[zombie67]

Team MacNN is currently in 12th place.

The good news: In about 4 days, we will pass "Housing and Food Services" to take 11th.

The bad news: in about 16 days, we will be passed by SETI.USA, to fall back to 12th. In the past couple of weeks, they jumped from about 50k/day to 90k/day. Kevint and Fish are the two primary culprits. Clearly they're on a mission, and successful. Check out their monthly chart at:

http://www.boincstats.com/stats/team...r=rosetta&id=6

We need to add resources, and fast. Let's get on it!

[Shaktai]

Originally Posted by zombie67

Team MacNN is currently in 12th place.

The good news: In about 4 days, we will pass "Housing and Food Services" to take 11th.

The bad news: in about 16 days, we will be passed by SETI.USA, to fall back to 12th. In the past couple of weeks, they jumped from about 50k/day to 90k/day. Kevint and Fish are the two primary culprits. Clearly they're on a mission, and successful. Check out their monthly chart at:

http://www.boincstats.com/stats/team...r=rosetta&id=6

We need to add resources, and fast. Let's get on it!

Well, I'm back for at least a bit. Not a lot of crunch power, but i'll add a few points a day to Rosetta. Still on hold waiting for my job relo to the Philippines.

[zombie67]

Originally Posted by Shaktai

Well, I'm back for at least a bit. Not a lot of crunch power, but i'll add a few points a day to Rosetta. Still on hold waiting for my job relo to the Philippines.

We'll take it!

Sorry about the bump in stats today. I got a bit OCD with jedi_mstr and just can't stop.

[ritzl]

In the FWIW category:

1) The science that is being done by all the DC projects has merit. The very nature of supercomputing/DC is to harness the power of the here and now, such that the future is enabled. In the case of SETI, this is just fun as who wouldn't want to be the first to discover "the signal." Einstein is probably also in the "meaning of life, the universe, and everything else" category. Basic science pure and simple.

In the case of the protein structure projects, this has a more relevant implication - Rational Drug Design (RDD). RDD is to electronics as electricity is to "chemical/shotgun" drug design, but in its infancy (i.e. the transistor hasn't been invented yet?). I have spent some time with people that are trying to produce unknown-structure protein crystals in microgravity with some success. That top-down experimental process is very expensive and time consuming. I see this DC, bottom-up, refinement of protein structure as narrowing the gap between/pointing the way toward more productive analytical-to-empirical protein structure experiments.

This is a VERY good thing, and there is a REAL benefit to humanity in this (protein structure) research.

2) Some projects are better suited to the processing advantages of different platforms. I am a Mac user. Rosetta does not seem to be conducive to that effort. My results (viewed in terms of credits) are worth 40% of a comparable windows machine.

3) Windows is the vast majority of the DC target market and will get initial apps tailored to that platform. As a Mac user, I understand that. And I also understand that a project needs to cater to, and rely upon the familiarity with, that platform. This does not change my belief that a killer Mac worker is out of reach of the project. If a/this project wants that 10-20% increase in DC processing, they know where to look.

4) Back to RDD. Given the financial/business implications of success in this area, who owns the product of the DC results? If a protein goes on to be a source for a successful therapy (drug), who gets the financial reward? At least in SETI and Einstein and all other basic science projects, the answer to that question is simply the joy of contribution.

In the RDD area, while I realize that protein structure characterization is all basic science now, it is leading to [potentially, very] lucrative drug production. It would be my suggestion that a statement be made that some financial distribution construct be put forth so that people get some small, long-term return on their investment in these projects.

There may be ZERO financial benefit given the vagaries of drug design, but there may be some. That's the risk.

I believe that a lot of people have invested real money to support the goals of these projects, and deserve a real return -- IFF it is oriented to a commercial (however far in the future) outcome, and a return is in the offing.

-------

Personally, (and I know I have low credits just having started DC in March, 2006) I will support the projects that are conducive to my platform's capabilities and/or those that are non-profit, basic/diagnostic in nature (WCG: Fighting Cancer). I leave it to the protein structure folks to determine what is the right course for their project[s] and I will revisit (particularly Rosetta, because I believe that you all have made a great attempt at inclusion in support of your project.).

