Posts by HAL9000

1) Message boards : Number crunching : Panic Mode On (109) Server Problems? (Message 1913877)
Posted 3 days ago by Profile HAL9000
Post:
MW is fine, but IIRC to do something productive there you need to have an AMD GPU, NV stuff has troubles to work with DP used by MW.

Does that changes?

GPUGrid with it's long time to crunch WU is not really a project to use as a backup. IMHO

Milkyway used Double Precision calculations.
Most GeForce GPUs are allowed DP performance of 1/32 SP
Radeon GPUs are allowed DP performance of 1/16 SP

So if both GPUs were 6000 GFLOPS in Single Precision the GeForce would be 188 GFLOPS DP and the Radeon 375 GFLOPS

If you move to the workstation GPUs they will have DP performance of up to 1/2 SP. Which is likely why they have 4 digit price tags.
2) Message boards : Number crunching : 100 WU limit for GPUs is too low. (Message 1913285)
Posted 6 days ago by Profile HAL9000
Post:
Is there a 'better' database to use than Informix? What does Google use, for example? SQL by MS? And is it practical to switch? Though I have heard of Informix for quite some time, I just Googled it and found that it is an IBM product. I guess I am old, but I thought that IBM had a reputation for doing big iron, and a solid DB? Or has that ossified over the years?

*edit* Just took a closer look at the quick blurb on Google, and this caught my eye: Stable release: 12.10.xC7 / June 15, 2016. I can appreciate having something just work if it ain't broke, but I don't know, it seems to me that over 18 months between stable releases seems like a pretty long time in the compressed IT world we live in now. But then again, what do I know? ;-)

It is the BOINC master database which uses mySQL rather than the science databases that use Informix. The issue is more how SETI@home uses the database.
For all of the tasks assigned to hosts, "results out in the field", the task information is stored in a single table.
So when tasks are assigned, reported, or a host doesn't an update then that table is accessed.
For the hardware running the db the limit of that table turns out to be around 11,000,000 rows. Then it can not longer complete a query fast enough. Which causes the db server to stop and the project to be offline for several hours if not days which recovery is run.

So the solutions are
1) Have more hefty master and replica db servers.
2) Recode how the results are sent to hosts are stored.
3) Limit the number of tasks sent to hosts.

#1 is a stop gap until that hardware hits its limit.
#2 is probably the best solution, but requires dedicated time that isn't available to the project.
#3 easiest to implement and can be adjusted if the db server gets near the limit again.
3) Message boards : Number crunching : 100 WU limit for GPUs is too low. (Message 1913217)
Posted 6 days ago by Profile HAL9000
Post:
As I just learned there is a maximum of 100 WUs per compute device, or device class, and with an average of about 4-5 minutes per WU, that's a maximum of about 8 hours of work I can store, and with the splitters out all the time recently, this really doesn't cut it.

I'm already out of work for two hours with the current outage.

A cache of 100 tasks is only about an 8 hour cache for my dual E5-2670, or R9 390X as well. When we have shorties it is even less.
There are a few options to handle running out of work during an outage.
1) Have a backup project in BOINC.
2) Let your system(s) have a break when they run out of work.
3) Not run the project.
4) Message boards : Number crunching : Panic Mode On (109) Server Problems? (Message 1912538)
Posted 9 days ago by Profile HAL9000
Post:
I don't know what you guys are doing, the time estimates on both of my boxes with GTX1060s are very accurate. I run 2 at a time and they process at the rate of 4 an hour, so they estimate 30 min a task.

I was running my R9 390x 24/7 on it for a few months until I rotated to Milkyway. The only work fetch issue I had was not filling my 5+5 day cache queue. I was only getting ~8 days worth of tasks.
5) Message boards : Number crunching : A very steep decline in Average Credits!!! (Message 1911959)
Posted 13 days ago by Profile HAL9000
Post:
My RAC is still going up after the server issues in December.
6) Message boards : Number crunching : Anything relating to AstroPulse tasks (Message 1911012)
Posted 16 days ago by Profile HAL9000
Post:
Since the statement has been made there is no more Arecibo work in hopper, or available, I would say the future of this thread is also in grave doubt.........

Plenty of AP resends still to come, then people can discuss how long it was since they last had an AP WU, and how much Credit they gave compared to MB..
I reckon it's got years of life, even if there is no more AP data to be split.

