I agree I only have 256Mb and am now 91% through 3-4 hrs to go and I will be uploading the WU

Do you have your settings to leave the app in memory or to swap it out?

-- Skip

I had Rosetta running on my 3 'normal' machines that have 512MB or better for a couple days w/o problems so I added rosetta to the slowest of my dedicated crunchers (XP 2000 OC'd to 1.99GHz). This 'basket cruncher' has a single 256MB stick of PC3200 and let it run for 48hours (WinXP Pro SP2). For the first few hours I suspend everything except Rosetta and the 1st two W/Us averaged out at about 2 hours to finish. Haven't seen any problems and so have just added Rosetta to the other 256MB XP based 'basket crunchers'.

If I get bored later this week I might try running it on one of the old 128MB boxes just to see what it does.

[SOAP BOX MODE ON]
256MB is overkill for the majority of the production BOINC based projects. I have two old slow machines w/ 128MB and they run 4 or 5 of the major projects (no CPDN on them, no E@H on the other but not due to memory). The PIII 550 w/ 128MB runs 5 BOINC project and DIMES! Not fast, but runs fine. So these 1G and 2G recommendations have nothing to do with BOINC (and yes, I have a GIG of sweet OCZ platinum 2-3-2-10 dual channel in my desktop).

The point I'm driving at here is IF / WHEN Rosetta really requires 512MB they will lose some hosts. Would it be a significant number? I think so. I'm sure I'm not the only one who will not go buy 'X' sticks of memory for 'the farm' just to run Rosetta on the crunchers. Roughly, you have to get past the 1st page of ranked participant to find one that's running less than 6 machines. It'd be interesting to run some sort of query to see how many hosts have under 499MB despite this projects recommendation of 512MB and what their credits total to for an idea of what the drop off for the project would be.

If you run BOINC as a service on a service/application stripped cruncher w/o graphics I'm sure I don't understand the need for more than 256MB of physical memory on a windows box and probably less on a Linux box.
[SOAP BOX MODE OFF]

Oh well, I started this out just to say, at least today, all is running fine on 256MB boxes.

Comments welcome, flames ignored. -- Skip

---

- da shu @ HeliOS,
"A child's exposure to technology should never be predicated on an ability to afford it."

512MB is recommended because we plan on doing tests for larger proteins and also plan to include protein design and docking tests, which could take more memory. It is not a requirement, just recommended.

Seems to me that the 512 recommendation is for WUs to come and not for WUs we are processing now. I have looked at my task manager and with four Rosetta's running I use a lot of memory at times. My 3 gig is over kill but I am ready to run the bigger and badder WUs

---

BOINC Wiki

The point I'm driving at here is IF / WHEN Rosetta really requires 512MB they will lose some hosts. Would it be a significant number? I think so. I'm sure I'm not the only one who will not go buy 'X' sticks of memory for 'the farm' just to run Rosetta on the crunchers. Roughly, you have to get past the 1st page of ranked participant to find one that's running less than 6 machines. It'd be interesting to run some sort of query to see how many hosts have under 499MB despite this projects recommendation of 512MB and what their credits total to for an idea of what the drop off for the project would be.

If you run BOINC as a service on a service/application stripped cruncher w/o graphics I'm sure I don't understand the need for more than 256MB of physical memory on a windows box and probably less on a Linux box.
[SOAP BOX MODE OFF]

It depends what they're doing in the WU, and the size of the WU itself. Likely, as of this beta phase the code isn't exactly optomized. Optomizing the code isn't first priority in development, and bug fixing tends to come first. When all is running well, then devs might look at ways to improve the efficiency of the code. However this can also be offset by the adding of more features, or in the case of this project more work being done to the unit.

In any case, CPDN (as I remember) and Einstein@home have been pointed out in times past as I remember when someone says that some computers will be lost. Those projects do run, and also set their requirements/recommendations to what they need/are looking for. Do they lose hosts in that CPDN takes so long to process, and Einstein also uses a lot more CPU then many other projects? Probably. But the argument has gone that it is for the users to determine what they can run with each box.

I for instance have a couple removable HDs available to me for this Linux apps class I'm currently taking. (Setting up a honey pot, or actually tar pit for the current project. And then a box to do, probably, best left unmentioned things to test the honey pot with, with a few routers to serve as firewalls to isolate it...) For now I installed BOINC to get some extra crunch time. But given they only run while I have class or am working on them; there are some projects I definitely would not set them up on. Main reason, they only have about 5 hours a week to crunch...

