summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAge
...
* Run test primals with Compile (not all succeed yet)Tom Smeding2025-03-03
|
* Compile: sum1inner, unit, idx0Tom Smeding2025-03-03
|
* Compile: Support EShape and EBuildTom Smeding2025-03-02
|
* Compile.Exec: Show number of loaded kernels in debug printsTom Smeding2025-03-02
|
* Compile: index into arraysTom Smeding2025-03-02
|
* compile: Compile constant array literalsTom Smeding2025-03-02
|
* compile: WIP reference-counted arraysTom Smeding2025-03-02
|
* Compile: More precise monads in genStructsTom Smeding2025-03-01
|
* simplify: Halfway beta-reduce product projections in effectful contextTom Smeding2025-03-01
|
* Fix some pretty-printingTom Smeding2025-03-01
|
* test: Simplify and make it a bit fasterTom Smeding2025-02-28
|
* Compile: some cleanupTom Smeding2025-02-28
|
* Remove stupid old nonworking cuda stubTom Smeding2025-02-28
|
* Compile to C and load using dlopenTom Smeding2025-02-28
|
* Comment unhandled cases instead of getting 1e6 warningsTom Smeding2025-02-25
|
* UnMonoid: Properly recurseTom Smeding2025-02-25
|
* Pretty: Allow colouring exts (currently not)Tom Smeding2025-02-25
|
* Compile: Emit structs in proper orderTom Smeding2025-02-25
|
* Purge Example.FormatTom Smeding2025-01-28
|
* Pretty-printer that supports extension fieldsTom Smeding2025-01-28
|
* Complete identity analysisTom Smeding2025-01-27
|
* Add ext field to remaining AST constructorsTom Smeding2025-01-27
|
* WIP identity analysisTom Smeding2025-01-20
|
* Improve Compile a little (still only scalars)Tom Smeding2024-12-12
|
* Somewhat working CompileTom Smeding2024-12-12
|
* Start WIP compile-to-CTom Smeding2024-12-06
|
* Rename Compile -> CompileCuTom Smeding2024-12-06
|
* UnMonoidTom Smeding2024-12-06
|
* Simplify: Reduce plus of literalsTom Smeding2024-12-06
|
* Simplify: Some forgotten 'acted'Tom Smeding2024-12-06
|
* WIP UnMonoid (to be used for compiling to C)Tom Smeding2024-12-06
|
* Fix indentationTom Smeding2024-12-06
|
* Working argument accum mode (...)Tom Smeding2024-11-26
| | | | | | | | | | | | | | | The derivative of 'neural' in full accum mode is pretty atrocious now; I think this is because when you have code like this: \(a :: Arr 1 R) -> let b = a in let c = b in sum d then because the argument, as well as both let bindings, bind a value of array type, each will introduce an accumulator, hence resulting in three (!) nested `with` clauses that each just contribute their result back to their parent. This is pointless, and we should fix this.
* WIP accum top-level argsTom Smeding2024-11-26
|
* Prepare for introducing top-level args in accum modTom Smeding2024-11-23
|
* Use accum storage for Case tooTom Smeding2024-11-23
|
* drevScoped returns a data type, not CPSTom Smeding2024-11-23
|
* Factor let storage-dependent scoping logic into separate functionTom Smeding2024-11-22
|
* Configuration for CHADTom Smeding2024-11-14
|
* Benchmark GMMTom Smeding2024-11-14
|
* Remove warningsTom Smeding2024-11-13
|
* Make EBuild derivative aware of zero cotangent arraysTom Smeding2024-11-11
|
* Test gmmTom Smeding2024-11-10
|
* Fix forward derivative of 'exp'Tom Smeding2024-11-10
|
* Test GMM; it failsTom Smeding2024-11-10
|
* Fix bug in simplifier (effects are hard)Tom Smeding2024-11-10
|
* Complete GMM implementationTom Smeding2024-11-10
|
* Cleanup, more Language operationsTom Smeding2024-11-09
|
* Add Custom to LanguageTom Smeding2024-11-09
|
* Some more primitive operatorsTom Smeding2024-11-09
|