Petalisp VS one-more-re-nightmare

While not "as fast as C" (C is not the absolute pinnacle of performance), Common Lisp is incredibly fast compared to the majority of programming languages around today. There is even a huge amount of ongoing work being done to make it faster still. We are seeing many interesting projects that make better use of the hardware in your computer (e.g. https://github.com/marcoheisig/Petalisp).

i think lisp-stat library is actually being developed. however one numerical cl library that doesnt get enough mention and is being constantly developed is petalisp for HPC

https://github.com/marcoheisig/Petalisp

However, if you have a lisp library that puts those semantics to use, then you could get it to employ magicl/ext-blas and cl-bmas to speed it up. (petalisp looks relevant, but I lack the background to compare it with APL.)

> his means cl pagckages can be "done".

this is true if there is nothing functional that can be added to a package. however its very much not true for ml frameworks right now. new things are being added all the time in the field. however even in the package i linked you have the necessary ingredients for any deep learning model: cuda and back propagation. the other person mentioned convolution which i think is pretty trivial to implement but still, if you expect everything for you to be ready made then you should probably stick to tf and pytorch. if you want to explore the cutting edge and push the boundaries then i think common lisp is a good tool. as an aside it might also be interesting to note that a common lisp package (Petalisp) is being used for high performance computing by a german university

https://github.com/marcoheisig/Petalisp

One trick is that Common Lisp can generate and compile code at runtime, whereas static languages typically do not have a compiler available at runtime. This lets you make your own lazy person's JIT/staged compiler, which is useful if some part of the problem is not known at compile-time. Such an approach has been used at least for array munging, type munging and regular expression munging.

https://github.com/telekons/one-more-re-nightmare

And the pretty hard to find blog post about it:

Depends on your regex engine, and your non-regex solution. My engine (shameless self-plug https://github.com/telekons/one-more-re-nightmare) rivals hand-written automata, having to load each character more-or-less* only once, and throws in vectorisation for simple search loops too. I would not want to write or maintain the generated code.

one-more-re-nightmare used to let you write Σ, but I then tried to search Greek stuff with it and it went wrong. So now there's...$ for all characters (since that's not used for end-of-line assertions).

One trick is that Common Lisp can generate and compile code at runtime, whereas static languages typically do not have a compiler available at runtime. This lets you make your own lazy person's JIT/staged compiler, which is useful if some part of the problem is not known at compile-time. Such an approach has been used at least for array munging, type munging and regular expression munging.

> One of my favorite examples is the regex library cl-ppcre. Thanks to the nature of Lisp, the recognizer for each regex you create can be compiled to native code on compiler implementations of CL.

That is not true - cl-ppcre generates a chain of closures. Experimental performance is in the same ballpark as typical "bytecode" interpreting regex implementations.

(Disclosure: I wrote another regex library at <https://github.com/telekons/one-more-re-nightmare>, which does do native code compilation.)

Someone else didn't tell me that before, so it can't be true. But I don't publish papers on toys, nor do I think toy projects are awfully fast. Though the x86-64 backend I wrote was in someone else's repository and thus was several PRs :(

Regular expressions

Also try this https://github.com/telekons/one-more-re-nightmare

A Common Lisp system has the compiler around at runtime, so if you can figure out how to profitably stage/specialise a computation, then you can roll your own cheap JIT of sorts. This can be useful for array munging and regular expressions at the least. You can do this in C, of course but you would need to use another compiler as a library (e.g. LLVM, TCC, libgccjit) or write your own (e.g. PCRE2's sljit).

Agree to disagree - I don't have the energy to remember operator precedence. One file from the regular expression compiler has most of the rewrite rules I read from the papers, except in S-expression syntax. There were a few bugs due to misreading precedence. Also c.f. Gerald Sussman talking about physics notation being a pain in the butt.