FYI, if one wants to do work with protein and one's box doesn't have enough memory; there is still predictor@home. Peeps on that board were pointing out that the project there is ideal for older computers. As of late the WUs we're getting now have been taking less time to crunch (about half of what they were) and the memory used is about 12,000 Kbyte according to task manager.

---

I've got a small display on my wall next to my workstation -- on it I've mounted a fan of UNIVAC punch cards, a small reel of 1 inch tape (from IBM 360 days), an 8-1/2 inch floppy disk, a 3-1/2 inch diskette and a 3 inch CD disk. Wonder what will be next ...

(edited for content)

hehe, believe it or not, silicon based semidonuctors could be on the way out. The problem is essentially this. We've been operating under the premise of Moore's law for sometime, which assumes that both clock rate and chip complexity doubles every 18 months. Course we're up in the several GHz level with CPU clock, and these chips are already rather complex.

With this comes heat, and lots of it. Another (though not so long ago) "I remember when peeps refered to the Pentium II 300 MHz Klammath as an Intel space heater". But it had nothing on either the Pentium 4 or the Athlon 64 in heat generation. Funny thing, in my apartment I rarely need heat to keep it warm in the winter. My computer can keep my apartment warm until the outside temp drops below 45 degrees with my current AXP 1900+. When I get the A64, unknown. On the other hand, it's October and I still need AC. However I do live in New Mexico.

Heatsinks have also gotten a lot better but are also pushing things. The Swiftech MCX6400-V heatsink I got for my Athlon 64 weighs in at 750 grams, and pretty much all A64's I gather use a backing plate to help hold the HSF on there. The fan is of course more. Then there's water cooling, things like Kryotech (or it's equivalent now adays). I remember when CPUs didn't need fans :D

On the other hand, die shrinkages have been a main way to combat heat generation, with a .09 micron manufacturing process out there currently. My A64 (using the Venice core) also uses things like SOI (silicon on insulator), AMD's equivalent to strained silicon, and the like. But in the end, all the stretching of things hits the limitations of physics. And one of these is that the silicon atom is a certain size. One just won't be shrinking the die size to sub-atomic sizes without going to the sub-atomic (aka quantum computers).

As such, people are already looking for a replacement, be it optical computers or whatever else, and also at replacing Si with carbon (as the carbon atom is smaller, but carbon and silicon also share many of the same properties). In 10-20 years, computers could be quite different.

---

If you run BOINC as a service on a service/application stripped cruncher w/o graphics I'm sure I don't understand the need for more than 256MB of physical memory on a windows box and probably less on a Linux box.

You may not. However, you trade processing time for swaping time. Efficiency goes way down. But, that is a choice only you can make.

I recall the first time they added processing to SETI@Home... I was taking 30+ hours per work unit at the time ... now we have much more processing and still the run time has fallen to 2-5 hours on the "average" system out there... and again we are looking at doubling the sensitivity at the expense of trippling the processing time for a work unit.

So, here are the choices... they recommend x, you have y, will it work? Probably ... but it may not work that well and that is their point ...

Not quite 2 years ago I got a top of the line PowerMac ... now it is near the bottom (I think they still sell the 1.8s, if not, it is the bottom G5 now) ... again, the point being, as time moves on the speed/throughput rises ...

Most of the systems I see being advertised are 256M min and more common 512 with some now standard at 1 G ... is all that RAM needed? Beats me, I been putting 1G in for a long time now ... why? heck, why not, RAM is cheap ... and I get more cobblestones ...

Any way ... it is not that big of a thing ...

---

If you run BOINC as a service on a service/application stripped cruncher w/o graphics I'm sure I don't understand the need for more than 256MB of physical memory on a windows box and probably less on a Linux box.

running as a service does not really affect (much) memory usage. so if you have to hold data in ram, you will have to hold it no matter if it runs as service or as stand alone app.

---

--
Florian
www.domplatz1.de

OK, I didn't check out the RAM requirements before I signed up for Rosetta. I turn out to have 256MB RAM. I now paused the current Rosetta WU, and let my computer work on other Boinck projects. I'm interested in Rosetta very much, so I would like to contribute to this project. Is that possible with my computer?
This is what I have:
AuthenticAMD, AMD Athlon(tm) XP 1900+
Microsoft Windows XP Home Edition, Service Pack 1
1,60 GHz, 256 MB RAM
My harddisk has more than enough free diskspace, so that's no problem.
Can I still run Rosetta? (I noticed that the WU I downloaded still was sitting at 1% after running for an hour. I know that this problem is already addressed in another thread, but could it have something to do with my lack of RAM?)
Thanks in advance!

