aboutsummaryrefslogtreecommitdiffstats
path: root/searchlib
Commit message (Collapse)AuthorAgeFilesLines
* Name the threads so it is easier to see who is doing what.Henning Baldersheim2020-11-125-14/+25
|
* Add name and type to exception.Henning Baldersheim2020-11-122-2/+3
|
* Improve protection of _parts map.Henning Baldersheim2020-11-112-6/+13
|
* just compute our resultArne Juul2020-11-101-11/+6
|
* fix copyright, untabifyArne Juul2020-11-101-2/+3
|
* add script wrapperArne Juul2020-11-102-0/+80
|
* add a new Java application for evaluating tensor conformance testsArne Juul2020-11-101-0/+160
| | | | | | * based on TensorConformanceTest unit test class * reads JSON into Slime structure * annotates with actual results from vespajlib evaluation
* Removed unused IFieldBase of Serializer/Deserializer interface.Henning Baldersheim2020-11-0928-187/+131
| | | | | Many years ago there were a big plan. But it went nowehere. So let us GC it instead of lying around polluting and bloating the code.
* Merge pull request #15221 from vespa-engine/balder/compute-updateability-onceHenning Baldersheim2020-11-082-4/+6
|\ | | | | - Move utility methods from proton::attribute to search::attribute.
| * - Move utility methods from proton::attribute to search::attribute.Henning Baldersheim2020-11-082-4/+6
| | | | | | | | - Compute isUpdateableInmemoryOnly once.
* | No need to inherit Identifiable.Henning Baldersheim2020-11-073-10/+4
|/
* Support using DirectTensorAttribute for both sparse and mixed tensors.Geir Storli2020-11-061-1/+1
|
* Merge pull request #15135 from vespa-engine/bratseth/median-aggregatorJon Bratseth2020-11-043-9/+15
|\ | | | | Add median aggregator
| * Add median aggregatorJon Bratseth2020-11-023-9/+15
| |
* | balance second phase ranking workloadHåvard Pettersen2020-11-024-63/+52
|/ | | | | | ... by first giving each thread the same number of results to re-rank and then exchanging results back to the threads finding them during matching/first phase ranking.
* GC unused visibility-delay, fsync and compression for tls flags.Henning Baldersheim2020-10-281-2/+2
|
* dedup common code.Henning Baldersheim2020-10-201-0/+7
|
* Fix forward declarations.Tor Egge2020-10-169-9/+9
|
* Remove search::StateFile and search::StateBuf.Tor Egge2020-10-1617-1364/+0
|
* Remove search::IOErrorHandler.Tor Egge2020-10-167-476/+0
|
* Remove search::SigBusHandler.Tor Egge2020-10-167-345/+0
|
* Merge pull request #14906 from vespa-engine/arnej/maybe-shrink-sparse-tensorArne H Juul2020-10-165-10/+32
|\ | | | | Arnej/maybe shrink sparse tensor
| * more robust test when tensors are memory efficientArne Juul2020-10-151-3/+28
| |
| * cosmetic cleanupsArne Juul2020-10-154-7/+4
| |
* | Merge pull request #14880 from ↵Arne H Juul2020-10-151-3/+25
|\ \ | |/ |/| | | | | vespa-engine/arnej/avoid-internal-typed-binary-format avoid using code that should be internal to TypedBinaryFormat
| * avoid using code that should be internal to TypedBinaryFormatArne Juul2020-10-141-3/+25
| |
* | fix specArne Juul2020-10-151-1/+1
| |
* | use "return {};" to return empty unique_ptrArne Juul2020-10-154-4/+4
| |
* | use qt_type againArne Juul2020-10-151-2/+2
| |
* | simplifyArne Juul2020-10-157-19/+7
| |
* | handle both engine- and factory-based tensorsArne Juul2020-10-1551-353/+291
| | | | | | | | | | | | * use EngineOrFactory::get() instead of DefaultTensorEngine::ref() * avoid direct use of DenseTensorView etc where possible * use eval::Value instead of tensor::Tensor where possible
* | GC unuse code and use std::mutex/std:condition_variable over vespalib::MonitorHenning Baldersheim2020-10-144-18/+11
|/
* vespalib::Monitor -> std:.mutex/std::condition_variableHenning Baldersheim2020-10-143-28/+30
|
* Merge pull request #14852 from vespa-engine/arnej/some-eval-additionsArne H Juul2020-10-144-8/+8
|\ | | | | Arnej/some eval additions
| * use cells() not cellsRef()Arne Juul2020-10-144-8/+8
| |
* | Use std::mutex/std::condition_variableHenning Baldersheim2020-10-1410-136/+139
|/
* - GC the last usages of vespalib::Lock.Henning Baldersheim2020-10-084-3/+2
| | | | - Now it is only vespalib::Monitor left
* Use std::mutex for the hash lockHenning Baldersheim2020-10-082-4/+4
|
* Use std::mutexHenning Baldersheim2020-10-088-61/+62
|
* Make vespalib hash functors noexceptTor Brede Vekterli2020-10-081-1/+1
| | | | | | Make ducktyped functor conditionally noexcept on `hash()` member function. Follow up on `-Werror=noexcept` warnings indicating where `noexcept` can be trivially added for such functions.
* Merge pull request #14774 from vespa-engine/balder/std-mutex-2.6Henning Baldersheim2020-10-088-62/+59
|\ | | | | Use vespalib::Lock -> std::mutex
| * Use vespalib::Lock -> std::mutexHenning Baldersheim2020-10-088-62/+59
| |
* | Use vespalib::Lock -> std::mutexHenning Baldersheim2020-10-082-23/+13
|/
* - GC unused code.Henning Baldersheim2020-10-075-13/+12
| | | | - vespalib::Lock -> std::mutex
* Merge pull request #14763 from ↵Henning Baldersheim2020-10-076-1/+7
|\ | | | | | | | | vespa-engine/balder/remove-dangerous-broke-move-and-copy-constructor-operators-on-sync Remove broken copy and move constructor and assignment operatoos on v…
| * Remove broken copy and move constructor and assignment operatoos on ↵Henning Baldersheim2020-10-076-1/+7
| | | | | | | | | | | | vespalib::Lock and vespalib::Monitor. Also repair broken usages of the same.
* | Range iterators over std::vector<bool> and storage::JudyMultiMap do notTor Egge2020-10-071-1/+1
|/ | | | return references.
* Add noexcept as indicated by -WnoeexceptHenning Baldersheim2020-10-0749-171/+149
|
* Merge pull request #14737 from ↵Henning Baldersheim2020-10-061-0/+2
|\ | | | | | | | | vespa-engine/toregge/include-hpp-files-to-get-definitions-of-needed-member-functions Include hpp files to get definitions of needed member functions.
| * Include hpp files to get definitions of needed member functions.Tor Egge2020-10-061-0/+2
| |