Документ взят из кэша поисковой машины. Адрес оригинального документа : http://hea.iki.rssi.ru/conf/hea2007/bp-0.2.97/tests/TIMING Дата изменения: Mon Dec 9 11:58:55 1996 Дата индексирования: Tue Oct 2 00:36:57 2012 Кодировка: Поисковые слова: п п п п п п п п п п п п

Here are some times for converting alldec.ref (2608 records) from refer
to bibtex. This used bp-0.1.1 and Perl5.001. bp-0.2.0 is about 20% faster.

Assuming alldec.ref is in the ref directory:

time ../bin/tconv.pl -convert alldec.ref -to foo.bib -format=refer,bibtex

(139 warnings, 802524 bytes)

Dell Pentium Pro 150MHz 28.8
Pentium Pro 133MHz BETA 32.6
HyperSPARC/125 45.7
Pentium 180 MHz 46.0
Sun SparcStation 20/61 46.1
Pentium 133 MHz 47.3
Pentium 100 MHz 61.0
HP 9000/735 83.0 (?? This number seems too slow)
Sun SparcStation 5/85 88.8
Gateway 2000 PentOD/83 96.2
Gateway 2000 486/66 137.2
Sun IPX 155.6
HP 9000/710 224.4 (?? also way too slow)



It is interesting to note that there was almost no gain from overclocking
the 133 MHz Pentium machine to 180 MHz. This same machine benefits greatly
from overclocking on many other tests, so we're guessing it's memory bound.
The SPARC machines do quite well, and the HP's do very poorly. This may
be caused by a bad compilation of Perl on the HPs. The Pentium Pro machines
do extremely well.