octane 2x250 vs 1x300 vs 1x400 CPU's

Hi all,
I just upgraded to a newer octane, and played around with trying to
compare my old dual 250 mhz R10k to a new single 300mhz R12k and single
400mhz R12k.

I didnt really think the single 300mhz CPU would be much faster than a
dual 250 mhz CPU setup, but I was wrong :) Had I known I would have
upgraded much sooner.

I did two types of benchmarks to give me a general idea, a gzip/gunzip
comparision on a 1.3Gb tarball, and an xmame test :)

I should add the fact the the dual 250mhz octane is an older model
octane, and the 300mhz and 400mhz tests were done on a newer model
octane2, but with older octane1 gfx.

here are the results using the sgi-freeware built "xmame (x11) version
0.61.2 (Dec 2 2002)" running "Pulstar" - my favorite Neo-Geo shooter
game :) (note - audio was turned off).

on the dual 250mhz, frame rate is about 34 fps
on the single 300mhz, frame rate is about 45 fps
on the single 400mhz, frame rate is about 55 fps


the second test involved gzip'ing and gunzip'ing a 1.3Gb tarball. The
same system disk and option drive were used (both 10k rpm).

==================================

dual 250mhz

octane 14# ls -lh
total 2823200
-rw-r--r-- 1 root sys 1.3G Dec 3 14:26 IRIX6.5.22m.tar

octane 15# time gzip IRIX6.5.22m.tar
985.1u 31.4s 17:18 97% 0+0k 22074+0io 1pf+0w

octane 17# time gunzip IRIX6.5.22m.tar.gz
166.3u 21.9s 3:17 95% 0+0k 19931+0io 0pf+0w

==================================

single 300mhz

octane 12# ls -lh
total 2823200
-rw-r--r-- 1 root sys 1.3G Dec 3 15:38 IRIX6.5.22m.tar

octane 13# time gzip IRIX6.5.22m.tar
802.1u 22.7s 13:59 98% 0+0k 22079+0io 1pf+0w

octane 14# time gunzip IRIX6.5.22m.tar.gz
131.4u 17.0s 2:40 92% 0+0k 19998+1296io 3pf+0w

==================================

single 400mhz

octane 9# ls -lh
total 2823200
-rw-r--r-- 1 root sys 1.3G Dec 3 17:07 IRIX6.5.22m.tar

octane 10# time gzip IRIX6.5.22m.tar
600.1u 18.4s 10:29 98% 0+0k 22075+0io 1pf+0w

octane 11# time gunzip IRIX6.5.22m.tar.gz
98.3u 14.2s 2:11 85% 0+0k 19944+1300io 1pf+0w

==================================

I hope this info can help someone.

cheers!
mike

mike wrote:[color=blue]
> Hi all,
> I just upgraded to a newer octane, and played around with trying to
> compare my old dual 250 mhz R10k to a new single 300mhz R12k and single
> 400mhz R12k.
>
> I didnt really think the single 300mhz CPU would be much faster than a
> dual 250 mhz CPU setup, but I was wrong :)[/color]

Much of the speedup can be attributed to the 2MB L2 (secondary) cache
in the R12K CPU module. Most of the R10K-250 duals had only 1MB
L2 cache. The clock frequency increase helps, but not nearly as
much as the secondary cache.

--
Greg Douglas
Reputable Systems
[url]http://www.reputable.com[/url]

whst do you think of dual 360MHz vs single 400MHz ?

thanks

Rick

In article <3FD3A50C.FB415E47@comcast.net>, mike wrote:[color=blue]
> Hi all,
> I just upgraded to a newer octane, and played around with trying to
> compare my old dual 250 mhz R10k to a new single 300mhz R12k and single
> 400mhz R12k.
>
> I didnt really think the single 300mhz CPU would be much faster than a
> dual 250 mhz CPU setup, but I was wrong :) Had I known I would have
> upgraded much sooner.
>
> I did two types of benchmarks to give me a general idea, a gzip/gunzip
> comparision on a 1.3Gb tarball, and an xmame test :)
>
> I should add the fact the the dual 250mhz octane is an older model
> octane, and the 300mhz and 400mhz tests were done on a newer model
> octane2, but with older octane1 gfx.
>
> here are the results using the sgi-freeware built "xmame (x11) version
> 0.61.2 (Dec 2 2002)" running "Pulstar" - my favorite Neo-Geo shooter
> game :) (note - audio was turned off).
>
> on the dual 250mhz, frame rate is about 34 fps
> on the single 300mhz, frame rate is about 45 fps
> on the single 400mhz, frame rate is about 55 fps
>
>
> the second test involved gzip'ing and gunzip'ing a 1.3Gb tarball. The
> same system disk and option drive were used (both 10k rpm).
>
>==================================
>
> dual 250mhz
>
> octane 14# ls -lh
> total 2823200
> -rw-r--r-- 1 root sys 1.3G Dec 3 14:26 IRIX6.5.22m.tar
>
> octane 15# time gzip IRIX6.5.22m.tar
> 985.1u 31.4s 17:18 97% 0+0k 22074+0io 1pf+0w
>
> octane 17# time gunzip IRIX6.5.22m.tar.gz
> 166.3u 21.9s 3:17 95% 0+0k 19931+0io 0pf+0w
>
>==================================
>
> single 300mhz
>
> octane 12# ls -lh
> total 2823200
> -rw-r--r-- 1 root sys 1.3G Dec 3 15:38 IRIX6.5.22m.tar
>
> octane 13# time gzip IRIX6.5.22m.tar
> 802.1u 22.7s 13:59 98% 0+0k 22079+0io 1pf+0w
>
> octane 14# time gunzip IRIX6.5.22m.tar.gz
> 131.4u 17.0s 2:40 92% 0+0k 19998+1296io 3pf+0w
>
>==================================
>
> single 400mhz
>
> octane 9# ls -lh
> total 2823200
> -rw-r--r-- 1 root sys 1.3G Dec 3 17:07 IRIX6.5.22m.tar
>
> octane 10# time gzip IRIX6.5.22m.tar
> 600.1u 18.4s 10:29 98% 0+0k 22075+0io 1pf+0w
>
> octane 11# time gunzip IRIX6.5.22m.tar.gz
> 98.3u 14.2s 2:11 85% 0+0k 19944+1300io 1pf+0w
>
>==================================
>
> I hope this info can help someone.
>
> cheers!
> mike
>
>[/color]

> whst do you think of dual 360MHz vs single 400MHz ?

dual CPU's are better I would think in this case - they make the system seem
a little more snappy. At the time when I had the impulse to buy the CPU
upgrade, the dual 360's were going for around $1200, and the single 400mhz
was $400. I just went for the cheaper deal.