Why are we free-floating?

This thread is "free floating" until GPUUG sorts out an internal staffing problem, and the new priorities for equipment/general funding are agreed between GPUUG and SETI@Home.

---

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

Yes, things are being worked on at the GPUUG.
A new fundraiser shall be forthcoming fairly soon.
It looks to be another rather ambitious one.
Since the move to the colocation, what the staff thinks they now need most might be a bit different than what was originally set out.

Please stay tuned.

Meow.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

msattler wrote:

Yes, things are being worked on at the GPUUG.

That's good to hear.

Unfortunately, it appears that Brad/Slavac isn't going to be returning. So, his roles will be distributed amongst other GPUUG members.

I've unofficially picked up a couple of them... I've been communicating with the SAH team regarding fundraising requirements, due to my interest in doing another matching-fundraiser weekend to get the Green Bank project operational ASAP, and since there have been some changes to the GPUUG site I think the info on recent updates should be passed on...

---

The project for the Upload / Download / NTPCKR servers has been removed from the GPUUG site as they are no longer required. In addition to numerous other benefits provided by the recent move to the colocation facility, there is now sufficient hardware to perform these functions. NTPCKR is, as far as I know, still in the development phase, which is why the existing data is not accessible, but when it goes public there should also be sufficient hardware to get it going. If not, the SAH team should certainly pass on any requirements for it.

The donations to GPUUG for the servers will be applied to a future project or to the "General Donation to The GPU Users Group" fund, which is used for emergency requirements such as hardware to repair server breakdowns. If you were a donor to the server project and wish to choose where your funds are directed, please advise and I will pass on the info.

As far as (near) future fundraising goes, the SAH team is working out the requirements. Right now, it appears as though (finally!) there is enough hardware in Berkeley to keep everyone's crunchers happy with work and handle the incoming data from Arecibo. The focus is therefore to get the Green Bank project up and running. The previous fundraiser's goal to buy hard drives et al for this has been put to good use. The SAH team released some preliminary Green Bank results in February. So, to those that donated, soon we should be seeing some actual work units from this!

It would appear that what is currently required to finalize Green Bank, with sufficient capacity to provide us with enough work, is a spectrometer array for processing the signals so that they can be analyzed and eventually divvied up into work units. At first it looked like this would require about 8 cores worth of machines plus GPUs to be purchased, but this requirement may have been mostly fulfilled with machines donated from another project.

So, we're (unofficially!) currently looking at only needing about 8 GPUs worth of compute requirement at the Green Bank site to get the project underway. This would be ideal for a matching weekend fundraiser.

---

I'm going to try to update this thread as things progress. It isn't pinned anymore, so this will also prevent it from getting locked again due to inactivity!

If anyone has any inquiries, I'll try to answer them as best as I can, or pass them on. The site admin. (Deltoid) just left for vacation so I can't promise promptness. :^)

---

Right now, it appears as though (finally!) there is enough hardware in Berkeley to keep everyone's crunchers happy with work and handle the incoming data from Arecibo.

What would it take to enable caches to be made large enough to get through the weekly outage again? And large enough to cover a day or two of down time?
A four day cache would be ideal- large enough to get through a long weekend outage & not require the staff to intervene during the break to keep crunchers crunching.

---

Grant
Darwin NT

What would it take to enable caches to be made large enough to get through the weekly outage again? And large enough to cover a day or two of down time?
A four day cache would be ideal- large enough to get through a long weekend outage & not require the staff to intervene during the break to keep crunchers crunching.

Unfortunately, nothing that we could fix... as far as I know it wasn't a storage/hardware limitation but instead was imposed due to some hosts filling the database with ghosted workunits. I know they're aware of the ever-present requests to enlarge the cache... don't know if it can be done without causing the same problems to reoccur.

I'd like to see the limits increased as well, but then one of my cheapo crunchers needing a reboot flips out and errors out every WU in its cache so I can understand why the limit is there. :^)

---