guile-user
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Playing with guile (vs python). Generate file for GDP suitable for g


From: Andy Wingo
Subject: Re: Playing with guile (vs python). Generate file for GDP suitable for gnuplot.
Date: Mon, 06 Mar 2017 22:02:25 +0100
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.1 (gnu/linux)

On Mon 06 Mar 2017 14:50, Thien-Thi Nguyen <address@hidden> writes:

> FWIW, we can also use SRFI 13 to save a few cycles:
>
>  (use-modules (ice-9 rdelim) (srfi srfi-13))
>  
>  (define (file->lines filename)
>    "Return a list of lines contained in a file."
>    (call-with-input-file
>        filename
>      (lambda (p)
>        (let loop ((acc '()))
>          (let ((line (read-line p)))
>            (if (eof-object? line)
>                (reverse! acc)            ; rv
>                (loop (cons (string-drop-right line 1)
>                            acc))))))))
>
> This variant is also more stack-conserving (amenable to TCO),
> which should afford some additional speedup.

Interestingly the timings are not dissimilar in 2.2; consider this most
microbenchmarky of cases:

    (use-modules (ice-9 match))
    (define test (iota #e1e6))
    (define (copy-list/accum xs)
      (let lp ((xs xs) (out '()))
        (match xs
          ((x . xs) (lp xs (cons x out)))
          (() (reverse! out)))))
    (define (copy-list/recur xs)
      (let lp ((xs xs))
        (match xs
          ((x . xs) (cons x (lp xs)))
          (() '()))))

Doing a ,time (length (copy-list test) at the REPL for both these two
copy-list impls gives me:

                 real times (s)
    ------------------------------------
    copy-list/recur   |  copy-list/accum
    ....................................
    0.067212s           0.046110s
    0.077659s           0.059750s
    0.065923s           0.040371s
    0.066098s           0.062963s
    0.062718s           0.040980s

The recursive version is still slower ATM but I think that will go away
once I manage to avoid running the function prelude each time; not
sure.  Anyway, I am happy recommending recursing now :)

Andy



reply via email to

[Prev in Thread] Current Thread [Next in Thread]