglas Rumbaugh2025-02-141-8/+7
|
* Parallel background reconstructions appear to be working!Douglas Rumbaugh2025-02-147-98/+49
|
* progress towards resolving asynch structure mergesDouglas Rumbaugh2025-02-136-138/+131
|
* more updatesDouglas Rumbaugh2025-02-113-16/+29
|
* Progress: began adding parallel merging and locking of levelsDouglas Rumbaugh2025-02-1116-87/+180
|
* Migrated remaining tail latency benchmarks to new interfaceDouglas Rumbaugh2025-02-103-19/+19
|
* Tiering: adjusted layout policy to properly handle L0Douglas Rumbaugh2025-02-061-1/+1
|
* CMakeLists updateDouglas Rumbaugh2025-02-061-0/+1
|
* Background compaction stuffDouglas Rumbaugh2025-02-0615-98/+587
|
* Bug fixesDouglas Rumbaugh2025-02-043-3/+8
|
* PSUDB updateDouglas Rumbaugh2025-02-031-0/+0
|