Obviously, it isn’t processor speed or FSB.

Team Stats :: http://setiweb.ssl.berkeley.edu/team_di ... amid=30229

Moderator: HAL6000

Post Reply
alnielsen
Posts: 38
Joined: Wed Jul 24, 2002 10:33 am
Location: Villa Park, IL
Contact:

Obviously, it isn’t processor speed or FSB.

Post by alnielsen » Thu Nov 06, 2003 11:44 am

As you can probably tell by the stats, I have started a new job and have been running SETI on a couple of computers. I found something I think is interesting.

IBM Netvista
Processor
Processor manufacturer: Intel
Processor (CPU)2: Intel Pentium 4
Processor internal clock speed2: 2.40GHz
Front Side Bus: 533MHz
512KB L2 Cache (for 2A GHz and Faster)
Processes a unit in 3.0-4.0 hrs.

IBM Thinkpad T40
Processor
BIOS type: Flash ROM
Processor manufacturer: Intel
Processor (CPU): Intel Pentium M processor
Processor internal clock speed2: 1.30GHz
Front Side Bus: 400MHz

1 MB On-Die Level 2 (L2) Cache
Processes a unit in 2.5-3.5 hrs.

The Pentium M processor, even tho it is almost ½ as fast and has a slower FSB, process the SETI unit faster. Must be the cache. Any comments?

Derek
Site Admin
Posts: 2477
Joined: Tue Jul 23, 2002 3:55 pm
Location: Canada
Contact:

Post by Derek » Thu Nov 06, 2003 11:59 am

I agree, it's probably the cache.

I'm not surprised to read this; I have heard that the Pentium M processor outperforms any other mobile chip out there by far.
-Derek

HAL6000
SETI Guru
Posts: 246
Joined: Wed Jul 24, 2002 7:11 pm
Location: Bloomfield, NJ U.S.A.
Contact:

Post by HAL6000 » Thu Nov 06, 2003 2:44 pm

The laptop I had was a 1Ghz P3 M with 512k, man that would smoke the 1Ghz coppermines :)

kuun
Post-O-Matic
Posts: 1566
Joined: Sat Nov 02, 2002 5:26 pm
Location: Tennessee
Contact:

Post by kuun » Fri Nov 07, 2003 12:20 am

ok...

here's the thing

larger cache will dramatically cutdown on data lag...

you give the CPU larger cache.. it has a shorter wait time for data to be stored and then transfered for processing.. also it can store and transfer and proccess more at once with larger cache :P
!!! WARNING !!!
The following forums: www.bp6.com
are infected with the following VIRUS(s): Kuun.infected.all.posts.Win2K.user

The following IRC servers has been exploited: irc.bp6.com
with the Following Exploit: Kuun.lurks.using.mIRC.v5.82.exploit

kuun
Post-O-Matic
Posts: 1566
Joined: Sat Nov 02, 2002 5:26 pm
Location: Tennessee
Contact:

Post by kuun » Fri Nov 07, 2003 12:21 am

ok...

here's the thing

larger cache will dramatically cutdown on data lag...

you give the CPU larger cache.. it has a shorter wait time for data to be stored and then transfered for processing.. also it can store and transfer and proccess more at once with larger cache :P



why do you tihnk XEONS being mostly slower than the top of the line pentiums will smoke a pentiums 2-3 times over

most new xeons have 2MB l2/l3 cache

cache is everything
!!! WARNING !!!
The following forums: www.bp6.com
are infected with the following VIRUS(s): Kuun.infected.all.posts.Win2K.user

The following IRC servers has been exploited: irc.bp6.com
with the Following Exploit: Kuun.lurks.using.mIRC.v5.82.exploit

estikid
Posts: 2
Joined: Thu Jul 25, 2002 5:11 pm
Location: United Kingdom
Contact:

Post by estikid » Sat Nov 08, 2003 11:53 am

Very true kuun, i remember reading an article about seti and cache usage, i think it may have been on this page - http://www.paradiseplace.org.uk/other/setiadvd.htm
I was wondering why my 1Ghz celeron was being outperformed by a p3 800mhz (on seti).
I dont know what to put here.

uallas5
Posts: 26
Joined: Mon Aug 05, 2002 10:09 am
Location: Watertown, MA

Post by uallas5 » Thu Nov 13, 2003 10:31 am

Also. it's the greatly improved and shortened pipelines combined with better branch prediction of the Pentium M. In most head to head tests on compareable hardware, the M will smoke the desktop P4 on a purely Ghz standpoint.

Anand has a really good review http://www.anandtech.com/mobile/showdoc.html?i=1800&p=1 from shortly after the M had been released.
"I Know Good From Bad and Where to Get the Refill" - Mr. Harding.

jives
Posts: 19
Joined: Fri Nov 01, 2002 8:43 am
Location: NE, USA
Contact:

Post by jives » Mon Dec 08, 2003 10:05 am

Part true part not.

The Anandtech article shows there is more to the Pentium M's gain than just seti.

Cache does help because yes my Xeon 450 w/1M out pace my PIII 600s w/512k, 750s w/256k and 850s w/256k. My Alpha EV6 500 w/4M can out my pace PIII 933 w/256k and comes within a few minutes of my Celeron 1.4G w/256k.

But it isn’t everything since I have PIII 1.2G w/256k and Athlon MP 1.2G w/384k (I think?) which are faster than the Alpha and Xeon.

So upping the CPU speed will gain you some, upping the FSB will get you some as well and upping the cache will speed it up. But as the CPU gets faster cache increases will grow in importance, the faster the CPU the more it is about the cache.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest