Posts by Mr. Kevvy


log in
11) Message boards : Number crunching : Average Credit Decreasing? (Message 1782337)
Posted 4 days ago by Profile Mr. KevvyProject donor
Validation pending is approaching 2000 due to these things, lol


7942 currently. :^)
12) Message boards : News : Breakthrough Listen data now flowing. (Message 1781949)
Posted 5 days ago by Profile Mr. KevvyProject donor
I remember reading that Green Bank will scan the local group of galaxies, including the Maffei galaxies.


I thought the targeted search would only be the most promising Kepler objects... the furthest I know of is Kepler-40B at 8,800 ly. M31 is about 2.5Mly, so a signal from M31 would have to be at least (2.5M/8800)^2 ~ 80,700x stronger to be detected. Assuming evenly distributed signal power, we're the reciprocal of this as likely to discover something at that distance. If signal powers are more commonly what we can produce (into the megawatt range) it's even more unlikely or impossible (plus at that angular resolution isn't it going to be drowned out by the rest of the galaxy?) Hrm.

Edit: Maybe for these observations SAH is just "going along for the ride" as we do with Arecibo.
13) Message boards : News : Breakthrough Listen data now flowing. (Message 1781684)
Posted 6 days ago by Profile Mr. KevvyProject donor
BTW, the MESSIER031 tasks are non-VLAR as far as I see and can be crunched on GPU, so no one is excluded.


Yes, getting them and very happy about that. :^) Unfortunately about 90% of them are going into -9 due to 30 or more triplets... hope it's just an initial hiccup. (These results are matching up with others', so not just my machines.)

I'm also curious what "MESSIER031" really means... Messier catalog object 31 is the Andromeda galaxy, and I doubt we are looking for ETI signals from that distance! Perhaps the area of?
14) Message boards : Number crunching : Average Credit Decreasing? (Message 1781007)
Posted 8 days ago by Profile Mr. KevvyProject donor
I've been watching this for a week or so, and there's no way it's an illusion. I started because I was trying to figure out why mine was dropping... ie machines not getting GPU work due to GUPPI? Technical issues ie bad cruncher spitting out Invalids? One of the ways it can be seen that RAC is constantly falling is to check the Top participants. This list updates infrequently, so is a snapshot of participant RAC at that point in time. Going into each profile shows their current RAC.

Here's what we see from the top 12 at the time I posted this:

Name TP RAC Current RAC Change Lee Klopp 714,808 714,145 -663 MJS 543,177 540,824 -2,353 BojanR 455,817 454,691 -1,126 WB5TUF - Glenn 449,755 448,701 -1,054 Igor_b 426,679 426,241 -438 Fawkesguy 381,914 381,269 -645 Mr. Kevvy 375,483 374,294 -1,189 msattler 308,949 308,329 -620 DARRENS 283,672 283,181 -491 Charles Long 270,878 271,526 +648 TSTC - Cloud... 270,009 269,847 -162 toaster 260,086 259,899 -187


One increase and 11 decreases. Usually I don't even see the one increase, either. And this has been pretty much constant for over a week.
15) Message boards : Number crunching : Average Credit Decreasing? (Message 1780995)
Posted 8 days ago by Profile Mr. KevvyProject donor
Hopefully you will see the graph which shows you what I am seeing

http://stats.free-dc.org/stats.php?page=user&proj=sah&name=71946


Don't see it there. I tried to post a graph but had the wrong one... I'll see if I can dig one one up, but yes, something has definitely changed.

Edit: The decrease can be seen here. Note the Credit per Day fall. (Ignore the Wednesdays as there is always a huge drop after the Tuesday outages.)
16) Message boards : Number crunching : Can Seti@home keep up with Breakthrough Listen data? (Message 1780679)
Posted 9 days ago by Profile Mr. KevvyProject donor
Well we can try to keep up


Ironic, as NVidia devices aren't getting any Breakthrough Listen data currently. Hope we are allowed to change that soon.
17) Message boards : News : Breakthrough Listen data now flowing. (Message 1780366)
Posted 10 days ago by Profile Mr. KevvyProject donor
I don't see why someone would want to do only 1/3 or 1/4 of what he can actually do for the project.


I estimate we have from hundreds to thousands of times greater of a chance finding ETI in a GBT work unit than an Arecibo one, because it is targeted to stars that are nearby, known to have exoplanets, and known (or thought) to have them in favourable positions for life (as we know it, Jim.)

It comes down to quality over quantity. And it's rather sadly ironic that the people who have contributed to this project the most over the years in terms of work done and time/money/effort invested in doing so are now mostly excluded from its most anticipated and promising venture.
18) Message boards : News : Breakthrough Listen data now flowing. (Message 1780362)
Posted 10 days ago by Profile Mr. KevvyProject donor
Efficient is to send a task to a processing unit, which can process it fastest compared to other type of work on the same PU. Sending a VLAR to a GPU, which could crunch 3-4 mid-AR tasks (according to message 1779961) in the same time is waisting of ressources, since a CPU does the VLAR in nearly same time as a mid-AR WU.


Again, this is only correct for CUDA. The VLAR issue does not affect OpenCL so GUPPI work units would complete in about the same time. Most people are going to have both platforms available, as far as I know. All we ask for is to be able to make the decision to run them ourselves.
19) Message boards : Cafe SETI : A little bit of the kittyman's mind. (Message 1780168)
Posted 11 days ago by Profile Mr. KevvyProject donor
In 1966, the mean temperature on earth was 58.6 degrees.
In 2016, the mean temperature on earth is still 58.6 degrees.

God bless Al Gore, the idiot.