We can also complain about why nobody is making an app for AP for work from GBT or Parkes.
That should keep the thread alive for years too :-)

I'm reasonably sure that it was mentioned they wanted to do AP for GBT data as well.
The first step requires the guys in the lab have the time to write the new splitter code, or have the budget for interns to tackle it.
Until that is done working on an app to processes non existent workunits wouldn't be much of a priority.
7) Message boards : Number crunching : Project/Stats/Tables.xml (Message 1910795)
Posted 16 days ago by Profile HAL9000
Post:
The individual user/host/team credit values are being exported normally - that's where BOINCstats and the other similar sites get their data from - but it's the project aggregate values which are missing. BOINCstats obviously recover the aggregate total by summing the individual values as they process them.

Yeah I was searching the other files to see if <nusers>, <nteams>, <nhosts>, and <total_credit> were moved to one of the other files.
Since SETI@home is often a test bed for BOINC changes it seemed reasonable that they might have been moved to a new location.
8) Message boards : Number crunching : Project/Stats/Tables.xml (Message 1910785)
Posted 16 days ago by Profile HAL9000
Post:
I wasn't sure if it was missing due recent server issues SEI@home has had, but that info is not present in any of the saved stat dumps I have going back to 2017-07.
It doesn't appear to be in the host, team, or user output files either.
9) Message boards : Number crunching : Intel security flaw (Message 1910694)
Posted 17 days ago by Profile HAL9000
Post:
I am not really worried about these issues for any of my home systems. As my systems would have to already become compromised by an attacker. At which point they could just access any data directly.

How it effects environments like AWS is really my only concern.
10) Message boards : Number crunching : E5-2670 -> E5-2680v2 (Message 1910689)
Posted 17 days ago by Profile HAL9000
Post:
When I have that problem (compound, or even finger prints) after opening up some of the used CPU's I've bought in the past, I just go to the medicine cabinet and get some 92% rubbing (Isopropyl) alcohol and a lint free cloth, and a couple wipes with fresh spots on the towel each time & it's all gone. At least that's been my experience. I like the higher proof, more concentrated I guess.

I normally start with dry wipes to get whatever will come off free first. Then hit it with rubbing alcohol and I normally use q-tips until they come back clean. I haven't had any issues leaving any q-tip fluff behind, but it is something I watch out for when cleaning contacts that way.
11) Message boards : Number crunching : Intel security flaw (Message 1910453)
Posted 18 days ago by Profile HAL9000
Post:
So far the details seem to be that some parts of protected kernel memory can be read.
Some sites are reporting that the issue is also present in ARM processors.

MS has issued the patch in November to users in the "fast ring" of updates and Apple pushed out an initial patch in early December.
12) Message boards : Number crunching : The BOINC site is now "Science United". What the...... (Message 1910446)
Posted 18 days ago by Profile HAL9000
Post:
It has also removed the BOINC download page. So no new users are currently able to join ANY projects.
Unless they with to participate in the Science United Alpha test.

The archived installer are still accessible via /dl at the moment.
13) Message boards : Number crunching : E5-2670 -> E5-2680v2 (Message 1910206)
Posted 19 days ago by Profile HAL9000
Post:

What motherboard do you have to put the CPUs into?

I have an Intel s2600cp2j from a Natex bundle and kind of wondering if this would work for me too. $59 per CPU upgrade seems pretty good.


I have 2 - one is an ASUS Z9PE-D16 (the one with 4(!)Ethernet ports), the other is an ASROCK Rack EP2C602.

Both have BIOS which should support the V2 CPUs (waiting for them to come to find out for sure!).

Check out the BIOS in yours to see if it will support the V2s; if not, check the MB webpage to see if such a BIOS is available (fingers crossed!).

EDIT: BTW, my machines both make 75-80K RAC, but both have dual GTX 1080s installed, running 3 WUs each.

Luckily the Intel S2600CP MBs do support the v2 processors, up to 135w, and also allow the use of 1866MHz memory with using v2 processors.
I grabbed a few of the E5-2670 S2600CP MB bundles from Natex last year.
I think the worst thing for me will be putting 256GB of 1866MHz in each board after upgrading the CPUs. That stuff still isn't doesn't come cheap.
14) Message boards : Number crunching : Are all CUDA GPUs supported? Is the GT 1030 unsupported? (Message 1910205)
Posted 19 days ago by Profile HAL9000
Post:
I imagine a GT 1030 rated for 942.3 GFLOPS will be a bit faster than the old 256MB 8500 GT rated for 28.8 GFLOPS I was using for a few years..
Plus the 1030 can run the OpenCL apps. Which tend to run better with some of the newer data from the GBT.

