Simpler Navigation for Servers and Operating Systems - Please Update Your Bookmarks
Completed: a much simpler Servers and Operating Systems section of the Community. We combined many of the older boards, so you won't have to click through so many levels to get at the information you need. Check the consolidated boards here as many sub-forums are now single boards.
If you have bookmarked forums or discussion boards in Servers and Operating Systems, we suggest you check and update them as needed.
General
cancel
Showing results for 
Search instead for 
Did you mean: 

I/O performance on HP-9000 D390 machine

Nandan gogate
Occasional Contributor

I/O performance on HP-9000 D390 machine

Hello,

I am in the process of comparing I/O performance on sun ultra sparc and HP.
I found that Hp-9000, D390's I/O performance is around 40 % inferior to the I/O performance on ultra-2 using solaris 2.5.

processor speed (which i dont think matters) is almost the same. I am measuring the performance using the following simple program.

#include
#include
#include

#define error
{
printf("Error , errno: %dn",errno);
return 1;
}

main()
{
char buf[8192];
int i=0;
for (i=0; i< 8000; i++)
buf[i]='a';
buf[8000]='0';
FILE *fp = fopen("/export/p2/dali/wrF", "w");
if(!fp) error;
for (i=0; i < 3000; i++)
{
fprintf(fp, "%s",buf);
fsync(fileno(fp));
}
}

Then i use timex cmd to find the time required. I make sure the executable is not on the nfs partition. The directory where i create the file (/export/p2/dali/wrF) is on a local partition. I make sure I delete the file before I run the program.

My questions are

1) Is this an acceptable way of measuring an approximate I/O performance ?

2) Has anybody seen such an inferior performance on HP machine compared to Sun machines ?

Any information in this regard is highly appreciated.

Thanks
3 REPLIES
Carol Garrett
Trusted Contributor

Re: I/O performance on HP-9000 D390 machine


What speed are the disks on your Sun and HP ?
This may explain the difference. Test their speeds using time dd if=/dev/rdsk/... of=/dev/null bs=1024k count=50 (for 50Mb)

Lets see if the disks are of similar performance.
Alexander M. Ermes
Honored Contributor

Re: I/O performance on HP-9000 D390 machine

Hi.
This may partially result from the bus topology difference between HP and SUN.
If you connected the SCSI interfaces to the
HP-PB bus, that already makes a lot of a difference. If you have done so, try to get interfaces for the HSC bus.
Rgds
Alexander M. Ermes
.. and all these memories are going to vanish like tears in the rain! final words from Rutger Hauer in "Blade Runner"
Dan Hetzel
Honored Contributor

Re: I/O performance on HP-9000 D390 machine

Hi,

It's difficult to say if you're not comparing apples with oranges...

Are the 2 disks (HP/SUN) of the same type and same capacity ? Comparing a 9Gb disk with a 36Gb one will always give speed measurements in favor of the large capacity disk.

As Carol suggested you, you should use dd and let the Operating System do the 'fsync' for you.

Dan
Everybody knows at least one thing worth sharing -- mailto:dan.hetzel@wildcroft.com