Hello Tom,
Things are blocking up again after a couple of days of it working quite well.
Not able to get work and validation pending on the rise.

Maybe consider a fund raiser to get things that are needed so that the project doesn't need to be such "low priority".

Hello Tom,
Things are blocking up again after a couple of days of it working quite well.
Not able to get work and validation pending on the rise.

Maybe consider a fund raiser to get things that are needed so that the project doesn't need to be such "low priority".

+1

I've just got hold of a couple of R9 Nano GPUs (for an absurdly low price in an auction), which are twice as fast at single and half as fast at double precision as my 280Xs. I intend to leave the six 280Xs on Milkyway continuously as double seems to be getting so rare. I will use very rude words if they aren't kept busy.

Hello Tom,
Things are blocking up again after a couple of days of it working quite well.
Not able to get work and validation pending on the rise.

Maybe consider a fund raiser to get things that are needed so that the project doesn't need to be such "low priority".

When you say "thinks are blocking up again," what do you mean? Things all appear fine on my end.

Hello Tom,
Things are blocking up again after a couple of days of it working quite well.
Not able to get work and validation pending on the rise.

Maybe consider a fund raiser to get things that are needed so that the project doesn't need to be such "low priority".

When you say "thinks are blocking up again," what do you mean? Things all appear fine on my end.

I'm getting "no new tasks" when requesting GPU work, as of 6:44pm Zulu (or possibly earlier, I wasn't asking then).

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

At any rate, it looks like we are sitting around ~10k unsent jobs for separation at any given time now, which is the expected behavior.

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

Yes it would help greatly, since I have machines with up to 4 cards. I thought it was set to 300 per card, 900 per host, but at any rate, please turn both of those up.

Every time I check server status it always says around 10,000, so I'm not sure how it's ever running out. I guess you could try increasing that since it only takes a dozen people to empty it.

I just made another request and got 79. Earlier in the day I was getting about 250.

Of course everything would be wonderful if we could fetch and receive on the same communication, but I'm guessing nobody ever worked out what was wrong there?

Things have been running smoothly for a few days, so I don't know what changed, but I'm not getting separation tasks even though the status page says there are plenty.

hopefully it is just a lag as the system does something else (generate work? backup?) and it will hand out tasks again.

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

Still not getting any work.

Can it be set so that you simply replenish the units you report after getting the initial batch, which can be set lower?

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

Still not getting any work.

Can it be set so that you simply replenish the units you report after getting the initial batch, which can be set lower?

There's some kind of bug nobody's ever worked out where you can't send and receive seperation tasks at once. Since a fast GPU will complete a task in under the minimum server contact time, there will always be some to report and you'll never get tasks until you run out. And Boinc doesn't try again for 10 minutes as it assumes there are none available.

Seems like the feeder is sleeping too long to feed the buffer. I would suggest reducing the amount of time the feeder sleeps for.


Either that or someone has just dumped a ton of tasks onto the server, and its working through validating all of them

Seems like the feeder is sleeping too long to feed the buffer. I would suggest reducing the amount of time the feeder sleeps for.

Either that or someone has just dumped a ton of tasks onto the server, and its working through validating all of them

The Boinc server software needs a complete rewrite. Looks like at the moment you have to fiddle with lots of numbers to make it work right. Kinda like a machine from the 50s. Please use your expertise and go to github and kick those "programmers" where the sun doesn't shine.

Seems like the feeder is sleeping too long to feed the buffer. I would suggest reducing the amount of time the feeder sleeps for.

Either that or someone has just dumped a ton of tasks onto the server, and its working through validating all of them

The Boinc server software needs a complete rewrite. Looks like at the moment you have to fiddle with lots of numbers to make it work right. Kinda like a machine from the 50s. Please use your expertise and go to github and kick those "programmers" where the sun doesn't shine.

And why would I want to do that? Every one of the BOINC devs are volunteers since they've lost NSF funding, so all they are doing is maintaining the repo.

For the GPU thing, I think one of the previous server admins made a change and Tom can't find it.

Seems like the feeder is sleeping too long to feed the buffer. I would suggest reducing the amount of time the feeder sleeps for.

Either that or someone has just dumped a ton of tasks onto the server, and its working through validating all of them

The Boinc server software needs a complete rewrite. Looks like at the moment you have to fiddle with lots of numbers to make it work right. Kinda like a machine from the 50s. Please use your expertise and go to github and kick those "programmers" where the sun doesn't shine.

And why would I want to do that? Every one of the BOINC devs are volunteers since they've lost NSF funding, so all they are doing is maintaining the repo.

Then help them out. You seem like someone who knows their way around the software. It's ridiculous the server doesn't just work.

For the GPU thing, I think one of the previous server admins made a change and Tom can't find it.

Assuming you mean send and receive at once, I asked Eric and then Tom, they both tried, but there's some "misconfiguration" that nobody can find a fix for. I've mentioned it to several people on github in Boinc but nobody knows.

I'm getting tasks again, so it must have been like stated... an influx of completed WUs that needed validation.

edit: WCG friends... hopefully we will be back at WCG on Friday April 22.

I'm getting tasks again, so it must have been like stated... an influx of completed WUs that needed validation.

I thought that, one computer grabbed a few hundred. But then I told the other machines to check and they got nothing.

edit: WCG friends... hopefully we will be back at WCG on Friday April 22.

Cool! I couldn't find any info on that, other than "ready for final testing" a week or so ago. I have one phone checking every day for it, unfortunately the rest of my machines need to be re-attached since I changed from pool to private Gridcoin mining.

I wonder if the feeder is just running out of tasks at this point because people are requesting lots of work. The number of WUs in the pool is set to 10k jobs, but I can double that and see if it improves things.

We also have the max WUs to send at one time limit set at 600, but it seems like people crunch through that fairly quickly. Would changing that to 1000 help you, so you don't have to request work as often?

Yes! please do! Just installed an R9 280X last night, and it has run dry several times already!

Just installed an R9 280X last night, and it has run dry several times already!

Those cards are wonderful, I have 6. Not sure what MW is going to do when they've all expired and everyone uses the modern pitiful ones without much DP.

deleted duplicate post

Just installed an R9 280X last night, and it has run dry several times already!

Those cards are wonderful, I have 6. Not sure what MW is going to do when they've all expired and everyone uses the modern pitiful ones without much DP.

yep