I know that much of what I have written above goes against the BOINC initial premise (non-profit), but at the same time the BOINC premise is being used for potentially for-profit activities. A clarification is in order.

Sorry for being long. Keep up the good work. Solve the problem!

I'll post this on Team MacNN forum.

Larry

[ritzl]

To clarify, I posted the previous on the Rosetta (good-bye) board.

---

To Zombie67 and Mikkyo, I don't have that much time to assimilate and comment, but I tried to condense my feelings on some of your previous posts in this reply.

You all (and others) are doing such great things by contributing to DC. Things I could never do, whether by time, aptitude, or expense. Please keep it up.

There IS real long-term benefit to these [all] DC projects. But it is exactly that. Long-term.

Protein structure, in particular has such profound implications. It is a pity that the Mac community is relegated to second tier in this pursuit.

---

I also don't even claim to know what the Mac v Windoze construct is in terms of worker development by project time, but I do know that it seems to be imbalanced, if only because no-one has (want's to allocate) the time to explore and develop a better way.

I'm sure I'm missing something in the global perspective, but these are my thoughts as a non-power user.

[jedimstr]

Originally Posted by zombie67

We'll take it!

Sorry about the bump in stats today. I got a bit OCD with jedi_mstr and just can't stop.

I just switched my clients to a 50/50 scheme from the previous 90-Rosetta/10-SETI. So you should be able to trounce me soundly in the Rosetta Stats. Of course the overall BOINC stats is another ballgame entirely (and wait until I finally get my Mac Pro... damn the 1900XT ship delay).

[zombie67]

Originally Posted by jedimstr

I just switched my clients to a 50/50 scheme from the previous 90-Rosetta/10-SETI. So you should be able to trounce me soundly in the Rosetta Stats. Of course the overall BOINC stats is another ballgame entirely (and wait until I finally get my Mac Pro... damn the 1900XT ship delay).

I'm getting there. I'm currently getting about 7200/day.

Looks like you are in the high 3k range with the new credit system and 90/10. Hard to tell. You vary a lot day to day. That will go up with increasing the SETI allocation and using the optomized application. The new MacPro is *sweet*! Which speed? That will add another ~2000 day at least.

Then again, I will be adding a Core2 Duo (E6600/2.4ghz) tomorrow.

I decided to wait until Clovertown (4-core) comes out, which may be as early as late Q4 now, and see what apple decides to do with it. Mmmmmmm....dual 4-core...

[jedimstr]

Originally Posted by zombie67

I'm getting there. I'm currently getting about 7200/day.

Looks like you are in the high 3k range with the new credit system and 90/10. Hard to tell. You vary a lot day to day. That will go up with increasing the SETI allocation and using the optomized application. The new MacPro is *sweet*! Which speed? That will add another ~2000 day at least.

Then again, I will be adding a Core2 Duo (E6600/2.4ghz) tomorrow.

I decided to wait until Clovertown (4-core) comes out, which may be as early as late Q4 now, and see what apple decides to do with it. Mmmmmmm....dual 4-core...

Here's the config I went with:
3.00 GHz Quad Xeon
4GB 667 DDR2 FB DIMM ECC-4x1GB
ATI Radeon X1900 XT 512MB (the cause of the delay...ugh)
500GB Serial ATA 3Gb/s drive (will add a RAID 0 pair once I find out which 500 they used)
16x SuperDrive DL
Airprt Extrm & BT 2.0+EDR
Apple Kybd & Wrls Mgty Ms
iWorks '06
30" Cinema HD Display (This is already here...in a box...waiting...double ugh)

Of course I need it for more than BOINC... my G4 PowerBook can barely keep up with Final Cut and Aperture (two of my most used apps).

