Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | minor cleanup | Håvard Pettersen | 2018-02-01 | 1 | -11/+3 | |
| | ||||||
* | pre-alloc optimized xw product output vector | Håvard Pettersen | 2018-02-01 | 2 | -18/+27 | |
| | ||||||
* | remove tensor function eval tests | Håvard Pettersen | 2018-01-31 | 1 | -14/+0 | |
| | | | | covered by normal interpreted function tests | |||||
* | use interpreted function to evaluate tensor function | Håvard Pettersen | 2018-01-31 | 1 | -2/+6 | |
| | ||||||
* | use interpreted function to evaluate tensor function | Håvard Pettersen | 2018-01-31 | 1 | -2/+3 | |
| | ||||||
* | remove tensor function eval from cross-language conformance test | Håvard Pettersen | 2018-01-31 | 1 | -21/+0 | |
| | ||||||
* | use interpreted function to evaluate tensor function | Håvard Pettersen | 2018-01-31 | 1 | -2/+6 | |
| | ||||||
* | added test constructor to make interpreted function from tensor function | Håvard Pettersen | 2018-01-31 | 2 | -0/+13 | |
| | ||||||
* | remove retained eval from conformance tests | Håvard Pettersen | 2018-01-31 | 1 | -85/+0 | |
| | | | | (as it is now equivalent to interpreted evaluation) | |||||
* | remove recursive eval from optimized tensor function operations | Håvard Pettersen | 2018-01-31 | 4 | -27/+0 | |
| | ||||||
* | remove recursive eval from tensor function | Håvard Pettersen | 2018-01-31 | 2 | -77/+0 | |
| | ||||||
* | allow any intermediate result as input for inner products | Håvard Pettersen | 2018-01-31 | 1 | -13/+11 | |
| | ||||||
* | optimized XW product now has children | Håvard Pettersen | 2018-01-30 | 5 | -75/+99 | |
| | | | | also remove instruction trampoline | |||||
* | optimized dot product now has children | Håvard Pettersen | 2018-01-30 | 6 | -26/+35 | |
| | | | | also remove instruction trampoline | |||||
* | special case for parameter fetching | Håvard Pettersen | 2018-01-30 | 3 | -25/+27 | |
| | | | | state replace -> pop_push and pop_pop_push | |||||
* | use optimized+compiled tensor function in interpreted function | Håvard Pettersen | 2018-01-26 | 17 | -450/+81 | |
| | | | | rename 'compile' to 'optimize' in tensor engine | |||||
* | compile tensor function | Håvard Pettersen | 2018-01-26 | 7 | -1/+266 | |
| | ||||||
* | Merge pull request #4763 from vespa-engine/bratseth/support-binary-tensor-files | Jon Bratseth | 2018-01-24 | 3 | -2/+13 | |
|\ | | | | | Support binary tensor files with ending .tbf | |||||
| * | Remove unnecessary includes | Jon Bratseth | 2018-01-24 | 1 | -2/+0 | |
| | | ||||||
| * | Spaces only | Jon Bratseth | 2018-01-24 | 1 | -8/+8 | |
| | | ||||||
| * | Use memorymapped file | Jon Bratseth | 2018-01-24 | 1 | -23/+19 | |
| | | ||||||
| * | Support binary tensor files with ending .tbf | Jon Bratseth | 2018-01-24 | 3 | -17/+34 | |
| | | ||||||
* | | improve type resolving for 'if' | Håvard Pettersen | 2018-01-19 | 8 | -18/+89 | |
|/ | | | | also verify inferred types against actual result in conformance test | |||||
* | run cross-language tensor conformance tests using tensor functions | Håvard Pettersen | 2018-01-18 | 13 | -54/+59 | |
| | | | | also pass tensor engine to tensor function eval function | |||||
* | run all tests from test_spec as tensor functions | Håvard Pettersen | 2018-01-18 | 1 | -16/+32 | |
