aboutsummaryrefslogtreecommitdiffstats
path: root/eval/src/tests/tensor
Commit message (Collapse)AuthorAgeFilesLines
...
* remove conformance test file (run tests directly instead)Håvard Pettersen2020-12-022-7/+26
|
* clean up tensor conformance testingHåvard Pettersen2020-12-011-19/+0
|
* test dense optimizers with FastValue instead of DefaultTensorEngineArne Juul2020-11-2712-54/+45
|
* Merge branch 'master' into arnej/add-simple-streamed-valueArne H Juul2020-11-256-476/+0
|\
| * GC unused codeArne Juul2020-11-256-476/+0
| | | | | | | | | | * remove DefaultValueBuilderFactory * remove PackedMixedTensorBuilderFactory and all packed mixed code.
* | run conformance test for StreamedValue alsoArne Juul2020-11-231-0/+6
|/
* Merge pull request #15410 from ↵Arne H Juul2020-11-211-4/+15
|\ | | | | | | | | vespa-engine/geirst/mixed-partial-update-remove-test Add partial update remove test with mixed tensor as input.
| * Add partial update remove test with mixed tensor as input.Geir Storli2020-11-201-4/+15
| |
* | Merge pull request #15412 from ↵Arne H Juul2020-11-211-29/+80
|\ \ | | | | | | | | | | | | vespa-engine/havardpe/improved-benchmarking-fairness Havardpe/improved benchmarking fairness
| * | use same loop_cnt when benchmarking if possibleHåvard Pettersen2020-11-201-6/+47
| | |
| * | each EvalOp gets its own stash, for more fairnessHåvard Pettersen2020-11-201-23/+33
| |/
* / track CellType moveArne Juul2020-11-203-5/+6
|/
* combine dimensions and split reduce operationsHåvard Pettersen2020-11-192-28/+109
|
* Add support for partial update remove operation where address is not fully ↵Geir Storli2020-11-171-0/+24
| | | | specified.
* move "keep as-is" optimizersArne Juul2020-11-1213-880/+6
| | | | | * from eval/tensor/dense to eval/instruction * minimal changes to track move in this commit
* move unit test for DenseXWProductFunctionArne Juul2020-11-112-164/+0
| | | | * and verify with FastValueBuilderFactory also
* move dense_xw_product_function to eval/instructionArne Juul2020-11-111-1/+1
|
* benchmark some forms of join with numberArne Juul2020-11-101-0/+18
|
* untangle factory-based optimization pipeline from DefaultTensorEngineHåvard Pettersen2020-11-031-22/+23
|
* drop BM of PackedMixedTensorBuilderFactoryArne Juul2020-10-261-3/+0
|
* use a run-time flag instead of conditional compilationArne Juul2020-10-264-27/+15
|
* partial duplicate of micro-bemchmarkArne Juul2020-10-254-7/+27
| | | | | | | * use conditional compilation to let the installed vespa-tensor-instructions-benchmark run its measurements only for the two most important implementations (old prod and new prod).
* Merge pull request #15012 from ↵Arne H Juul2020-10-221-20/+36
|\ | | | | | | | | vespa-engine/havardpe/better-dense-plan-for-generic-reduce improve generic dense reduce with more robust cell ordering
| * improve generic dense reduce with more robust cell orderingHåvard Pettersen2020-10-221-20/+36
| | | | | | | | | | | | | | - unroll reduce all loop - extend benchmark with optimized instructions (to catch slow ones) - drop SimpleTensorEngine from benchmark (always very slow) - tweak API on templated aggregators
* | rename to vespa-tensor-instructions-benchmark and install itArne Juul2020-10-222-2/+4
|/
* be more exact in memory accountingArne Juul2020-10-211-2/+1
| | | | | | | | | | | | | * override the delete with size and forward to global delete with actually-allocated size, thereby making it possible for AddressSanitizer to check our accounting * unit test explicitly made a float-cell builder for a double-value, that didn't work out and should not be allowed * put alignment adjustment in a common place * fix off-by-one leftover over-allocation * memory usage stats are no longer estimates, rename correspondingly
* drop SimpleValueEngineArne Juul2020-10-191-6/+0
|
* run conformance test also for SimpleValueEngineArne Juul2020-10-191-2/+8
|
* added mixed -> partial mixed peek casesHåvard Pettersen2020-10-161-0/+2
|
* added tensor peek benchmarkHåvard Pettersen2020-10-161-1/+94
|
* added tensor lambda benchmarkHåvard Pettersen2020-10-161-1/+58
|
* added encode/decode benchmarkHåvard Pettersen2020-10-161-0/+71
|
* added tensor create benchmarkHåvard Pettersen2020-10-161-41/+86
|
* extend map benchmark with number caseHåvard Pettersen2020-10-161-1/+6
|
* test via compatibility layerArne Juul2020-10-153-25/+5
|
* remove DenseTensorView::cellsRef(), just use cells()Arne Juul2020-10-141-1/+1
|
* benchmark GenericMap alsoArne Juul2020-10-131-0/+45
|
* allow interpreted function to use new generic operationsHåvard Pettersen2020-10-123-62/+36
| | | | update conformance test to also test with SimpleValue and FastValue
* Eliminate noexcept warnings.Tor Egge2020-10-111-6/+6
|
* Merge pull request #14757 from vespa-engine/add-tensor-partial-update-2Arne H Juul2020-10-096-0/+415
|\ | | | | Add tensor partial update 2
| * cleanup unit tests after reviewArne Juul2020-10-093-19/+7
| |
| * require sparse remove_specArne Juul2020-10-081-0/+1
| |
| * test larger mixed, cleanupArne Juul2020-10-081-3/+1
| |
| * add PartialRemove unit test for bad inputArne Juul2020-10-071-2/+26
| |
| * add PartialAdd unit test for bad inputArne Juul2020-10-071-3/+28
| |
| * add PartialModify unit test for bad inputArne Juul2020-10-071-2/+28
| |
| * add TensorPartialUpdate with add/modify/removeArne Juul2020-10-076-0/+353
| |
* | Merge pull request #14793 from vespa-engine/havardpe/fork-fast-and-simple-valuesArne H Juul2020-10-091-0/+1
|\ \ | | | | | | Havardpe/fork fast and simple values
| * | separate code for simple and fast valuesHåvard Pettersen2020-10-081-0/+1
| | |
* | | Merge pull request #14769 from vespa-engine/arnej/fix-concat-collapsingArne H Juul2020-10-081-0/+71
|\ \ \ | |/ / |/| | Arnej/fix concat collapsing