[zombie67]

Originally Posted by jedimstr

Here's the config I went with:
3.00 GHz Quad Xeon
4GB 667 DDR2 FB DIMM ECC-4x1GB
ATI Radeon X1900 XT 512MB (the cause of the delay...ugh)
500GB Serial ATA 3Gb/s drive (will add a RAID 0 pair once I find out which 500 they used)
16x SuperDrive DL
Airprt Extrm & BT 2.0+EDR
Apple Kybd & Wrls Mgty Ms
iWorks '06
30" Cinema HD Display (This is already here...in a box...waiting...double ugh)

Of course I need it for more than BOINC... my G4 PowerBook can barely keep up with Final Cut and Aperture (two of my most used apps).

What an awesome machine. I feel your pain. How long is the delay?

The Mac Pro will be at the top RAC for all the projects very quickly. The only thing out there that can beat it are things like 8+ way machines.

I recently built a dual dual-core xeon 5060 3.2ghz. It is the previous generation to the version in the Mac Pro. Anyway, it supports HT, so it runs 8 jobs at once. Very cool. But the performance is a bit worse than I expected. It is getting only about 1600/day with this new credit system.

I've been looking into building a cruncher with the new xeon chips for about $1700 (2x5130(2gbz), and no graphics). Apples to Apples (heh), you really can't build one for cheaper than what you can buy from apple. And then you miss out on OSX! Great price...very competitive!

[jedimstr]

Originally Posted by zombie67

What an awesome machine. I feel your pain. How long is the delay?

The Mac Pro will be at the top RAC for all the projects very quickly. The only thing out there that can beat it are things like 8+ way machines.

...snip...

3-4 weeks from time of order. Check the 1900XT and Mac Pro delay threads in the PowerMac and MacPro forum for details. They recently dropped the price by $100 for the 1900XT video card and updated all delayed orders accordingly, but there's still a delay.

It says mine is shipping Sept. 19 but that's at the 4th week mark, so hoping I get mine next week instead (or even earlier).

[zombie67]

Good news! Apparently there were some expiration problems with RAC on Rosetta. Seems RAC only gets updated when people report in. So if they never report a job, their RAC never decays. They've run some script that updated everything. We're 7th (in RAC) now! Considering our overall credit place is 9th, that gives us the ability to move up 2 slots. Cool beans.

Spam, deleted.

[ChillyWilly5280]

Originally Posted by seababy

Spam, deleted.

OK, I don't know how or why, but I was able to edit seababy's post before I logged in. In this case it's no big loss (an ad for some site selling Photoshop w/o the box & manual), but obviously this is not as it should be. Have I been promoted to Administrator?

[reader50]

No, your user privs are normal. Can you edit this post?

I'll do some checking. The first thing that comes to mind is that regular members can edit guest posts, which isn't how it's supposed to be.

[reader50]

Update. Unregistered can edit unregistered posts. You must have been logged out when you did the edit? This has been fixed.

Good work on the spam, btw. The mark of a good moderator - nail the spam fast.

[jedimstr]

mmm... spam...

In other news...

Good News: I finally have my Mac Pro (mentioned above) fully setup and crunching away.
Bad News (at least for Rosetta): I was getting nasty nasty power per point return in Rosetta so I changed my 50/50 Rosetta/SETI split to 1/99. So effectively, I'm back to almost fulltime SETI with Rosetta taking over only when SETI decides to bring servers down for an extended period.

Which means that Zombie67 is effectively trouncing me in Rosetta stats, but overall Boincstats I'm making a killing in the team (except for two other people with their massive Distributed Computing Farms... Bernhard and Jamsoft... who I'll probably never be able to touch).

[reader50]

I recently did a 10-day run on Rosetta, for benchmarking purposes. Near the end, I started having workunits (or worker threads) sort-of crash. Apparently, BOINC was ready to switch away to another client. It would do so, but the previous Rosetta worker remained, wanting 100% CPU. Quitting and restarting BOINC solved it, and no work units were lost.

But this hang happened 4-5 times during the last day or so of benchmarking. Only on switching away from Rosetta workers, not when switching away from Einstein or SETI. Latest Rosetta worker, latest stock BOINC. Considering I was only a day away from finishing Rosetta benchmarking, I didn't investigate further. Just restarted BOINC as needed until all Rosetta units finished.