-
integer conversion rules just like in C
-
shape (gets vector of dimensions)
-
math operators from cmath.h
-
reshaping of vectors
-
vertical and horizontal concat. should be able to concat along any dimension: x # y = vec i in 0:(n+m) -> if i < n then x[i] else y[i-n] -- "horizontal" x & y = vec i in 0:_ -> vec j in 0:(n+m) -> if i < n then x[i,j] else y[i,j-n] -- "vertical"
-
multidimensional transpose? x^T = vec i_1 in _ -> vec i_2 in _ -> ... vec i_d in _ -> x[i_d,...,i_1]
but... x^T for a 1-d vector remains a 1-d vector, so outer product x^T * y does not work.
but... maybe vec(x) * y is the correct way to write this (since then first vector becomes row vector)
-
REPL (and associated intepreter)
-
fix struct representation. right now structs are always pointers, so an array of structs becomes a double pointer. this is incorrect.
-
external/internal function names (extern "name" foo ...)
-
Implement:
- dgelsy - Least squares using QR
- dgeqp3 - QR fact.
- dgetrf - LU fact.
- dgetrs - Linear system solver using LU
- dorgqr - Q from QR?
- dpotrf - Cholesky fact. of real symmetric
- dpotri - inverse of real symm. pos. def.
- dtrtri - inverse of triangular
- lsame - char comparison
-
const global variables (and make better integration with C macro constants)
This repository has been archived by the owner on Aug 5, 2024. It is now read-only.