This version has a working gpu acceleration for inner product with the use of the -gpu flag, gpu not activated for other operations
More work done on the gpu compiler, inner product not yet working
Fixed library to work with more than one gpu present
first working version of clvecmat tested with cpu version of opencl under xorin on an intel cpu
modified conf153 to reflect delphi semantics for mod
fixed bugs in truncation of floats in AVX512 and rationalised the sharing of instructions between sse and AVX512 using ifdefs in the latter
remove tests which would never work and change them into ones that do work with turbo/delphi semantics
update the antlr version
more example programs for the book
Generalise nabla so it takes and expression not an identifier as argument
Added nabla ∇ as an operator to pascal
Fix persistentce linking for arm64
Add a collection of tutorial examples
Fixed bug in erase file
Update manual to describe persistence
First working version with persistent store
Integratation of the persistent heap with rtl.c not fully working yet
First version of persistent store for vector pascal
Add memory mapped file routine
Fix up assignment of fp16 on intel, not tested yet due to not having a granit rapids cpu
Updated the apassing of static arrays to formal schematic ones
Fixed a bug in the libray arcsin function
fix bug in arm64 round to integer