performance measurements

Each table row shows performance measurements for this Racket program with a particular command-line input value N.

 N  CPU secs Elapsed secs Memory KB Code B ≈ CPU Load
5000.460.4620,816532  0% 2% 98% 0%
3,00011.3911.3928,032532  1% 0% 100% 0%
5,50037.8237.8230,068532  1% 0% 0% 100%

Read the ↓ make, command line, and program output logs to see how this program was run.

Read spectral-norm benchmark to see what this program should do.


Welcome to Racket v6.1.1.

 spectral-norm Racket #2 program source code

#lang racket/base

;;; The Computer Language Benchmarks Game

;;; Translated directly from the C# version by Isaac Gouy
;;; contributed by Matthew Flatt

;; This version uses unsafe operations

(require racket/cmdline
	 racket/require (for-syntax racket/base)
           (lambda (name) (regexp-replace #rx"unsafe-" name ""))
          [fx->fl ->fl])
         (only-in racket/flonum make-flvector))

(define (Approximate n)
  (let ([u (make-flvector n 1.0)]
        [v (make-flvector n 0.0)])
    ;; 20 steps of the power method
    (for ([i (in-range 10)])
      (MultiplyAtAv n u v)
      (MultiplyAtAv n v u))
    ;; B=AtA         A multiplied by A transposed
    ;; v.Bv /(v.v)   eigenvalue of v
    (let loop ([i 0][vBv 0.0][vv 0.0])
      (if (= i n)
          (flsqrt (fl/ vBv vv))
          (let ([vi (flvector-ref v i)])
            (loop (add1 i)
                  (fl+ vBv (fl* (flvector-ref u i) vi))
                  (fl+ vv (fl* vi vi))))))))

;; return element i,j of infinite matrix A
(define (A i j)
  (fl/ 1.0 (fl+ (fl* (->fl (+ i j))
                     (fl/ (->fl (+ i (+ j 1))) 2.0)) 
                (->fl (+ i 1)))))

;; multiply vector v by matrix A
(define (MultiplyAv n v Av)
  (for ([i (in-range n)])
    (flvector-set! Av i 
                   (for/fold ([r 0.0])
                       ([j (in-range n)])
                     (fl+ r (fl* (A i j) (flvector-ref v j)))))))

;; multiply vector v by matrix A transposed
(define (MultiplyAtv n v Atv)
  (for ([i (in-range n)])
    (flvector-set! Atv i
                   (for/fold ([r 0.0])
                       ([j (in-range n)])
                     (fl+ r (fl* (A j i) (flvector-ref v j)))))))

;; multiply vector v by matrix A and then by matrix A transposed 
(define (MultiplyAtAv n v AtAv)
  (let ([u (make-flvector n 0.0)])
    (MultiplyAv n v u)
    (MultiplyAtv n u AtAv)))

(printf "~a\n"
         (Approximate (command-line #:args (n) (string->number n)))

 make, command-line, and program output logs

Sat, 04 Apr 2015 04:22:32 GMT

/usr/local/src/racket-6.1.1/bin/racket  spectralnorm.racket-2.racket 5500


Revised BSD license

  Home   Conclusions   License   Play