This repository has been archived by the owner. It is now read-only.
Please sign in to comment.
Simplify KnownNat constraints in for Numeric type operations
Currently, this slows down the performance of the operations because wrap-around points are computed inside of the operation, and are not specialised due to the NOINLINE pragmas. In a planned patch, these NOINLINE pragma's will be removed when the clash-prelude is compiled for fast compilation, and the performace lost by the current patch will be recovered.
- Loading branch information
Showing with 309 additions and 265 deletions.
- +14 −10 src/CLaSH/Class/BitPack.hs
- +3 −3 src/CLaSH/Prelude/BitIndex.hs
- +2 −2 src/CLaSH/Prelude/BlockRam.hs
- +6 −6 src/CLaSH/Prelude/BlockRam/File.hs
- +4 −5 src/CLaSH/Prelude/DataFlow.hs
- +2 −2 src/CLaSH/Prelude/RAM.hs
- +4 −4 src/CLaSH/Prelude/ROM.hs
- +9 −9 src/CLaSH/Prelude/ROM/File.hs
- +79 −77 src/CLaSH/Sized/Internal/BitVector.hs
- +2 −1 src/CLaSH/Sized/Internal/Index.hs-boot
- +126 −89 src/CLaSH/Sized/Internal/Signed.hs
- +58 −57 src/CLaSH/Sized/Internal/Unsigned.hs
Oops, something went wrong.