Work Shortage
2 Aug 2020 14:16:43 UTC
Topic 223209
(moderation:
It looks like the project is out of Gamma Ray GPU tasks. I noticed all my AMD tasks finished about 07:30 EDT and I only have 3 NVIDIA tasks left in the queue. I supose I could go over to GW tasks with the AMD, but I'll need to disable the NVIDIA card, as it only has 2 GB of memory. GW files do not like that. :)
Clear skies,
Matt
My most recently downloaded
)
My most recently downloaded _0 or _1 GRP GPU tasks came about 4 hours ago, so I suppose new WU generation died a few hours ago.
I suspect you will find that GRP GPU work generation stoppage is a temporary failure, likely fixed early in the new working week, while GW GPU work seems to need a whole new series to be set up, and quite likely will take rather longer.
Makes sense. I shifted my GPU
)
Makes sense. I shifted my GPU box over to Milky-way@home while I wait for more work to be available here. My CPU box is not hungry, yet. :)
Matt White wrote: Makes
)
I've got my GPUs (apart from the one that expired yesterday - no overheat protection on the voltage regulator, seriously?) on Milkyway aswell, they normally do half of each, so they shifted automatically :-)
CPU Gamma work is running as normal, so you won't run out of that.
I noticed a few days ago that GPU Gamma only had 3.5 days of work left, and would run out last night. Then the next day two months of work was added to the queue, but that has since been removed, perhaps there was an error in the new batch?
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
I hope this gets fixed early
)
I hope this gets fixed early tomorrow because I only have ~ 17hrs of pulsars on both boxes then the well will be dry.
archae86 wrote:... while GW
)
Are they working on a new (more efficient) GPU app that you know of? I can't find anything definite.
Betreger wrote: I hope this
)
I only have a 3+3 hour queue. It makes it quicker for any adjustments I make (like project weighting) to take effect, and it means the scientists get the jobs back quicker.
As for running out, impossible, I run enough projects so there's always something else to do. GPUs for example go off to do Milkyway.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
I find a 1 day cache works
)
I find a 1 day cache works well for me.
Betreger wrote:I find a 1
)
What's the advantage of a larger cache? The only reason I have a cache at all is Milkyway work units can be 30 seconds! And I can't contact the server that often. P.S. it's "buffer" :-)
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
The advantage is this week I
)
The advantage is this week I lost internet service for 10hrs before it could be restored.
Betreger wrote: The
)
Ok that's a reason (if it's regular, otherwise it's no big deal if it doesn't run now and then). I used to have internet like that, but that was about 15 years ago and due to a leaking junction which an engineer had installed incorrectly across the road.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.