---

This is what I have:
AuthenticAMD, AMD Athlon(tm) XP 1900+
Microsoft Windows XP Home Edition, Service Pack 1
1,60 GHz, 256 MB RAM
My harddisk has more than enough free diskspace, so that's no problem.
Can I still run Rosetta?

Christoffel,

Welcome! While 512 MB of RAM is the recommended minimum to participate, you will still be able to do so with 256 MB. The disadvantage will occur when they start moving towards complex protein structures that will require more memory than what we are currently crunching. Your system will then start using virtual memory from your hard disk. It will slow things down for you, but you should still be able to run ok.

For the 1% bug, you should find your system moving past 1% onto 8.33% before an hour. If it doesn't, suspend the project and then immediately resume it and you should be back in business.

---

I didn't check the demands on the computers neither as I assumed my computer would be able to deal with almost anything!

But I tried Predictor, and it interfered with my Danish - English - Danish electronic dictionary, a program, I bought many years ago, and now have the latest version of. My dictionary became unstable, as it stopped working after a while, telling me that there were not enough system ressources for it, and I had to reboot everytime! :-( The problem came with Predictor, was there while running Predictor and disappeared, when I detached from Predictor, so as I need my dictionary more than they need my computer, I left Predictor. So not all can run Predictor without any problems!

But if the day comes, where my computer won't be able to keep up with the demands here at Rosetta, I'll ponder if I'll go out and buy more RAM. Yes, I will, if I'll need it in general, no, if it's only for Rosetta!

---

[b]"I'm trying to maintain a shred of dignity in this world." - Me[/b]

[quote]My laptop only has 256MB. It works, but work units take a relatively long time for the processor's speed (average 2 3/4 hours on a 2.8GHz Pentium 4). Have been looking to add extra RAM anyway (not just for BOINC/Rosetta), so I've ordered it now. Maybe it will make a difference.

Have you had any heat problems running you P4 laptop? Do you run it 24/7 or does it get a rest ?

I'm in the market for a new laptop and would appreciate any reccomendations regarding the P4 vs a centrino. Is the centrino adequate for DC projects or do i go for the horsepower?

Thanks

DP

---

Well now!
I have an IBM Think Centre 2.8GB Desk Top just about 1 year old. Windows XP Pro 2 with 256 Ram and have set Virtual Ram at 750mb. Average time for a WU is about 2 1/2 hrs, range is between 2hr 02min and 4hr 40m. That is over the last 50 WUs. There are some 20 GB unused so no problem using Virtual Ram, or so it seems. Not so sure I see the need for the latest uptodate super super high memory and all the gizmos. PC runs more or less 24/7 on Rosetta. No WUs have exited early since Beta testing ended.

---

For the 1% bug, you should find your system moving past 1% onto 8.33% before an hour. If it doesn't, suspend the project and then immediately resume it and you should be back in business.
=========

I've never had a WU yet that was Stuck advance any further by just Suspending the Project. The only way I've been able to sometimes get them to advance is to shut the BOINC Program down & then Re-Start it again. This works sometimes but not always.

These new 1acf_ Type WU's seem worse than the previous ones, almost every morning I get up I have to abort 2 or 3 of them & Re-Start the Manager on a few others. And it seems they can get Stuck @ any % too, 8.33% & 75% seems to be the going trend now for them to get Stuck, at least on my PC's anyway.

I've had to abort 3 WU's this morning already totaling 20 Hour's CPU Time for one's stuck @ either 8.33% or 75%. And I'm watching 3 other WU's that I'm pretty sure are Stuck @ the same % ... I'll let them run for awhile yet and see if they advance.

@ Fuzzy HollyNoodles

Off topic, but I'm curious, what disctionary software are you using that Predictor screwed up?

Med venlig hilsen,
Adrian...

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

For the 1% bug, you should find your system moving past 1% onto 8.33% before an hour. If it doesn't, suspend the project and then immediately resume it and you should be back in business.
=========

I've never had a WU yet that was Stuck advance any further by just Suspending the Project. The only way I've been able to sometimes get them to advance is to shut the BOINC Program down & then Re-Start it again. This works sometimes but not always.

These new 1acf_ Type WU's seem worse than the previous ones, almost every morning I get up I have to abort 2 or 3 of them & Re-Start the Manager on a few others. And it seems they can get Stuck @ any % too, 8.33% & 75% seems to be the going trend now for them to get Stuck, at least on my PC's anyway.

