LHC@home

I must say, I joined Rosetta, and the first six or seven units failed. I read up and realized a reboot was a necessity. I detached, (upon a reboot), I will reattach. I have yet to think that the scientific gains could outweigh the minor problems with the project... Granted I haven't rebooted in nearly a month, I will not heed this warning. I think, after seeing a discussion on C-SPAN, this project is incredibly worth while and will only improve in the future. It seems people in this thread have been suspect of the potential for several projects, including LHC, I don't share this negativism. It's interesting no-one has mentioned malaria control.net; they are great in my opinion.

Well Malaria is a closed beta at present, so wouldn't be a viable alternative for most, however there are other projects. As to Rosseta, from current indications in RALPH (rosseta alpha project), the remove from memory bug seems to have been worked out from what I'm gathering. Also the % of failed WUs has been reduced, about halved I'm gathering... No notice when the app will make it out of alpha test and into the rosseta project itself. There could be some more app versions that might come out first as well, as rosseta itself is still on 4.82 or so, not 4.93...

Speaking of which, I should probably stop by the RALPH boards again, and see what the lattest is from the posters, well beyond the news item updates...

Actually, thx for reminding me of this, just checked their site and new work is flowing again on the alpha... Re-enabled WU d/l for now... Might know more about where it's heading and where things now stand in a bit...

Edit: Actually it's app version beta 4.97 now, so it might be a little while still before we have the improved app out for you guys

---

And peterthomas, may I suggest ... Honestly, sign up at CPDN or some other such project if you're that worried about running out of work :D

for your info peterthomas is an active cruncher on Einstein - I wonder where he will go when Einstein run out of work on S4 - they are likely to have a gap before s5 is available.

I guess (like many crunchers) doesn't like to think of his client trying valiantly for WU that are not there, even if his boxes do have other work to keep them going.

It is not my viewpoint - 2 of my boxes are still asking orbit@home for work - but it takes all kinds to support a DC project.

River~~

It is not my viewpoint - 2 of my boxes are still asking orbit@home for work - but it takes all kinds to support a DC project.

River~~[/quote]

I don't think that is just DC, I think that is with anything that involves more than one person......

---

I actually spent the past 30 minutes reading this thread ... and it has meandered through so many topics I had to refresh b4 posting!!! (this is my first post here, so bear with me ...)

I am in hopes that peterthomas got the point, and that many others with that initial point of view realized also that DC and BOINC are THE means to our future. Many have probably joined multiple projects half-way through this thread!

I started with SETI back in '99, and have since joined 9 other projects through BOINC Synergy. I join the projects I'm interested in, to include LHC. Most people I speak with don't understand my passion, or explanation. I have, however, gotten at least 6 others involved with projects. I'm an IT guy, and my satisfaction is from using computers for something with an actual purpose. Spending 8 hours a day 'teaching' people how to cut/paste, add a printer, or describing what a virus or spyware is drives me nuts. So, I spend my leisure time (yeah right!) reading articles like this ....... and checking my stats. This thread has been not only informative, but should be in a disclaimer that everyone has to acknowledge before joining a project!!!!!!

Happy crunching, and btw I have SETI first, Einstein second, and LHC third in my priorities!! Yes, I'm happy and this is an awesome project, regardless of what each person contributes!

---

Some IT expert!

And peterthomas, may I suggest ... Honestly, sign up at CPDN or some other such project if you're that worried about running out of work :D

for your info peterthomas is an active cruncher on Einstein - I wonder where he will go when Einstein run out of work on S4 - they are likely to have a gap before s5 is available.

I guess (like many crunchers) doesn't like to think of his client trying valiantly for WU that are not there, even if his boxes do have other work to keep them going.

It is not my viewpoint - 2 of my boxes are still asking orbit@home for work - but it takes all kinds to support a DC project.

River~~

Yeah, but since about 4.45 or somewhere about there (OK, not including the beta versions of the CC), BOINC supported suspending a project. If it does bother one, or perhaps not bother one but seems silly to keep contacting the server when one knows there isn't work, it's only like 2 mouse clicks to suspend the project.

The feature set in the newer BOINC CCs already supports the ability for the user to remedy this. Once one knows there's work available, one can re-enable it. In fact, this is exactly what I ended up doing. Observing the pattern in the past, I noticed that we were running close to the end of WUs the project was looking for, so figured new work might be comming. I was a bit late in noticing it, because when I got done today BOINC wasn't exactly the first thing on my mind.

Lets just say I had a rotting wisdom tooth, and it caused some gum problems. So the dentist put me on anti-biotics on Monday, and I was saw the oral surgeon today (actually yesterday now). Still only about 12 hours since they surgically cut and drilled the tooth out, so understandably I had other things on my mind getting home... But when I checked back and saw "oh there are 93,000+ WUs", I just re-enabled the project... Actually I'm running little other then LHC now, now that work is available, but now I'll return to my other WUs in the not to distant future. Just don't see an issue with any of this personally. It's the nature of the beast, and what the project needs :cool:

---

But now we have new work here at LHC, so chill down and enjoy the crunching!

