I find it a bit frustrating, especially when available work load are tight, that some users are allowed to download over 1,000 workunits! Let's be realistic and limit user to having at most 100 or so outstanding workunits.
I've got one I complete, but the other user who has the other part has 4048 work units outstanding. Ridiculous that anyone would need to have that many work units before connecting to the internet.

---

Do you have any clue what a modern computer can do in 24h?

Helli

---

A loooong time ago: First Credits after SETI@home Restart

looks like he has completed it as the two you have pending have wing men with only 2-300 outstanding

---

Tim

I find it a bit frustrating, especially when available work load are tight, that some users are allowed to download over 1,000 workunits! Let's be realistic and limit user to having at most 100 or so outstanding workunits.
I've got one I complete, but the other user who has the other part has 4048 work units outstanding. Ridiculous that anyone would need to have that many work units before connecting to the internet.

Welcome to the dilemma

Limit buy quantity is not the answer

You may have units to crunch for days with less than 10 units cached
Others would blow thru 10 in a few hours

The better limit is so many days worth
Each to what they are capable of using

Bill, If you are running other projects you still more than likely wouldnt get 100 work units.

With the 3 day outage coming up 100 work units would not last my i7 more than a day.

---

[/quote]

Old James

To add to what was said...Taken just now from my modest GPU host:
SETI@home Enhanced (anonymous platform, nvidia GPU)
Number of tasks completed 5171
Max tasks per day 1290
Number of tasks today 476
Consecutive valid tasks 1191
Average processing rate 137.26068194655
Average turnaround time 1.52 days

That is just for ONE application...Many of these guys run lots faster rigs and multiple applications.

What I can't figure is why a person with a rac of 3 will complain the loudest about being out of work....*scratches head*

---

"The most exciting phrase to hear in science, the one that heralds new discoveries, is not Eureka! (I found it!) but rather, 'hmm... that's funny...'" -- Isaac Asimov

There are about 1,000,000 registered SETI crunchers, or 2,000,000 BOINC crunchers.
If we consider the BOINC cruncher base, then:
I have a RAC of about 5,000, and I come in at about 13,000th position.
That means there are about 1,987,000 whose RAC is lower than mine, but only 12,999 whose RAC is higher than mine, indeed I'm in the upper 99.3% of users.
Which means the vast majority of users have a lower RAC than me.

So we have a dilemma.
Do we feed the top 1% and sacrifice the other 99%, or do we have a system that keeps everyone supplied with data.

The answer is, we keep everyone fed, in a manner that keeps the majority supplied with their daily single WU. Those at the top, and I include myself in that category, must expect to have a reduced supply of data from time to time, perhaps even long term - the lack of data won't kill us, but it will keep the other 99% from killing us by not being around and keeping BOINC, and SETI as a high profile mass participation HOBBY.

I say HOBBY because we are not paid to take part in BOINC, or SETI, its only out egos that get polished by having massive scores, indeed our bank balances take a bit of a hammering to get those scores.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

As this is a Seti situation, not a Boinc situation, the stats might be changed somewhat if only ACTIVE Seti crunchers and your RAC on Seti were considered in your evaluation. This a per-project decision, not Boinc-wide.

Although your position may still be supported, I don't think it will change my position at all.

There were some stats a while back about how much credit the top 100 Seti crunchers contribute compared to the rest of the crowd. I don't remember the exact numbers. But we are in the minority.
But it was also pointed out that as far as the servers and bandwidth is concerned, it is far more efficient to give 1 host 1000 WUs than to give 1000 hosts 1 WU each.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Seti and GPU grid have become projects for those who can throw the most money at the project, no longer projects for spare computer cycles.

---

Sniff. Sniff.

Ewww. What's that smell.

Politics. Ugh.

Seti and GPU grid have become projects for those who can throw the most money at the project, no longer projects for spare computer cycles.

There's some truth in that. One of the reasons I like Milktway@home is you're only allowed 6 work units per cpu at a time, but "Maximum daily WU quota per CPU 10000/day".

Everybody gets a crack at the work units and while the project is up you can
work on projects 24/7 (at least I can).

When it's down, well, there are other projects to crunch.

I crunch on 9 projects with 3~4 machines, but only half the projects seem to be
handing out work units at any one point in time.

Sniff. Sniff.

Ewww. What's that smell.

Politics. Ugh.

you raised a question so don't be sarcastic ...

in my case, each box does over 300 wus per day. if you set your cache to 10/10 then you should be able to get enough work to last for 20 days which should also make you immune from just about all of the ups and downs. this is how i run my laptop (dual core with hyperthreading and Nvidia GPU) and it has enough work for around 3 weeks.

in the case of the above boxes, they only get enough work for 2-3 days so they get caught up in the outage and the subsequent scramble for work when the system comes back up (even though i have the caches set deep). if the limits went, my caches would fill over the subsequent 4-6 weeks and i would then not be prone to the weekly cycle or unscheduled outage. it's the weekly cycle coupled with the limits that is causing the problem.

---

If you have a PC (and you have the correct settings made) you get WUs.
There are no members which are preferred.