I've had to abort 3 WU's this morning already totaling 20 Hour's CPU Time for one's stuck @ either 8.33% or 75%. And I'm watching 3 other WU's that I'm pretty sure are Stuck @ the same % ... I'll let them run for awhile yet and see if they advance.

I do not pay much attention to percent completed on this project, because this statistic does not seem to mean much.

I guess I am very fortunate, I have been crunching on this project with 2 dedicated machines and have completed a total of 174 units to-date. I have not had any problems whatsoever with any workunit. :)

---

Regards,
Bob P.

I do not pay much attention to percent completed on this project, because this statistic does not seem to mean much.

I guess I am very fortunate, I have been crunching on this project with 2 dedicated machines and have completed a total of 174 units to-date. I have not had any problems whatsoever with any workunit. :)
____________
Regards,
Bob P.
=========

It's not that I want to or even like to have to do that Bob, but I'm almost forced to if I want to run this Project. If I didn't constantly monitor the WU's it would only be a matter of a few days and every one of my Copmuters would be doing nothing but running WU's that were Stuck at 1 Percentage or another.

I'm going to drop the Project here shortly because it's just way to much Time Consuming for me to run this Project if I want it to run smoothly. I'll keep monitoring the Boards and if things seem to be running better I'll try it some more later on ...

I do not pay much attention to percent completed on this project, because this statistic does not seem to mean much.

I guess I am very fortunate, I have been crunching on this project with 2 dedicated machines and have completed a total of 174 units to-date. I have not had any problems whatsoever with any workunit. :)
____________
Regards,
Bob P.
=========

It's not that I want to or even like to have to do that Bob, but I'm almost forced to if I want to run this Project. If I didn't constantly monitor the WU's it would only be a matter of a few days and every one of my Copmuters would be doing nothing but running WU's that were Stuck at 1 Percentage or another.

Sorry to hear that...each system is different as I (should) know. Stupid question, assume you have the 512mb of RAM specified for this project? Also, I am running Rosetta solo, which may help it overcome its crankiness....

Hope things work out for you longer term! :)

---

Regards,
Bob P.

Stupid question, assume you have the 512mb of RAM specified for this project?
=========

All my Computers have at least 1Gb of Ram and I even increased the Virtual Memory to try & see if that helped out. Funny thing is if I could only get those newer 1bmbA_ Type WU's I've been getting a few of lately I wouldn't have any problems. I haven't had a 1 of those hang yet, it's those 1acf_ & 1cfyA that are a problem for me from time to time.

They run in streaks it seems like, I can go a few days & only have 1 or 2 that hang & then I start to get them in bunch's across my Computers. I'm wondering if it might be a good thing to just reboot the Computers once a day because of all the Memory the WU's use up ... ???

I also run Rosetta by itself, I Suspend all the other Projects and run for part of the day, then I Suspend Rosetta and run the other Projects.

I'm going to drop the Project here shortly because it's just way to much Time Consuming for me to run this Project if I want it to run smoothly. I'll keep monitoring the Boards and if things seem to be running better I'll try it some more later on ...[/quote]
>It would be a shame if you pulled the plug, Bob......the AMD USER team would never recover! I have the same frustrations as you and the project people don't seem to acknowledge that they are doing anything about it. My experience with this 1% nightmare parallels yours....I have reduced my 'connect to project time' to .2 days so I can better monitor things and it's a pain. I never had problems with any WU's until they upgraded their servers to accomodate Boinc 5.2.1. After that, I was getting about 1 or so a day on 12 boxes(all with 512M RAM). It may be a coincidence but it makes one wonder. At the time I was using Boinc 4.19 on my boxes so I upgraded to 5.2.1 to see if that helped and it did. I have also increased our switching time to 120 minutes between 2 projects. I have had only one WU stuck at 1% in about three days. Since we both are on the AMD USER team it begs the question if AMD CPU's are more likely to have this problem for some obscure reason?? Seems unlikely, but in the weird digital world who knows? Bye the bye, so that you can see any comman factor between our units and yours, we are using (mostly) 1.8G Durons with WinXP. Hope this helps.... please don't leave us yet!....Cheers, Rog.

---

Anyone else seeing this stuck at 1% bug?

Anyone else seeing this stuck at 1% bug?

Hi David. Thanks for looking at this problem. Maybe your question needs a new thread?? Just a suggestion. There has been some discussion in other related threads about this 1% bug so others have seen it. We need a 1% solution as it were......with you playing the part of Sherlock Holmes of course! Cheers, Rog.

---