| | ||||||
* | added initial version of make_tensor_function | Håvard Pettersen | 2018-01-18 | 4 | -0/+325 | |
| | ||||||
* | test new tensor IR nodes | Håvard Pettersen | 2018-01-17 | 1 | -11/+126 | |
| | ||||||
* | added 'const_value' and 'if_node' to tensor IR | Håvard Pettersen | 2018-01-16 | 2 | -12/+83 | |
| | ||||||
* | add 'concat' and 'rename' tensor operations | Håvard Pettersen | 2018-01-16 | 2 | -41/+137 | |
| | | | | also add intermediate Leaf/Op1/Op2 classes | |||||
* | tweak tensor IR semantics and add more encapsulation | Håvard Pettersen | 2018-01-16 | 8 | -140/+170 | |
| | ||||||
* | move more param stuff to lazy_params.h | Håvard Pettersen | 2018-01-16 | 9 | -36/+41 | |
| | ||||||
* | Use lazy parameters in tensor functions. | Håvard Pettersen | 2018-01-16 | 17 | -79/+67 | |
| | ||||||
* | move LazyParams to separate file | Håvard Pettersen | 2018-01-16 | 6 | -12/+44 | |
| | ||||||
* | added test for push_children (tensor IR nodes) | Håvard Pettersen | 2018-01-15 | 1 | -0/+24 | |
| | ||||||
* | tensor function compilation based on bottom-up in-place optimization | Håvard Pettersen | 2018-01-15 | 4 | -12/+75 | |
| | ||||||
* | separate optimize and compile concepts | Håvard Pettersen | 2018-01-15 | 10 | -77/+85 | |
| | | | | | | | | optimize: tensor function -> tensor function compile: node tree -> tensor function node trees now contain tensor functions to support direct recursive mixed-mode evaluation. | |||||
* | Merge pull request #4612 from vespa-engine/balder/some-misc-old-cleanup | Henning Baldersheim | 2018-01-11 | 2 | -17/+13 | |
|\ | | | | | Use c++11 for loops and alignment. | |||||
| * | Avoid using temporaries. | Henning Baldersheim | 2018-01-10 | 2 | -17/+13 | |
| | | ||||||
* | | Merge pull request #4595 from vespa-engine/balder/track-common-cell-index | Henning Baldersheim | 2018-01-11 | 3 | -66/+93 | |
|\ \ | |/ |/| | Balder/track common cell index | |||||
| * | Put the address back in the combiner. | Henning Baldersheim | 2018-01-10 | 3 | -25/+24 | |
| | | ||||||
| * | Add comment explaining a bit about the right address. | Henning Baldersheim | 2018-01-10 | 1 | -0/+2 | |
| | | ||||||
| * | Move the right address context on the outside. | Henning Baldersheim | 2018-01-10 | 3 | -75/+68 | |
| | | ||||||
| * | Use the same AddressContext for the combinedAddress and compute the ↵ | Henning Baldersheim | 2018-01-10 | 3 | -30/+45 | |
| | | | | | | | | combinedIdx inline too. | |||||
| * | Hide the calculation inside the address context. | Henning Baldersheim | 2018-01-09 | 1 | -2/+3 | |
| | | ||||||
| * | Make and address context to keep code together. | Henning Baldersheim | 2018-01-09 | 2 | -30/+47 | |
| | | ||||||
| * | Rename to indicate it refers to right side. | Henning Baldersheim | 2018-01-09 | 2 | -10/+10 | |
| | | ||||||
* | | Revert earlier reorder. | Henning Baldersheim | 2018-01-10 | 1 | -2/+2 | |
| | | ||||||
* | | Use cheap and_modulator instead of prime modulator to avoid costly division. | Henning Baldersheim | 2018-01-10 | 2 | -2/+4 | |
| | | ||||||
* | | Different hashing gives different order. | Henning Baldersheim | 2018-01-10 | 1 | -2/+2 | |
| | | ||||||
* | | Add missing newline at eof and align members. | Henning Baldersheim | 2018-01-10 | 1 | -2/+3 | |
| | |