Rosetta@Home - 11/05: Stats & milestones

[BlackMountainCow]

No new members today


[TA]R@H is now 219 members strong. That number excludes some additional 30 people combined in TeAm Enterprise and about 11 in the Rebel Alliance.




Need some stats? Take a look at Free-DC, BOINC-Stats and BOINC-Synergy

Got some questions? Try the Official R@H FAQ (work in progress).

Want to know something about the medical relevance of Rosetta? Take a look at this article.

Interested in alpha testing for Rosetta? Take a look at RALPH.




Yesterday's highest overall climber: James Wisnieski (2,202 spots)

Yesterday's highest TeAm climber: CD (4 spots)


Yesterday's production for TA: 90,163 (day before: 90,200 = -37)

TA's RAC 88,391

TA's total score 30,588,268

TA's world rank is: #4

TA's production rank today #2




Milestone Makers:

none






Yesterday's top 25 producers:

• TeAm Enterprise (15,695 creds)
  
• Dennis (12,002 creds)
  
• uallas5 (6,638 creds)
  
• BadThad (4,705 creds)
  
• teriba (2,966 creds)
  
  
• Dulanic (2,635 creds)
  
• TA_TheReaper (2,555 creds)
  
• Night181 (2,332 creds)
  
• Insidious (2,175 creds)
  
• strider7 (2,133 creds)
  
  
• Pokey (1,963 creds)
  
• aaronpriest (1,893 creds)
  
• Neurodog (1,715 creds)
  
• Fardringle (1,639 creds)
  
• ken008 (1,607 creds)
  
  
• Rudy Toody (1,599 creds)
  
• The Arns Law Firm (1,459 creds)
  
• Paul (1,187 creds)
  
• fusuikan (1,183 creds)
  
• CrimsonWolf (1,120 creds)
  
  
• jw.middleton (1,018 creds)
  
• Rebel Alliance (964 creds)
  
• OhioDude (836 creds)
  
• kmmatney (820 creds)
  
• Plox (803 creds)

SitRep for the TeAm and Rosetta News:

• Yesterday, we had 93 active members (42.47%).
  

:beer::beer:

BMC with thx to CyGoR

 

[Rudy Toody]

Thanks for the stats, BMC!

 

[BadThad]

Thanks BMC!

 

[Insidious]

:moon:

 

[apriest]

Well, I got one of the new Dual CPU, Dual Core Xeon 5130s we are building fired up temporarily today. Four Woodcrest cores @ 2GHz w/ 1333MHz FSB and 4MB shared cache (per socketed CPU), quad channel DDR2-667MHz memory bus with 4x512MB (2GB Total) memory, and 5x Seagate NL35 250GB SATA-II 16MB cache drives in RAID 5 baby! It is in testing and will have the OS swapped out for a final build later on, but I wanted to get R@H up and running on it to see what it's capable of.

I'm running a custom-compiled x86_64 SMP Linux kernel (2.6.18.2) and 32-bit Boinc since 64-bit apparently has issues with R@H and they say there's no 64-bit advantage with this project, unlike SETI.

I'm seeing 1225.97 million ops/sec for measured floating point speed and 2568.58 million ops/sec for measured integer speed so far in Linux. I had floating point up to 1841.79 million ops/sec when I put Vista RC2 on it, so I'm not sure what the discrepancy is there. My Pentium M 2GHz laptop puts out 3477.55 million ops/sec for integer speed, so I'm surprised it can't at least match that. In fact, I have three slower Pentium M laptops at 1.73GHz and 1.86GHz that also beat the new beast in benchmarks. Oh well. 4 cores cranking away when it's idle will blow away the laptops soon enough. And when the second server gets built... wheeee!

http://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=349097

 

[BlackMountainCow]

OMG, what a freakin' nice addition to the fleet! :thumbsup:

 

[Smoke]

Nice addition there, apriest!

The measurement I like for comparison purposes is the RAC number. It is too bad you have to wait 30 days for this number to reach the true level for any new system (or newly installed).

The floating point and integer speed numbers always seem sort of arbitrary to me.

The real measure is "how much meat a system puts on the table" ... RAC.

I'm about 10 days into a fresh new OC on one of my systems and can't wait to see the final maintained RAC.

 

[apriest]

Hey Smoke,

Yeah, I agree, the RAC is what counts at the end of the day! I've got an older dual Athlon MP 1900+ in service that has "RAC'd" up to 301.22. It's #1 in my fleet. 2nd place is a Pentium D 3GHz with a RAC of 274.93, and #3 is another dual Athlon MP 1900+ with 272.61. Those old Athlons really crank out the WUs. The highest measured speeds all come from the laptops, but since they don't run 24x7 and are often on battery, the RACs are pretty low on them. I've also had to limit CPU throttling to 50% so they don't overheat. The Pentium D in 2nd place for my fleet also is limited to 50% to keep the fans from sounding like a turbo jet. It would probably have a higher RAC otherwise.

So yeah, I've got high hopes for the new Xeon 5130 servers and we'll see how they look in a month or so. I have them set to acoustic mode instead of performance mode in the BIOS to keep the fan speed low so I don't have to limit the CPU to 50% or fewer cores like the Pentium Ds. The BIOS throttles back the bus and CPU if they start getting hot before ramping fan speed. I've found I get the best server performance this way as when Rosetta is not running the CPUs never get hot and they run at full speed for whatever task they need, but when it's idle and Rosetta goes all out on 4 cores the BIOS downclocks some and things don't get hot. It appears to crank out WUs much faster than setting limits in Rosetta itself as far as CPU throttling or limiting CPU#s. I wish every BIOS had this feature!

What are you OC'ing?

 

[Smoke]

I've got one of the PentiumD 805 systems built to the Tom's Hardware specs. I've had it running as high as 4.05 GHz but nothing useful could be done at that level. I've been trying to find the sweet spot ... where it can run as fast as it can go without causing errors when crunching Rosetta.

Smoke-Vista

It's been running under the current configuration since late in the day of October 26th. What makes me think I may have found the sweet spot is the number of "Results" being reported. I think this number is based on a much shorter time span (maybe 10 days or so). 194 - 196

 

[apriest]

Nice! I've heard the 805 is quite a value, many people have OC'd that sucker to the stratosphere.

I've got 58 results in the past 2-3 days on the Xeon. It's lower than I think it could be because for the past 8hrs all of the results have had client or compute errors. I'm not sure why. None of the rest of my machines have errors.

 

[imaheadcase]

I ordered my parts for my new Core 2 duo system today, be running BOINC on that hopefully this weekend.

Hopefully i get a decent OC to 3ghz or more with the 1.86ghz one. Its a win win for the team, cause now i have to work to make up for all the money I spent so computer can use BOINC more!

 

[Smoke]

lol @ imaheadcase

Some people work to put food on the table and others are like us.

Crunch On! :beer:

 

[JWMiddleton]

Hi guys! Light day, eh? Well, I am partially to blame. I left Saturday about noon to drive to Charlotte, NC, for 10 nights, while working on a network implementation. When I checked my R@H stats I found them to be down. So, I look at the computers on my account and found that my two dedicated boxes have not sent in resulte since the day I left. There must be some corollary to Murphy?s Law that applies here.

Cruch on!