It wouldn't be a bad idea if they updated the System Requirement section on the SETI@home site to indicate GPU requirements.
15) Message boards : Number crunching : Weekly Schedules on Boinc (Message 1908659)
Posted 29 days ago by Profile HAL9000
Post:
Add in as many cards as will fit in your computer. Then add this option <use_all_gpus>1</use_all_gpus> to your cc_config.xml file in your BOINC directory with Notepad.Then use Re-read config files from the Options menu in the Manager. You should see all the gpu's that are installed picked up in the Event Log. Boinc will use all available resources.

Weeeelllll... Maybe. I found that I had hit the limit of either 5 or 6 on one of my systems with dual proc Xeons and 10 4x PCI-E slots that I am using USB expanders on, and it has I believe 8 or 9 cards hooked up to it that way, but it ran out of memory space if I remember correctly, and I had to disable them one at a time until it would finally boot. I also had to go into the BIOS and disable everything that I could that wasn't necessary, to free up as much as I could. But in the 'normal' world, I'd say yeah, as many as you can get to fit onto the MB. ;-)

I wonder if that would still be an issue with an UEFI MB & UEFI GPUs.
16) Message boards : Number crunching : Weekly Schedules on Boinc (Message 1908623)
Posted 29 days ago by Profile HAL9000
Post:
Using a start time of 24:00 and an end time of 00:00 will disable network activity.
That allows two ways to have network activity only on specific days.

You can use 24:00-00:00 for the general time parameters and override on the day you want to allow. The inverse can also be used.

Both of these examples will only allow BOINC to use the network for 1 hour on Friday between 13:00 and 14:00


Another option is set the time you want to allow in the general time parameters and then have a script run boinccmd --set_network_mode disable always and boinccmd --set_network_mode enable auto when you wouldl ike BOINC to stop and start network activity.
17) Message boards : Number crunching : Seti@Home or Atropulse? (Message 1908244)
Posted 21 Dec 2017 by Profile HAL9000
Post:
... and the MB credit is highly dependent on the angle rate (ar) of the packet. So is the runtime between different apps.
http://setiathome.berkeley.edu/workunit.php?wuid=2784838125
MB can pay more than one credit per second. I have not seen AP do that.

Petri

I have, but it requires a Radeon GPU
18) Message boards : Number crunching : Multicore smartphones - how to use their full potential (Message 1907237)
Posted 15 Dec 2017 by Profile HAL9000
Post:
By enabling Developer options/mode they then make other options, such as screen staying on while on charger, available.
Anything similar for Android phones?

Thanks for the idea!
The phone is now lying face-down on my night stand, continuously running two Seti tasks. I also know now why the CPU would clock down as the temperatures in this state are high. For just two cores running at full blast (and a third being used by the OS), temps have gone up from 25C to 45C and they're still rising. Glad it's cold in the bedroom.

The option is Preferences->Developer options->Stay awake (Screen will never sleep while charging)

I ran No Screen Off on that phone before but it would do so for only 10 minutes (600 seconds). I didn't dare try the other one when I read one or more people bricked their device with it. So this is the easier thing to do.
Of course enabling the Developer options is a pre (Open Settings> About> Software Information> More. Then tap “Build number” seven times to enable Developer options in the main preferences menu.)

Where I wouldn't mind leaving the phone alone and walk out the house, I'll unplug it before doing so later on. Especially with the higher heat build-up of the CPU and charger, I'd rather be home when it's doing stuff.

Reducing the screen brightness can help lower the heat build up on some devices.
19) Message boards : Number crunching : BOINC Manager replacement (Message 1907183)
Posted 15 Dec 2017 by Profile HAL9000
Post:
Personally I find the Processes tab in the Win 10 task manager as useful as the Notices tab in BOINC manager.
20) Message boards : Number crunching : Bug in Top Computers List (Message 1907096)
Posted 14 Dec 2017 by Profile HAL9000
Post:
It looks like your other machine has appeared. With your two hosts holding #30 and #40.


Next 20


 
©2018 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.