summaryrefslogtreecommitdiffstats
path: root/eval/src
Commit message (Collapse)AuthorAgeFilesLines
* avoid abstract value typesHåvard Pettersen2019-04-2528-756/+247
| | | | | | | | | | | | remove basic value type ANY remove concept of tensors with unknown dimensions disallow value types with unbound indexed dimensions remove predicates talking about abstract types type of unknown values are now ERROR (was ANY) require that overlapping indexed dimensions are of equal size type unification now requires types to be equal ('if' expressions) creating a tensor type without dimensions now gives a double type make rank feature setup fail on invalid types (query/attribute)
* support binary formats with cell type in reference implementationHåvard Pettersen2019-04-093-7/+248
|
* Merge pull request #9029 from ↵Henning Baldersheim2019-04-082-2/+42
|\ | | | | | | | | vespa-engine/havardpe/parse-and-ignore-tensor-cell-type parse (and ignore for now) tensor cell type in tensor type spec
| * Improve handling of malformed cell type specifiersHåvard Pettersen2019-04-081-6/+24
| | | | | | | | | | | | | | | | | | Do not consume any input when encountering a malformed cell type specifier. This avoids parse conflicts between 'less' operators and tensor cell type specifiers in type expressions in the type resolution test which parses partial input into value types. example expression: 'tensor < double'
| * parse (and ignore for now) tensor cell type in tensor type specHåvard Pettersen2019-04-052-2/+24
| | | | | | | | example: tensor<float>(x[10])
* | Use bit number 3 mean 'with cell type' to carry forward the current schema ↵Henning Baldersheim2019-04-052-5/+7
|/ | | | | | where the bits hav meaning. Update expected result.
* Add a method that will extract cells only to a favoured cell type. This is ↵Henning Baldersheim2019-04-036-11/+72
| | | | temporary until full typed tensor support is in place.
* Keep the serialzation of the cells with the serialization for now.Henning Baldersheim2019-04-037-82/+79
| | | | Clean up code for better understanding and reuse.
* Instead of specifying type of tensor, specify serialization.Henning Baldersheim2019-04-0313-86/+74
|
* Remove wrappers not neeed.Henning Baldersheim2019-04-023-12/+1
|
* Remove duplicate typedefs.Henning Baldersheim2019-04-022-12/+8
|
* Handle dense tensors encoded as floatHenning Baldersheim2019-04-0211-137/+251
|
* Add cell_typeHenning Baldersheim2019-04-024-19/+18
|
* Only include what you need.Henning Baldersheim2019-04-024-11/+4
|
* Unify on nested namespaces.Henning Baldersheim2019-04-0221-107/+54
|
* Adjust build setup for Darwin.Tor Egge2019-03-141-1/+1
|
* Check for assignable tensor type when setting tensor in TensorFieldValue.Tor Egge2019-03-071-3/+3
|
* Add '()' to macro definition.Geir Storli2019-03-013-3/+3
|
* Merge pull request #8604 from ↵Tor Egge2019-02-255-22/+156
|\ | | | | | | | | vespa-engine/geirst/remove-and-modify-for-mixed-tensors Geirst/remove and modify for mixed tensors
| * Verify that previous value is used by testing with an operation that ↵Geir Storli2019-02-251-17/+8
| | | | | | | | combines old and new.
| * Support modify operation on mixed tensors.Geir Storli2019-02-252-2/+64
| |
| * Support remove operation on mixed tensors.Geir Storli2019-02-254-5/+86
| |
* | Merge pull request #8596 from vespa-engine/toregge/eliminate-some-gcc-9-warningsHenning Baldersheim2019-02-252-2/+2
|\ \ | |/ |/| Eliminate some gcc 9 warnings.
| * Eliminate some gcc 9 warnings.Tor Egge2019-02-252-2/+2
| |
* | Support add operation on mixed tensors.Geir Storli2019-02-252-2/+74
| |
* | Migrate unit tests to gtest.Geir Storli2019-02-256-41/+48
|/
* Implement remove operation for sparse tensor.Geir Storli2019-02-2012-1/+154
|
* Remove files used by old unit test runner.Tor Egge2019-02-1912-12/+0
|
* Eliminate clang warnings in eval.Tor Egge2019-02-102-3/+3
|
* Return unique pointer to const tensor instead.Geir Storli2019-02-081-2/+2
|
* Create common test function for creating a tensor using default tensor engine.Geir Storli2019-02-084-38/+50
|
* Merge pull request #8439 from ↵Tor Egge2019-02-0827-35/+224
|\ | | | | | | | | vespa-engine/geirst/add-operation-support-on-sparse-tensors Geirst/add operation support on sparse tensors
| * Update Tensor forward declarations.Geir Storli2019-02-088-8/+8
| |
| * Rename unit test tensor_modify -> tensor_modify_operation.Geir Storli2019-02-083-8/+8
| |
| * Implement add() function on vespalib::tensor::Tensor for adding a set of ↵Geir Storli2019-02-0816-19/+208
| | | | | | | | | | | | cells to the tensor. Currently only supported for sparse tensor.
* | Eliminate some clang warnings in evalTor Egge2019-02-083-5/+2
|/
* Merge pull request #8348 from vespa-engine/toregge/tensor-update-end-to-endGeir Storli2019-02-0419-48/+395
|\ | | | | Tensor modify update end to end
| * Test that updates to missing cells are ignored.Tor Egge2019-02-011-0/+24
| | | | | | | | Test that updates to out of range cells are ignored.
| * Apply tensor modify update to given field value.Tor Egge2019-02-0119-48/+371
| |
* | Revert "Revert "Use OS provided xxhash.""Henning Baldersheim2019-02-033-2/+16
| |
* | Revert "Use OS provided xxhash."Håkon Hallingstad2019-02-033-16/+2
| |
* | Use OS provided xxhash.Henning Baldersheim2019-02-013-2/+16
| |
* | Adjust forward declarations in eval.Tor Egge2019-02-0119-25/+25
|/
* Merge pull request #6518 from ↵Henning Baldersheim2018-08-081-4/+15
|\ | | | | | | | | vespa-engine/havardpe/avoid-old-style-varargs-when-specifying-function-types Avoid using the 'convenient' way to specify parameter types when
| * Avoid using the 'convenient' way to specify parameter types whenHåvard Pettersen2018-08-081-4/+15
| | | | | | | | | | | | | | | | | | | | resolving application functions we want to call from an llvm-compiled function. LLVM 3.9 uses varargs with a nullptr sentinel at the end while newer versions of LLVM (5/6) uses variadic templates. This leads to 'inconvenience' when upgrading where the code still compiles, but the no longer needed/wanted nullptr sentinel triggers a signature mismatch leading to not being able to call functions like 'sin'/'cos' etc.
* | dump -> print(stream)Henning Baldersheim2018-08-083-13/+11
|/
* newer llvm releases have dropped virtual.Henning Baldersheim2018-08-071-7/+4
| | | | Use llvm::dyn_cast to be future safe.
* Remove stringref::c_str()Tor Brede Vekterli2018-07-252-2/+2
| | | | | | | | | | | | | The expected semantics of c_str() (a null-terminated string) cannot be satisfied with a string reference, so remove the function entirely to prevent people from using it in buggy ways. Replaces c_str() with data() in places where it is presumed safe, otherwise constructs temporary string instances. Certain callsites have been de-stringref'd in favor of regular strings, in particular where C APIs have been transitively called. The vast majority of these were called with string parameters anyway, so should not cause much extra allocation.
* some more cleanupArne Juul2018-06-201-3/+0
|
* add and use HDR_ABORT insteadArne Juul2018-06-202-5/+5
|