This is mainstream-science, peer-reviewed, universally accepted (among scientists at least) data correlated from about 6,300 weather stations worldwide. It's also published by NASA. (Source)

If you don't trust this level of reliability, what will you do when SETI@Home detects and verifies an ETI signal using the same level of rigor, the same scientific principles, the same methodology of peer review? Why would it be any more reliable than this?

I'm sure when it happens you won't take kindly to the people who found it being called idiots by outsiders.
20) Message boards : Cafe SETI : What If... (Message 1779785)
Posted 13 days ago by Profile Mr. KevvyProject donor
21) Message boards : Cafe SETI : Anyone´s leaving the project ? (Message 1779514)
Posted 13 days ago by Profile Mr. KevvyProject donor
Mark,

If one of the criteria of this decision is that you feel left out of the GBT data as you have an exclusively NVidia farm, as I do, I'd request you add your voice to the Breakthrough Listen News thread asking that we have the option to allow it on NVidia if we choose to. I think by now we know the risks. :^)

I hope you can stick around though. As I have told myself many time, at this points: what's left to do? You could dial it back a bit and save some $ though... I wouldn't be able to stomach those power bills.
22) Message boards : News : Crowdfunding to support our undergrads (Message 1779495)
Posted 13 days ago by Profile Mr. KevvyProject donor
Thanks for noticing. :^) Glad it's over the $5K threshold for at least a basic setup.
23) Message boards : News : Breakthrough Listen data now flowing. (Message 1779282)
Posted 14 days ago by Profile Mr. KevvyProject donor
To be able to run VLARs on my Nvidia GPUs I'd live with a "health warning" as suggested by Mr.Kevvy. Although I suspect some would just ignore the warning and moan like crazy.


Enstein@Home, rather than using app_info.xml to set the <count> parameter, has it in the project preferences as follows:

GPU utilization factor of BRP apps
DANGEROUS! Only touch this if you are absolutely sure of what you are doing!
Wrong setting might even damage your computer! Use solely on your own risk!
Min: -1.0 / Max: 1.0 / Default: 1.0


This seems to mitigate the moaning, so a similar caveat would certainly be sufficient here where there's less chance of overheating/hardware damage.

(My Linux machines have both CUDA and OpenCL. If they were given GUPPI VLARs, wondering how it is determined which of the two would pick them up.)
24) Message boards : Cafe SETI : Thor is the Designer Cafe! (Message 1779208)
Posted 14 days ago by Profile Mr. KevvyProject donor
Thor flunked grade three twice...


He passed the thord time!

(You're making this too easy for me. :^D)
25) Message boards : Cafe SETI : Thor is the Designer Cafe! (Message 1779185)
Posted 15 days ago by Profile Mr. KevvyProject donor
As a warning you should not be so flippant, remember Thor's power. As an example the poster known as ID left the project after Thor arrived.


He underestimated the power of Nathoral Thorlection!
26) Message boards : Cafe SETI : Anyone´s leaving the project ? (Message 1779041)
Posted 15 days ago by Profile Mr. KevvyProject donor
Credits at different projects are very much like national currencies: I could make a higher number elsewhere, but as they are worth correspondingly less, it isn't worth moving there over. :^)

When I was crunching for Einstein@Home, for example, I noticed that my daily production/RAC topped out at almost exactly 4x what it peaked at here with the same farm. SETI@Home, being run by the developers of BOINC, uses a credit system based the actual FLOP count of the work unit (you can see how many operations it took in your account -> valid tasks). Some projects pay a flat rate per work unit, and of course this is open to inflation.
27) Message boards : News : Breakthrough Listen data now flowing. (Message 1778572)
Posted 16 days ago by Profile Mr. KevvyProject donor
The VLARs go out to ATI GPUs, but are held back from NVIDIA. If I could find a way to send VLAR to only NVIDIA OpenCL and not NVIDIA CUDA, I would...


What would be a possible workaround in the interim (until someone cracks the issue of why VLAR/VHAR(?) bogs CUDA pulse detection) is a default-off setting in SETI@Home Preferences ie "Accept VLAR work units on NVidia CUDA (warning: these units may complete very slowly and cause the computer to become unresponsive)"

(Those with both dedicated crunchers and workstation computers they actually use could prevent the ones they use from slowing down by adding "home"/"school"/"work" separate preferences, putting the dedicated crunchers in one "location" setting with accepting VLAR on CUDA on, and the workstations in another "location" setting with it off.)

Here's to hoping someday someone can figure out a way to make that happen. In the meantime, looks like the need for some CPU heavy rig building.


I'll have to look at it that if the work that my NVidia GPUs do frees up CPUs and AMDs to do the work to make The Great Discovery, it's still worth it... I'll be happy if anyone makes it. :^)
28) Message boards : News : Breakthrough Listen data now flowing. (Message 1778415)
Posted 17 days ago by Profile Mr. KevvyProject donor
I'd expect about 50/50 for the time being.


Unexpectedly large amount... excellent: doubled workload (hope the infrastructure can keep up.)

Also it was noted in NC that much of this may be VLAR due to being targeted so may be minimal(?) GPU involvement.
29) Message boards : Number crunching : New Greenbank Files (Message 1778410)
Posted 17 days ago by Profile Mr. KevvyProject donor
Guess it's time for me to invest in a CPU only cruncher, lol...


This thread may prove helpful :^)
30) Message boards : News : Crowdfunding to support our undergrads (Message 1778396)
Posted 17 days ago by Profile Mr. KevvyProject donor
I noticed that possibly because this campaign uses crowdfund.berkeley.edu rather than give.berkeley.edu that the donation doesn't seem to cross over to our profile... mine still shows the same number and amount of donations as before I donated, and this means that our regular "little green star" probably won't appear or be extended by it (is this correct?)


Previous 20 · Next 20

Copyright © 2016 University of California