Re: [Tinycc-devel] modern c++ compiler written in C (food for thought)

[Christian Jullien]

This my last post about OpenLisp to not pollute tcc mailing Lisp.

Yes, OpenLisp is probably the world fastest lisp interpreter. For example, it interprets lisp code faster than CLISP compiler.

 

It has 3 modes interpreter (the default), compiled to LAP with interpreted by a VM (~ 2times faster), LAP translated and optimized in C (~20times faster than interpreter).

 

Here is what you can get with standard Lisp benchmarks when compiled in C (known as gabriel's benchmarks):

 

01Fib     : ok, time =   0.000s. ( 0 GC)

02Tak     : ok, time =   0.000s. ( 0 GC)

03Stak    : ok, time =   0.002s. ( 0 GC)

04Ctak    : ok, time =   0.004s. ( 0 GC)

05Takl    : ok, time =   0.003s. ( 0 GC)

06Takr    : ok, time =   0.001s. ( 0 GC)

07Boyer   : ok, time =   0.007s. ( 2 GC)

08Browse  : ok, time =   0.015s. ( 6 GC)

09Destru  : ok, time =   0.001s. ( 1 GC)

10Travini : ok, time =   0.004s. ( 0 GC)

11Travrun : ok, time =   0.023s. ( 0 GC)

12Deriv   : ok, time =   0.003s. ( 2 GC)

13Dderiv  : ok, time =   0.003s. ( 2 GC)

14Divit   : ok, time =   0.001s. ( 1 GC)

15Divrec  : ok, time =   0.002s. ( 1 GC)

16FFT     : ok, time =   0.008s. ( 0 GC)

17Puzzle  : ok, time =   0.012s. ( 0 GC)

18Triang  : ok, time =   0.079s. ( 0 GC)

19Fprint  : ok, time =   0.004s. ( 0 GC)

20Fread   : ok, time =   0.001s. ( 0 GC)

21Tprint  : ok, time =   0.000s. ( 0 GC)

22Frpoly  : ok, time =   0.023s. (12 GC)

                       0.194 s.

 

Speed can be optimized even more if compiled using profile guided optimizations (+ ~15%)

 

I’ll be glad to reply other questions related to OpenLisp in private emails.

 

Christian

 

-----Original Message-----
From: address@hidden [mailto:address@hidden On Behalf Of Sergey Korshunoff
Sent: dimanche 1 novembre 2015 12:36
To: address@hidden
Subject: Re: [Tinycc-devel] modern c++ compiler written in C (food for thought)

 

> Microsoft C 18.0  Win32 (32 bits) |  1.865 Microsoft C 18.0  Win32

> (Compile) |  1.289

OpenLisp is so fast! Interpreted code runs as fast as LAP

 

> Legend: (I)nterpreted, (L)ap, (C)ompiled, (H)alf-compiled.

> Test    OpenLispv7.9(L)   TISLv4.0(H)

> Fib            0.008                            0.010

And LAP looks like a Half-compiled

 

_______________________________________________

Tinycc-devel mailing list

address@hidden

https://lists.nongnu.org/mailman/listinfo/tinycc-devel