Having had the machines run virtually dry last week and on life support over the weekend till the flood gates were opened, i thought i might share some observations.

1. Total work that went into queue, circa 2800 wus.
2. At the time of the outage swith on, circa 50% of wus had still not been downloaded (they are pending)
3. It's now 5:30pm on Thursday in Oz and 20% of wus have been processed and are now pending upload.

Not sure how many others have had a similar experience however it will be interesting to see what happens when things come back up...

L.

---

On this system, which is a virtual machine running OpenSolaris, I have 2 WUs in progress (but one has finished and is trying to upload) and 2 awaiting validation, notwithstanding the fact that this is a slow machine. On the real machine, which is a Linux box, I have one in progress and 4 awaiting validation, but this was true even before the outage. People load too many WUs, in my humble opinion.
Tullio

---

Hopefully they will allow uploads to settle in a bit before the downloads are enabled. While I may be have an empty machine or two by then, I would hate to see everything falling into error and crashing things in too much of a rush.

The initial "20" limit last time really was a good idea.. It just left a lot of people with empty GPU's or CPU's. Hopefully it will be squared away, and moderate increases in it will not crash systems.

---

Janice

I would agree about the rolling 20 unit process queue being a good idea.

I think the problems were that it seemed to ignore the GPU serving on a first basis the Host CPU and the other problem being 20 units was unlikely to provide enough work to cover the new 3 day schedule.

the advantages every machine has work to process either already queued. or are given up to the magic number* of workunits so they have something to do. Outstanding work is hopefully reduced as people with queues above this limit return completed work and are not replenished until the magic number* is involved.

I survived the first 3 day knock-down round having 78 WUs approximately 2 dozen of these were shorties (~1 hr processing time) connected during Friday Berkeley Time and got a top up to approximately 60 and then over the weekend actually worked the queue down to 19 and got a top up from this rolling 20 unit queue a few hours before they lifted it so everyone could get enough to last this second 3 day knock-down.

I lowered my queue from 6 days to 5 and got a total processing queue of 61 WUs only 6 short WUs in these. Currently have processed 29WUs (includes shorties) so I've enough to see me through round 2. May or may not lower queue further depends on estimates and shortie spread.

BTW I'm pleased to see you still around soft^spirit having moved beyond 100K.

* magic number may or may not be 20.

I do not think anyone really WANTS 10 days+ of work.. But people have been creative to try and get enough to last through the outage.

It looks like I will start running dry soon.. but not badly before tomorrow.
And yes.. still crunching away for now.

We all have limits on the torpedos we can take. Mine got stretched. Still at it to see what the new normal will be.

---

Janice

Still at it to see what the new normal will be.

That pretty much sums up life these days - as well as S@H. Grinding through a bunch of shorties right now, but it looks like I will make through the outrage with WUs to spare. I agree that the upload rush when they turn back on will be interesting.

An interesting side effect: my S@H queue was one or two WUs right up to 18 hours before the outage, and my other projects seemed to take this as an opportunity to overstock. Once I got a whack of S@H work, WCG started jumping in and out of high priority mode.

---

On this system, which is a virtual machine running OpenSolaris, I have 2 WUs in progress (but one has finished and is trying to upload) and 2 awaiting validation, notwithstanding the fact that this is a slow machine. On the real machine, which is a Linux box, I have one in progress and 4 awaiting validation, but this was true even before the outage. People load too many WUs, in my humble opinion.
Tullio

I have a comparatively slow machine by modern standards, certainly not a super cruncher, with a Q6600 & a GTX8800 that is mildly overclocked. This gets through around 6-10 MB WU an hour depending on type. This means it needs around 240 WU per day to survive a 3+ day outage I need around 500-700 WU - fewer if I snag some AP or the WU are longer but I still need hundreds of units.

Are you saying I should only have a cache of 10 or so :)?

I have a comparatively slow machine by modern standards, certainly not a super cruncher, with a Q6600 & a GTX8800 that is mildly overclocked. This gets through around 6-10 MB WU an hour depending on type. This means it needs around 240 WU per day to survive a 3+ day outage I need around 500-700 WU - fewer if I snag some AP or the WU are longer but I still need hundreds of units.

Are you saying I should only have a cache of 10 or so :)?

I am simply asking why my two machines, an average CPU (Opteron 1210 at 1.8 GHz running Linux) and a much slower virtual machine running Solaris must always wait for their wingmen, who usually have much faster CPUs. I can only think that my wingmen have downloaded too many WUs. And yes, I am running 5 other projects on the Linux box, only SETI on Solaris, so I am not even running SETI all the time. All my projects are equally shared.
Tullio

---

it's all about the cache. those with small caches will always complete their work faster than those with a large cache. I like to think I hit it at a medium sized cache. That way folks with small a small cache like yourself don't have to wait long and I dont have to wait long for the folks with a large cache

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

its now Sat 12:30pm in Oz. Just looked at Pending Credits, they now sit at 345k+, whereas before the changes they sat around 140k...

---