A perfect power, and then?

Niels Möller nisse at lysator.liu.se
Sun Oct 28 11:05:15 CET 2012


nisse at lysator.liu.se (Niels Möller) writes:

> 0. Support in speed, for benchmarking.

Not checked in yet, but here are some benchmark numbers, comparing to
binvert:

$ ./speed -s 1-50 -r mpn_binvert mpn_broot.3 mpn_broot.5 mpn_broot.0xbbbbbbbb
overhead 0.000000008 secs, precision 10000 units of 1.25e-09 secs, CPU freq 800.00 MHz
          mpn_binvert   mpn_broot.3   mpn_broot.5 mpn_broot.0xbbbbbbbb
1        #0.000000137        3.8366        4.4029       21.8508
2        #0.000000198        5.1909        5.8947       32.5246
3        #0.000000241        6.7340        7.7864       46.7029
4        #0.000000297        6.8619        8.1066       52.9341
5        #0.000000351        8.0714        9.6218       65.0285
6        #0.000000416        7.9729        9.4530       65.6215
7        #0.000000491        7.7748        9.4784       67.4275
8        #0.000000570        7.9890        9.5140       69.0510
9        #0.000000665        8.6231       10.4859       75.3280
10       #0.000000757        8.5470       10.2203       75.0891
11       #0.000000866        8.3232       10.1256       75.5589
12       #0.000000970        8.3588        9.9857       75.2369
13       #0.000001094        8.1834        9.9811       77.2853
14       #0.000001215        8.1890       10.3640       77.3365
15       #0.000001341        8.1409       10.2097       78.4372
16       #0.000001471        8.1653       10.0144       78.1555
17       #0.000001616        8.7680       10.6265       82.2390
18       #0.000001755        8.6346       10.5235       82.5256
19       #0.000001918        8.5545       10.5469       82.5374
20       #0.000002086        8.3482       10.2555       82.1251
21       #0.000002269        8.3587       10.2612       83.0050
22       #0.000002437        8.3138       10.2000       82.8108
23       #0.000002624        8.3805       10.2959       82.9302
24       #0.000002834        8.4132       10.1681       82.8163
25       #0.000002984        8.5935       10.4927       89.1252
26       #0.000003197        8.4106       10.3507       87.5543
27       #0.000003422        8.4921       10.3770       86.1490
28       #0.000003651        8.2085       10.0955       83.4299
29       #0.000003863        8.3407       10.2859       85.9442
30       #0.000004099        8.1107       10.0274       83.3120
31       #0.000004380        8.1464        9.9863       84.3408
32       #0.000004613        8.0016        9.8276       82.2488
33       #0.000004830        8.3851       10.3905       88.6530
34       #0.000005124        8.0764        9.9676       84.8014
35       #0.000005375        8.2228       10.2147       85.2712
36       #0.000005682        8.1296       10.0968       85.1667
37       #0.000005946        8.3624       10.2367       84.8217
38       #0.000006200        8.0952        9.8462       82.9712
39       #0.000006561        8.2121       10.0474       83.2341
40       #0.000006827        8.1598       10.0154       81.7764
41       #0.000007147        8.3116       10.2083       84.0677
42       #0.000007412        7.9661        9.9489       82.4589
43       #0.000007772        8.1481       10.0188       83.1875
44       #0.000008153        7.8436        9.6670       80.9926
45       #0.000008471        8.3276       10.1567       84.5383
46       #0.000008921        8.1530        9.8974       82.0601
47       #0.000009171        8.4420       10.2143       85.7628
48       #0.000009446        8.3430       10.1862       88.0095
49       #0.000009823        8.5707       10.5054       90.2245
50       #0.000010344        8.1809       10.0330       83.6643

Regards,
/Niels

-- 
Niels Möller. PGP-encrypted email is preferred. Keyid C0B98E26.
Internet email is subject to wholesale government surveillance.


More information about the gmp-devel mailing list