#grin_compiler 搜尋結果
Just compiled the first running program with this pipeline: GHC 8.11.0 -> ExternalSTG -> Lambda -> GHC 8.8.2 #grin_compiler
Finished the first version of the Haskell Cost Centre Stack viewer. (similar to call stack) It uses GHC-WPC to show the source code of the functions. Web UI is powerful and easily accessible! github.com/grin-compiler/… #grin_compiler #Haskell
#grin_compiler In the Datalog model of a GRIN program, each program point must be uniquely identified. Right now, I'm working on extending the syntax of the GRIN IR so that it can be processed by Datalog analyses.
I finally know how to build GHC/STG programs manually! github.com/grin-compiler/… #grin_compiler
#grin_compiler Currently measuring execution times and memory usage of small Idris programs compiled with GRIN. Both seem to improve compared to the binaries generated by the Idris compiler. This is to refine the measurements in our previous paper. nbviewer.jupyter.org/github/Anabra/…
I added part of Idris-Dev tests to idris-grin. I am fixing the the bugs they found. #grin_compiler
Finally I've fixed the Control Flow Analysis Datalog implementation. I wrote a unit test for each Datalog rule. The implementation is still unoptimized but it gets the correct result. It takes 14 seconds to complete the CFA for this program: raw.githubusercontent.com/grin-compiler/… #grin_compiler
The next task I'll do is to feed back the whole program Lambda2/STG IR to GHC/STG and generate working executable with GHC's regular codegen backend and RTS. #grin_compiler
It seems that the simplest type system for GHC/STG is the following: newtype StgType = UnboxedTuple [PrimRep] data PrimRep = IntRep | WordRep | AddrRep | FloatRep | DoubleRep | LiftedRep | UnliftedRep In short everything is an unboxed tuple. #grin_compiler
I implemented something similar in the mini-grin repository. I am really glad that you made it as a reusable library... I see possible applications in the #grin_compiler. :)
For now I'm planning to use it for analyses in a compiler I'm writing, maybe later I will find other usecases. Also hoping the #grin_compiler team will use it at some point. And it was simply a lot of fun to glue it all together. 😁
I am working on a Heap Points-To implementation using Souffle datalog as part of a code experiment in github.com/andorp/mini-gr… . When done it will be the model how to integrate abstract interpretations written in datalog in the github.com/grin-compiler/… #grin_compiler
#grin_compiler I still need to fix GHC-GRIN CFA Datalog implementation. So I work on unit tests. github.com/grin-compiler/…
Finished the first version of the Haskell Cost Centre Stack viewer. (similar to call stack) It uses GHC-WPC to show the source code of the functions. Web UI is powerful and easily accessible! github.com/grin-compiler/… #grin_compiler #Haskell
Just compiled the first running program with this pipeline: GHC 8.11.0 -> ExternalSTG -> Lambda -> GHC 8.8.2 #grin_compiler
It seems that the simplest type system for GHC/STG is the following: newtype StgType = UnboxedTuple [PrimRep] data PrimRep = IntRep | WordRep | AddrRep | FloatRep | DoubleRep | LiftedRep | UnliftedRep In short everything is an unboxed tuple. #grin_compiler
I finally know how to build GHC/STG programs manually! github.com/grin-compiler/… #grin_compiler
I implemented something similar in the mini-grin repository. I am really glad that you made it as a reusable library... I see possible applications in the #grin_compiler. :)
For now I'm planning to use it for analyses in a compiler I'm writing, maybe later I will find other usecases. Also hoping the #grin_compiler team will use it at some point. And it was simply a lot of fun to glue it all together. 😁
I added part of Idris-Dev tests to idris-grin. I am fixing the the bugs they found. #grin_compiler
Finally I've fixed the Control Flow Analysis Datalog implementation. I wrote a unit test for each Datalog rule. The implementation is still unoptimized but it gets the correct result. It takes 14 seconds to complete the CFA for this program: raw.githubusercontent.com/grin-compiler/… #grin_compiler
The next task I'll do is to feed back the whole program Lambda2/STG IR to GHC/STG and generate working executable with GHC's regular codegen backend and RTS. #grin_compiler
#grin_compiler In the Datalog model of a GRIN program, each program point must be uniquely identified. Right now, I'm working on extending the syntax of the GRIN IR so that it can be processed by Datalog analyses.
#grin_compiler Currently measuring execution times and memory usage of small Idris programs compiled with GRIN. Both seem to improve compared to the binaries generated by the Idris compiler. This is to refine the measurements in our previous paper. nbviewer.jupyter.org/github/Anabra/…
#grin_compiler I still need to fix GHC-GRIN CFA Datalog implementation. So I work on unit tests. github.com/grin-compiler/…
I am working on a Heap Points-To implementation using Souffle datalog as part of a code experiment in github.com/andorp/mini-gr… . When done it will be the model how to integrate abstract interpretations written in datalog in the github.com/grin-compiler/… #grin_compiler
Just compiled the first running program with this pipeline: GHC 8.11.0 -> ExternalSTG -> Lambda -> GHC 8.8.2 #grin_compiler
Something went wrong.
Something went wrong.
United States Trends
- 1. #AEWDynamite 18.3K posts
- 2. Will Richard 3,849 posts
- 3. #Survivor49 3,035 posts
- 4. #SistasOnBET 2,196 posts
- 5. #ChicagoPD 1,023 posts
- 6. Klay 6,492 posts
- 7. #AmphoreusStamp 2,398 posts
- 8. Godzilla 29.5K posts
- 9. Unplanned 4,713 posts
- 10. Nico Harrison N/A
- 11. Binnington 2,328 posts
- 12. Harrison Barnes N/A
- 13. Kent State 1,327 posts
- 14. Athena 10.2K posts
- 15. Savannah 5,446 posts
- 16. Cade 30.6K posts
- 17. Jovic 1,029 posts
- 18. Pat Spencer N/A
- 19. Mitchell Robinson 1,366 posts
- 20. Ovechkin 7,850 posts