---

Yeah, but since about 4.45 or somewhere about there (OK, not including the beta versions of the CC), BOINC supported suspending a project. If it does bother one, or perhaps not bother one but seems silly to keep contacting the server when one knows there isn't work, it's only like 2 mouse clicks to suspend the project.

Well actually, even 4.19 supported suspending projects.
The new thing about 4.43/4.45 compared to 4.25 (only taking the official releases) was the "no new work button" to flush the work buffers.

Some IT expert!

...

Thanks for your contribution to the thread Mike, just sharing the love?! btw - I fixed my signature, so thanks!

Some IT expert!

...

Thanks for your contribution to the thread Mike, just sharing the love?! btw - I fixed my signature, so thanks!

Yeah, then like an idiot I left of my own sig! Ha!

The post bearly made sense :(

---

Click here to join the #1 Aussie Alliance on LHC.

Actually, I found 4 other projects that I had to update/fix, so a little embarrassment here saved me some in a bunch of other forums! Now all my projects are up to date, no harm done - AND I finally got some LHC wu's!! (yes, I've been slacking and/or have missed out when there was work avail ...)

Looks like we have a good lot of workload, so we'll be busy for a little.

---

Yeah, but since about 4.45 or somewhere about there (OK, not including the beta versions of the CC), BOINC supported suspending a project. If it does bother one, or perhaps not bother one but seems silly to keep contacting the server when one knows there isn't work, it's only like 2 mouse clicks to suspend the project.

Well actually, even 4.19 supported suspending projects.
The new thing about 4.43/4.45 compared to 4.25 (only taking the official releases) was the "no new work button" to flush the work buffers.

Yeah, it might have gotten back ported to 4.19, which as I remember was an interum release for some projects. After 4.13 I was using the beta CC, not the official release, so might not have looked at 4.19 in the specific...

In fact I seem to remember mention of some things having gotten back ported, including retry on downloads or something, that helped, don't remember if it was Rossetta (which I had when it was in beta), predictor, or whatever. Was definitely a while ago however that peeps had this...

---

...[snip]...
Looks like we have a good lot of workload, so we'll be busy for a little.

I have temporarily shutdown all my other projects while LHC has WU's ;-). All I have to do now is check each day to make sure the suspended ones are not getting too close to deadlines!

---

Click here to join the #1 Aussie Alliance on LHC.

Yeah, I've largely done the same, with the exception of RALPH which had WUs for a few and needed some testing... There's like 10 days before I hit the first deadlines and considering we're now < 40k available units I don't think that'll be an issue...

---

Yeah, I have about 10 days to my next dealine too.

I just have to remember to look ;-)

---

Click here to join the #1 Aussie Alliance on LHC.

Just a note for those that suspend lhc when it has no work available: I believe that boinc does not modify the debt figures for suspended projects, so you may find that you crunch less lhc work than you would by allowing your host to poll the project (which really wastes very little cpu/network)

You also may find that occasionally you get other peoples errored or timed out work if your host is lucky while there is supposedly no work.

---

Just a note for those that suspend lhc when it has no work available: I believe that boinc does not modify the debt figures for suspended projects, so you may find that you crunch less lhc work than you would by allowing your host to poll the project (which really wastes very little cpu/network)

You also may find that occasionally you get other peoples errored or timed out work if your host is lucky while there is supposedly no work.

I was wondering about that. What if I set it to no new work?

---

Click here to join the #1 Aussie Alliance on LHC.

Just a note for those that suspend lhc when it has no work available: I believe that boinc does not modify the debt figures for suspended projects, so you may find that you crunch less lhc work than you would by allowing your host to poll the project (which really wastes very little cpu/network)
...

Correct, boinc does not modify debt for suspended projects. Whether this matters depends on whether you went into EDF mode while running LHC.

In short, if you have a small cache there is no issue here, if you have a large cache then it does matter. People who left the "connect interval" at the BOINC install settings can suspend and resume as much as they like without causing themselves problems -- apart from remembering to resume of course ;-)

Explanation

If last time round you downloaded enough LHC work to go into earliest deadline first (EDF) mode, then you will be left with a negative debt which will inhibit the fetching of new LHC work till that debt is paid off. For this to happen you probably have a connect interval of 4 days or more.

In that case, by suspending the project you prevent that debt being paid off. This means you will experience a frustrating delay in downloading new LHC work when you do resume.

If you leave the project runnable then the debt will get paid off at the rate set by your resource shares. Providing LHC has about 50% of the share then the debt should be clear before the next work becomes available.

Once the debt is clear, the client will not go past zero in the absence of work. This means that if you never go into EDF there is no problem suspending work. And if you still have a cache set to 0.1 days you won't get into EDF and you won't get projects building up negative long term debt.

---

...

I was wondering about that. What if I set it to no new work?

ditto, I think.

The long term debt is incremented only when the client is about to download new work and holds back *solely* because of the debt.

But on the bright side, if you keep your cache size below around 3.5 days this should never be an issue for you.

I just let my hosts poll from time to time. If you have enough projects that something else goes into EDF mode then the debt may increase some for LHC.

---

BOINC WIKI

BOINCing since 2002/12/8