🚀 go-pugleaf

In article <1992Oct31.055451.3552@netcom.com> wolfgang@netcom.com (Wolfgang Henke) writes:
>
>Charles R. Hoynowski (charles@netcom.netcom.com) wrote:
>: In article <1992Oct29.174104.8803@netcom.com> wolfgang@netcom.com (Wolfgang Henke) writes:

[stuff deleted]

Here are some benchmark comparisons between the Sun SS10/41 and HP9000/735:

Dhrystone 1.1 (Yes I know 2.1 is better but the vendors use 1.1 so that's
what we use too to check their numbers):

With registers and optimization, 500000 iterations
HP9000/735	203252 Dhrystones/sec
Sun SS10	118577 Dhrystones/sec


Whetstone:

Single precision
HP9000/735	161290 KWhetstones/sec
Sun SS10	 65217 KWhetstones/sec

Double precision,with optimization,50M inst):
HP9000/730	147059 KWhetstones/sec
Sun SS10	 62500 KWhetstones/sec


Double precision, rolled
HP9000/735	36140 Kflops
Sun SS10        12875 Kflops

Double precision, unrolled
HP9000/735	34333 Kflops
Sun SS10         8240 Kflops

Time for make on a given file (GKS):
                 real           user             syst
HP9000/735	 11.00          10.23            0.47
Sun SS10         36.00          34.05            1.12


Musbus I/O test:

File Size: 250 blocks
  Write:
HP9000/735	12500.0 kBytes/s
Sun SS10         2483.7 kBytes/s
  Read:
HP9000/735	25000.0 kBytes/s
Sun SS10        10869.6 kBytes/s
  Copy:
HP9000/735	 7638.9 kBytes/s
Sun SS10         2016.9 kBytes/s



Bo
-- 
   ^   Bo Thide'--------------------------------------------------------------
  |I|        Swedish Institute of Space Physics, S-755 91 Uppsala, Sweden
  |R|      Phone: (+46) 18-303671.  Fax: (+46) 18-403100.  IP: 130.238.30.23
 /|F|\          INTERNET: bt@irfu.se      UUCP: ...!mcvax!sunic!irfu!bt  
 ~~U~~ ----------------------------------------------------------------sm5dfw-