TAS's member thread for July 2007: Leiden classical.

[petrusbroder]

Edit July 22: Here are the stats for 3rd week of crunching.

Edit July 15: Here are the stats for the second week of crunching!

Edit July 7: Here are the stats for the first week of crunching ...

Edit July 1: The starting stats are here ...


The poll is closed.
Leiden Classical won by the closest possible margin: 1 vote (of 11 votes).
Starting stats collected - will be posted @ TAS later today.

Lets get crunching!

 

[TallBill]

Sweetness, just switched over Looks like the results showing up are not instant. Gonna take a little getting used to that.

 

[BlackMountainCow]

Last time I did Leiden they were instant

 

[petrusbroder]

Now they do a quorum of two and two replications (if the quorum is not reached).
If there are 16 computation errors in a row, then that WU is dropped.
I had that happen on my Mac - 30 WUs errored out.
It took them only minutes - 15 minutes at the most - but still: errors do not give credits, so I took my Macs off that project.

No probs with PCs yet.

 

[petrusbroder]

Here is a proposal for the rules for the July challenge.
I would be happy for input and discussion.

I was thinking "as few rules as possible" - but I think OTOH that some rules have to be agreed upon. :Q

 

[Rudy Toody]

Originally posted by: petrusbroder
Here is a proposal for the rules for the July challenge.
I would be happy for input and discussion.

I was thinking "as few rules as possible" - but I think OTOH that some rules have to be agreed upon. :Q

As I also stated on that thread, "I like it."

Because this format requires some crunching expertise, a person can win the belt multiple times.

 

[BlackMountainCow]

Question: Does Leiden rely on benchmarks or do they have fixed credit?

 

[BlackMountainCow]

Originally posted by: BlackMountainCow
... do they have fixed credit?

Guess they don't

 

[caferace]

Seven home cores attached. Usual slow start expected until I clear out the previous project WU's and get the office comps involved. IIRC the PIII army is no go on Leiden (from last fall). Maybe I'll try again. They were strong contributors on RCN...

-jim

 

[BlackMountainCow]

My PIII-600 works fine with Leiden.

 

[petrusbroder]

Originally posted by: BlackMountainCow
My PIII-600 works fine with Leiden.

Yeah, so does mine too - but it is the limited Win XPe which may or may not play nice with the different applications ...

 

[petrusbroder]

Here are the starting stats for Leiden Calssical.
Lets get going ...

 

[caferace]

Two more cores on.

-jim

 

[TallBill]

Blah, Boinc is crashing again..

 

[Rudy Toody]

Leiden Classical penalizes the faster host!

Two of my hosts received the same WU. The credit awarded was that requested by the slower cpu.

As I see it, the fastest host on the project will never get the credits it requests!

--Fred

 

[BlackMountainCow]

That's true, but with stock BOINC, the difference should be around 2 creds max; at least that's what I can tell from my results.

 

[TallBill]

Biggest difference I've seen is 2 credits, just browsing through random Wu's.

 

[petrusbroder]

Originally posted by: Rudy Toody
Leiden Classical penalizes the faster host!

Two of my hosts received the same WU. The credit awarded was that requested by the slower cpu.

As I see it, the fastest host on the project will never get the credits it requests!

--Fred

Yupp, that is how things have been for a long time: When the quorum was three results, the highest and lowest requested credits were discarded and the the middle one was awarded. When the quorum was two often the mean was awarded - or the lower ... depending on project. The diff was often very small. OTOH a difference of 2 credits is a lot when each WU gives 10 - 20 credits, and very small when each WU gives 60 - 120 credits ...

 

[petrusbroder]

Quick news:

TeAm Anandtech just climbed from 8:th rank to 7:th rank in Leiden Classical. :beer:

Look at the TAS-affect!

 

[rabrittain]

When I attached Leiden to my BOINC Managers, the account was named "Rex" instead of "rabrittain". All of my other BOINC accounts (R@H, E@H, etc.) are still named "rabrittain". Does anyone know why this may have happened. Also, the name at the top of my BOINC signature has changed to "Rex". Can someone please explain this to me?

;

 

[Rudy Toody]

Originally posted by: rabrittain
When I attached Leiden to my BOINC Managers, the account was named "Rex" instead of "rabrittain". All of my other BOINC accounts (R@H, E@H, etc.) are still named "rabrittain". Does anyone know why this may have happened. Also, the name at the top of my BOINC signature has changed to "Rex". Can someone please explain this to me?

;

Just to let everyone know we fixed this. When one attaches to a project for the first time, the Windows account name is the default, so this meant he was logged on as Rex and forgot to change it to rabrittain.

 

[BlackMountainCow]

Hehe, just wanted to write the same thing. Happened to me sometimes as well, when asslaminating machines.

 

[caferace]

For many of you guys, +103F isn't hot. And I can take it every so often. Unfortunately, we have it predicted for the next 4-5 days or so. And the comps being upstairs with minimal A/C, conditions are ripe for machine failure, not to mention the uncomfortable and expensive part of it.

So, in the morning I'm shutting all boxes at home down with a couple of exceptions. I'll bring 'em back on when weather stabilizes..

http://www.wrh.noaa.gov/foreca...&textField2=-121.21875

-jim

 

[Rudy Toody]

It's not that hot in Seattle, but with no AC I've had to UC the OC on the PC to keep doing DC!

--Fred

 

[BlackMountainCow]

:laugh: @ Rudy


Here in Germany it's pretty cold for the time of they year; just some 60F to 65F