Benchmark testy

[Jan Hubicka]

> Za to cecko od SunSoftu je asi potrebne si priplatit. :(
Cecko od sunsoftu dela u prumernyho programu asi 40% uryhleni na
Ultrasparcu a zpomaluje na normalich sparcech...
> 
> > 
> > Jinak je potreba se smirit s tim, ze chci-li pracovni stanici, pak PC
> > s 200 MHz Pentium Pro a Linuxem je nejen levnejsi, ale v nekterych
> > pripadech i rychlejsi nez Ultra-1 (167 MHz). Rada lidi si to stale
> > jeste neuvedomuje.
> 
> Ano, teraz to uz viem, rozpadli sa moje iluzie o vykone SPARCov.
> Pre zaujimavost, urobil som este jeden test, cisto len procesora.
Na druhou stranu Ultra-1 ma nesrovnatelne hezci architekturu. Srovnani
s gcc je ponekud nespravedlive, protoze podpora Ultry je zatim chaba.
Ultra-1 vede v jinych situacich nez je chroustani floatu (kde pentium
typicky vede, protoze ma floaty nelogicky rychlejsi nez integery).
Ultra-1 ma lepsi sbernici apod..takze vede v komplikovanejsich vecech
Navic solaris patri IMO k jednem z nejpomalejsich UNIXu a instalace Linuxu
casto pomuze :)
> 
> 	integer		floating point
> gcc	1.870		1.494
> cc	1.716		1.424
> 
> Index je porovnanie k vykonu pocitaca DELL s procesorom Pentium 90.
> (U alphy sa tie indexi pohybovali okolo 6)
Moje testy u XaoSe na Ultra-1 koncily asi: pentium 130-7.15 smycek/sec
ultra-10 smycek/sec gcc, 13 smycek/sec cc -fast

Honza
> 
> 
> 					jano
> 
>   BYTE UNIX Benchmarks (Version 4.1.0)
>   System -- SunOS ultra 5.5.1 Generic sun4u sparc SUNW,Ultra-1
>   Start Benchmark Run: Fri Oct 31 09:46:30 MET 1997
>    1 interactive users.
>   
>   -r-xr-xr-x   3 bin      root       89564 May  3  1996 /bin/sh
>   /bin/sh:	ELF 32-bit MSB executable SPARC Version 1, dynamically linked, stripped
>   /dev/dsk/c0t1d0s4     989246  145435  744891    17%    /apps
> Dhrystone 2 using register variables     267417.8 lps   (10.0 secs, 10 samples)
> Double-Precision Whetstone                   54.4 MWIPS (9.9 secs, 10 samples)
> System Call Overhead                      45471.7 lps   (10.0 secs, 10 samples)
> Pipe Throughput                           42019.5 lps   (10.0 secs, 10 samples)
> Pipe-based Context Switching              14410.0 lps   (10.0 secs, 10 samples)
> Process Creation                            322.6 lps   (30.0 secs, 3 samples)
> Execl Throughput                             99.0 lps   (29.7 secs, 3 samples)
> File Read 1024 bufsize 2000 maxblocks     27361.0 KBps  (30.0 secs, 3 samples)
> File Write 1024 bufsize 2000 maxblocks     3711.0 KBps  (30.0 secs, 3 samples)
> File Copy 1024 bufsize 2000 maxblocks      3518.0 KBps  (30.0 secs, 3 samples)
> File Read 256 bufsize 500 maxblocks        7371.0 KBps  (30.0 secs, 3 samples)
> File Write 256 bufsize 500 maxblocks       3771.0 KBps  (30.0 secs, 3 samples)
> File Copy 256 bufsize 500 maxblocks        3239.0 KBps  (30.0 secs, 3 samples)
> File Read 4096 bufsize 8000 maxblocks     71315.0 KBps  (30.0 secs, 3 samples)
> File Write 4096 bufsize 8000 maxblocks     3414.0 KBps  (30.0 secs, 3 samples)
> File Copy 4096 bufsize 8000 maxblocks      1640.0 KBps  (30.0 secs, 3 samples)
> Shell Scripts (1 concurrent)                 95.3 lpm   (60.0 secs, 3 samples)
> Shell Scripts (8 concurrent)                 18.0 lpm   (60.0 secs, 3 samples)
> Shell Scripts (16 concurrent)                 9.0 lpm   (60.0 secs, 3 samples)
> Arithmetic Test (type = short)            15809.4 lps   (10.0 secs, 3 samples)
> Arithmetic Test (type = int)              16510.2 lps   (10.0 secs, 3 samples)
> Arithmetic Test (type = long)             16535.2 lps   (10.0 secs, 3 samples)
> Arithmetic Test (type = float)            44153.1 lps   (10.0 secs, 3 samples)
> Arithmetic Test (type = double)           34550.8 lps   (10.0 secs, 3 samples)
> Arithoh                                  1172977.0 lps   (10.0 secs, 3 samples)
> C Compiler Throughput                       118.0 lpm   (60.0 secs, 3 samples)
> Dc: sqrt(2) to 99 decimal places           3764.0 lpm   (30.0 secs, 3 samples)
> Recursion Test--Tower of Hanoi             4995.3 lps   (20.0 secs, 3 samples)
> PerlBench                                  no measured results
> select()                                 460503.4 lps   (0.1 secs, 3 samples)
> 
> 
>                      INDEX VALUES            
> TEST                                        BASELINE     RESULT      INDEX
> 
> Dhrystone 2 using register variables        116700.0   267417.8       22.9
> Double-Precision Whetstone                      55.0       54.4        9.9
> Execl Throughput                                43.0       99.0       23.0
> File Copy 1024 bufsize 2000 maxblocks         3960.0     3518.0        8.9
> File Copy 256 bufsize 500 maxblocks           1655.0     3239.0       19.6
> File Copy 4096 bufsize 8000 maxblocks         5800.0     1640.0        2.8
> Pipe Throughput                              12440.0    42019.5       33.8
> Pipe-based Context Switching                  4000.0    14410.0       36.0
> Process Creation                               126.0      322.6       25.6
> Shell Scripts (8 concurrent)                     6.0       18.0       30.0
> System Call Overhead                         15000.0    45471.7       30.3
>                                                                  =========
>      FINAL SCORE                                                      18.2
> 
> 
> ----- I can't get no satisfaction -----
> 			Rolling Stones
>  
> 

------------------------------------------------------------------------------
                   Have you browsed my www pages? Look at:
                       http://www.paru.cas.cz/~hubicka
      Koules-the game for Svgalib,X11 and OS/2,  Xonix-the game for X11
      czech documentation for linux index, original 2D computer art and
              funny 100 years old photos and articles are there!