G
Guest
Guest
exerpts from technical forum at aceshardware entitled [bold]low performance of itanium[/bold]...
"C-program implementing 512x512-matrix to matrix multiplication with elements of long, float and double types":
itanium:
using "Linux64 OS and Cygnus IA64-gcc compiler"
long: 11 sec, float: 11 sec, double: 11 sec
using Linux64 OS and "SGI Pro64"
long: 2 sec, float: 1 sec, double: 1 sec
p4 1.4ghz, 100 qrd buss, pc 800 rdram, default msvc release optimiztaions (msvc 6.0):
long: 0.99, float: 1.18, double: 1.69
Tbird 900 mhz with PC133 RAM (Abit KT7-RAID):
long: 3.44, float: 4.23, double: 5.26
ES40 (EV67/667MHz), Tru64 v5.1, Compaq C (-fast):
long: 1, float: 1, double: 1
<A HREF="http://www.aceshardware.com/cgi-bin/ace/tech.pl?read=13858" target="_new">http://www.aceshardware.com/cgi-bin/ace/tech.pl?read=13858</A>
note: these are essentially non optimized code for the P4 and amd systems since msvc 6.0 doesn't offer sse(2)/3dnow! optimizations. if you take the time to read the thread you will notice also that simply switching compilers for the itanium increased performance by a factor of 10 and performance would increase even further by using intel's 5.0 compiler.
also anybody think it's kind of ironic that the 4 and the $ are the same key (at least us keyboard layout).
the results were taken for several different posts so i didn't use quotes too much but none of the above tests were done by me.
"C-program implementing 512x512-matrix to matrix multiplication with elements of long, float and double types":
itanium:
using "Linux64 OS and Cygnus IA64-gcc compiler"
long: 11 sec, float: 11 sec, double: 11 sec
using Linux64 OS and "SGI Pro64"
long: 2 sec, float: 1 sec, double: 1 sec
p4 1.4ghz, 100 qrd buss, pc 800 rdram, default msvc release optimiztaions (msvc 6.0):
long: 0.99, float: 1.18, double: 1.69
Tbird 900 mhz with PC133 RAM (Abit KT7-RAID):
long: 3.44, float: 4.23, double: 5.26
ES40 (EV67/667MHz), Tru64 v5.1, Compaq C (-fast):
long: 1, float: 1, double: 1
<A HREF="http://www.aceshardware.com/cgi-bin/ace/tech.pl?read=13858" target="_new">http://www.aceshardware.com/cgi-bin/ace/tech.pl?read=13858</A>
note: these are essentially non optimized code for the P4 and amd systems since msvc 6.0 doesn't offer sse(2)/3dnow! optimizations. if you take the time to read the thread you will notice also that simply switching compilers for the itanium increased performance by a factor of 10 and performance would increase even further by using intel's 5.0 compiler.
also anybody think it's kind of ironic that the 4 and the $ are the same key (at least us keyboard layout).
the results were taken for several different posts so i didn't use quotes too much but none of the above tests were done by me.