Seti@Home just killed ver 1.06

[Greebe]

big bummer

[DentyCracker]

why, was it very fast?

[ETHunter]

You should be runnning V3.0 now anyway..this version is faster than older version..on average..there is more diversity in unit speed than before though..however this is designed like this as it is better for the science..and it is that which counts when all the kicking and screaming has finished

------------------
PIII800@920MHz
Abit Be6-IIV2.0
256MB Crucial PC133 Cas2-2-2
4 of Seagate 20.2GB UDMA100 7200rpm 2MB HDD
2 of Maxtor 15.3GB UDMA100 7200rpm 2MB HDD
Hercules ProphetII GTS 64MB
Turtle Beach MontegoII Sound
Promise Ultra 100 UDMA Controller Card
Hitachi 12X DVD[40X CDROM]
Acer 52X CDROM
HP 9250i CDRW
Zip 250 Internal Drive
Elsa Video Capture Card

Support Seti@home, Support Science..support The Dataquest Trust

[Rags]

On my machine 1.06 was MUCH faster.

Rags

[Manex]

ETHunter, version 3.0 is *not* faster on every platform. I was running the version 2.4 text client on some HP 4-way compute servers with a 1 MB cache per processor and 2GB of system RAM. They used to process a unit in 2.75 to 3.25 hours, every time. Now, with the version 3.0 client, the same servers have not finished a unit in under 6.5 hours, and I've processed over 100 WU so far with the new client. I don't know what happened with HP-UX version of the client, but it is a factor of 2 times slower than the old client even when the angle_range parameter indicates it should be a fast unit....

[This message has been edited by Manex (edited 25 October 2000).]

[Guru]

Rags did you ever try 2.7 on your comp?

------------------
Join the MURC SETI team! | SETI @ MURC

"Cutting the space budget really restores my faith in humanity. It eliminates dreams, goals, and ideals and lets us get straight to the business of hate, debauchery, and self-annihilation."

[Greebe]

I tried 2.7 (and 3.0) both were slower for me than 1.06. This might be the time for me to say, Asta Seti!

[ETHunter]

I can't answer as to why your seeing what you do with the HP platform..although I suspect that the seti team have finally optimised the clients for average single cpu low level cache home pc's.

There have been similar slowdowns noticed with those running dual Xeon machines with large caches.

I have a Sun Ultra10 Graphics workstation..this has a 550MHz UltraSparcIIi cpu with 2MB full speed cache..yet it takes nearly 6 hours to do a unit..

Remember that they have added a lot more science to the newer clients so the fact that it is slower than 1.06 is not surprising really. They have probably found a way to truly utilise the cpu to it's bets abilities.

Greebe..If you are serious about dipping out because the client is slower then you are missing the point of this..It is a scientific experiment..firstly to prove that this type of project will work, but most importantly to analise research data in the hope of finding an artificial signal..thus proving we are not alone in this vast cosmos.

Stay and take part in what could be our future!

------------------
PIII800@920MHz
Abit Be6-IIV2.0
256MB Crucial PC133 Cas2-2-2
4 of Seagate 20.2GB UDMA100 7200rpm 2MB HDD
2 of Maxtor 15.3GB UDMA100 7200rpm 2MB HDD
Hercules ProphetII GTS 64MB
Turtle Beach MontegoII Sound
Promise Ultra 100 UDMA Controller Card
Hitachi 12X DVD[40X CDROM]
Acer 52X CDROM
HP 9250i CDRW
Zip 250 Internal Drive
Elsa Video Capture Card

Support Seti@home, Support Science..support The Dataquest Trust

[Rags]

Guru,

All versions after 1.06 I tried were slower.

Rags

[Rattledagger]

Rags, maybe you have some mild form of speed-bug in ver. 1.06 on your machine. One of TeamAnandtech got a heavy form with beta 2.70 spitting out was it 500 units on a day?

[jms]

I'm truly amazed that the SETI@Home team didn't kill v1.06 a long time ago. A lot of people where running this version with Ollie's modified FFT code. And NO, I am NOT suggesting that anyone on our team did this.

Jan M.

[Maggi]

Don't leave Greebe !

The 3.0 cmd-line is now below 5hours/WU on my Celery ... I bet you can't beat that on your Athlon ...

[nehalmistry]

i dont go as far back as the 1.x series, but i do know 3.0 is terrible for me.... i've been getting < 3%/hr.... that will take me forever to complete a WU(more than 30 hours)

after im done this last WU with 3.0 im moving back to 2.0....

------------------
P5A-B AMD K6-266@300
Matrox Milleniumm G200 AGP (oh, lets party)
Creative SB Awe32 (a classic, superb card)
Realtek 8029A NIC Card
64meg Ram
Ali V agp chipset
ICQ UIN: 24730025

[DentyCracker]

2.70 was the fastest on my Duron 700@850. 3.0 gives me 2 types 4.8 and 6+ hr wu, with win2k being on average a little faster than 98

[Helevitia]

Greebe! You can't stop crunching now! We are just getting started.

Dave