effectfully
effectfully
- [ ] **Generate terms of type other than Unit too in NEAT tests** Currently we only generate terms of type Unit in NEAT tests, which restricts the shapes of...
- [ ] **Automatically make the default builtins available to NEAT tests** We have this in the NEAT tests: ``` defaultFunTypes = Map.fromList [(TyFunG (TyBuiltinG TyIntegerG) (TyFunG (TyBuiltinG TyIntegerG) (TyBuiltinG...
- [ ] **NEAT type normalization is inefficient** The NEAT machinery uses small-step semantics for normalizing types. That is inefficient as the normalizer has to traverse the whole type over...
- [ ] **Develop a strategy for using Neat generator to test equivalence between Agda model and Plutus evaluator** As a Plutus developer I want to understand the testing strategy...
I doubt we're going to address these within the next several years or ever at all, so I'm closing the issue.
Closing in favor of #5967.
> That actually makes me wonder... could the particular list fold benchmarks be fast enough that they're being dominated by the cost of the additions or the recursion? That would...
[Latest](https://github.com/IntersectMBO/plutus/pull/5486#issuecomment-2323726065) numbers are: ``` sum/compiled-from-Haskell/sum-right-builtin/5000 5.467 ms sum/compiled-from-Haskell/sum-right-Scott/5000 5.410 ms sum/compiled-from-Haskell/sum-left-builtin/5000 5.126 ms sum/compiled-from-Haskell/sum-left-Scott/5000 5.077 ms sum/hand-written-PLC/sum-right-builtin/5000 4.209 ms sum/hand-written-PLC/sum-right-Scott/5000 4.042 ms sum/hand-written-PLC/sum-left-builtin/5000 3.452 ms sum/hand-written-PLC/sum-left-Scott/5000 3.563 ms ``` i.e....
> i.e. sops and builtins perform about the same. An approximate costing model for `matchList` copied from the one for `chooseList` gives us [budgets](https://github.com/IntersectMBO/plutus/pull/6530#discussion_r1783637694) that are also on par with...
Latest data: ``` plutus-benchmark/lists/test/Sum/9.6/left-fold-built-in.eval.golden CPU: 122_258_594 Memory: 523_832 Size: 78 (con integer 5050) ``` vs ``` plutus-benchmark/lists/test/Sum/9.6/left-fold-scott.eval.golden CPU: 69_848_900 Memory: 373_600 Size: 263 (con integer 5050) ``` I.e. builtins are...