4 ms·
It won't be extremely fast. This was discussed in comp.lang.forth a few years back. It's very interesting as an instructional tool but is not optimal for GFort
by lebuffon 5y ago
It won't be extremely fast.
This was discussed in comp.lang.forth a few years back. It's very interesting as an instructional tool but is not optimal for GForth.
For example the symbol table is written in Forth whereas it would be faster to use GForth wordlists which give you a hashed lookup method.
There is a string-to-number routine written but GForth has >NUMBER which can handle double precision conversions and is in the kernel.
However it is all there and it works. For an experience Forth user it wouldn't take too much to improve it.
If it was compiled on a native code Forth compiler like SwiftForth, iForth or VFX and with a few better uses of the internal system resources it would be fast enough to useful I suspect.