My E7600 need 24 WUs/day the additional GTX260 need 160+ WUs/day.
My 940 BE need 48 WUs/day the 4 additional GTX260 need 640+ WUs/day.
(normal AR WUs)
If shorties much more. Then x4.

So if my E7600 PC have ~ 550 WUs in the list this is a 3 day WU cache.
If my 940 BE PC have ~ 2,100 WUs in the list this is also a 3 day WU cache.
(normal AR WUs)


EDIT: Don't forget the currently ghost WUs. The SETI@home server think your PC have DLed the WUs, but they are not DLed. So BOINC request more. So it's possible you have 50 % or more in the list listed than DLed on your PC.

---

Sniff. Sniff.

Ewww. What's that smell.

Politics. Ugh.

you raised a question so don't be sarcastic ...

(snippage)

Hint: There is more than one Bill/bill on these fora.
Check your attributions.

How would limiting a machine that can run 1000 work units pretty swiftly be unrealistic? That's just about as realistic as saying anyone who can't crunch that much shouldn't get any work because you're holding back the big crunchers!!

In an ideal world where the splitters/blankers are keeping up with demand and having plenty of spare WU's on hand along with a large enough pipe to feed them out everyone's systems would fill their cache's and everyone would survive the ups and downs of the system. We have only been back active for what two weeks? Give the boys a chance to tune the systems in, everyone get some work, caches start to fill, and everything balances out. Then this debate will be moot.

Over the few days that the splitters have had plenty to do the bandwidth was more a concern than the number of work units. And just as the bandwidth was easing off, 3 days in, the WU's ran dry. Give em time.

---

Traveling through space at ~67,000mph!

How would limiting a machine that can run 1000 work units pretty swiftly be unrealistic? That's just about as realistic as saying anyone who can't crunch that much shouldn't get any work because you're holding back the big crunchers!!

In an ideal world where the splitters/blankers are keeping up with demand and having plenty of spare WU's on hand along with a large enough pipe to feed them out everyone's systems would fill their cache's and everyone would survive the ups and downs of the system. We have only been back active for what two weeks? Give the boys a chance to tune the systems in, everyone get some work, caches start to fill, and everything balances out. Then this debate will be moot.

Over the few days that the splitters have had plenty to do the bandwidth was more a concern than the number of work units. And just as the bandwidth was easing off, 3 days in, the WU's ran dry. Give em time.

Fair enough....and hopefully true soon.
These religious arguments usually only come up when the project is having problems of one sort or another and somebody thinks their ox is being gored more than the next guy's.

Hopefully the data to split can be blanked fast enough to keep things flowing and all shall be happiness and light around here again...LOL.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Fair enough....and hopefully true soon.
These religious arguments usually only come up when the project is having problems of one sort or another and somebody thinks their ox is being gored more than the next guy's.

Hopefully the data to split can be blanked fast enough to keep things flowing and all shall be happiness and light around here again...LOL.

I hope you're right because everyone has to varying degrees been after each other's throats the past month or so and it's getting worse I believe. So keep us out of the light and don't feed us after midnight or bad things happen. ;)

---

Traveling through space at ~67,000mph!

How would limiting a machine that can run 1000 work units pretty swiftly be unrealistic? That's just about as realistic as saying anyone who can't crunch that much shouldn't get any work because you're holding back the big crunchers!!

In an ideal world where the splitters/blankers are keeping up with demand and having plenty of spare WU's on hand along with a large enough pipe to feed them out everyone's systems would fill their cache's and everyone would survive the ups and downs of the system. We have only been back active for what two weeks? Give the boys a chance to tune the systems in, everyone get some work, caches start to fill, and everything balances out. Then this debate will be moot.

Over the few days that the splitters have had plenty to do the bandwidth was more a concern than the number of work units. And just as the bandwidth was easing off, 3 days in, the WU's ran dry. Give em time.

Fair enough....and hopefully true soon.
These religious arguments usually only come up when the project is having problems of one sort or another and somebody thinks their ox is being gored more than the next guy's.

Hopefully the data to split can be blanked fast enough to keep things flowing and all shall be happiness and light around here again...LOL.

It would be nice if we could work towards that "ideal world" in these discussions. Unfortunately it doesn't take long before somebody trots out their gored ox and the it's "You think your ox is gored? Take a look at my ox! Now that's gored!".

If people could remember that the new hardware is still in the "tweak" phase and have a little patience we could get a better picture of where to apply the next improvement.

I'd like to see a moratorium on gored oxen until February.

Hey on this machine Milkyway report 58 hours runtime when it downloads, actuals come in at 32 hours approx, so 6 WUs limit is not bad

Dual Intel CPU 2.3GHz no Nvidia card

---

Yes but lots of folks appear to grab anywhere from 100 to 500 WUs and SIT on them taking 4 weeks or longer to process them

I for one, have no problem with those that have the resources to process 1000 WUs a day, accurately and turn them back in 24 hours after receiving.

There are far too many who download 100 or more and take forever to return them, that is my only complaint

If you have hardware that can process 1000 per day, GREAT. Just so long as you play fair with the rest of us, and let us get a few of the WUs in the pool.

To save bandwidth, I do not refresh my connection to the server every 2 minutes looking for more WUs

---