Posts by ivan


log in
1) Message boards : Number crunching : SETI/BOINC Milestones [ v2.0 ] - XXIX (Message 1754816)
Posted 8 Jan 2016 by Profile ivan
Just got another notch on my pistol -- now up to 44th overall. I won't get any further for quite some time it seems, and in fact might get caught by some coming up behind me. Next stop, 200 mill.

Well, there's the 200 million at last. RAC's dropping like a stone as I lose GPUs to MB V8, but there are enough CPUs to keep me ticking along.
2) Message boards : Number crunching : Intel Compute Stick (Message 1748677)
Posted 12 Dec 2015 by Profile ivan
Thanks everyone for your input. Looks like I save myself eighty squid and burn gas rather than electricity to top up the heat in the Rechenzentrum.
3) Message boards : Number crunching : Intel Compute Stick (Message 1748117)
Posted 10 Dec 2015 by Profile ivan
Is anyone using an Intel compute stick for S@H? I bought myself a 4K monitor fot £200 in the Black Friday sales, and am driving it from my Linux Mate PC (GTX 660 Ti) via Display Port. This means the HDMI port is available for another device...
eBuyer are selling the Windows version of the Compute Stick very cheaply at the moment -- cheaper even than the less-provisioned Ubuntu version! If it managed around 2,000 RAC like my similarly-equipped J1900 it might be a cheap addition to my stable.
4) Message boards : Number crunching : SETI/BOINC Milestones [ v2.0 ] - XXIX (Message 1747931)
Posted 9 Dec 2015 by Profile ivan
Just got another notch on my pistol -- now up to 44th overall. I won't get any further for quite some time it seems, and in fact might get caught by some coming up behind me. Next stop, 200 mill.
5) Message boards : Number crunching : I'm Trying to Build an OSX CUDA App... (Message 1746153)
Posted 1 Dec 2015 by Profile ivan
I thought Apples were supposed to be all shiny and you just mash things with your forehead.... apparently not.

No, that's watermelons.
6) Message boards : Number crunching : Christmas Wish List! (Message 1742504)
Posted 16 Nov 2015 by Profile ivan
Well folks, it's that time of year again when we seriously start thinking what we want for Christmas. So, I'll get it started. http://www.guru3d.com/news-story/intel-core-i7-6950x-broadwell-e-has-10-cores-and-20-threads.html I'll even settle for the 6800K.

Hmm, I had ten-core Xeons over a year ago.
7) Message boards : Number crunching : NVIDIA getting ready to unleash Jetson TX1 module (Message 1741332)
Posted 11 Nov 2015 by Profile ivan
http://nvidianews.nvidia.com/news/tiny-nvidia-supercomputer-to-bring-artificial-intelligence-to-new-generation-of-autonomous-robots-and-drones

I can't wait to pre-order one of these... to crunch SETI WUs, what else?

I hope they don't mark it up in the UK as much as they did the TK1. Mine's happily crunching MB on four 32-bit cores and the Kepler GPU, at a RAC of just over 2,000. Assuming all 8 64-bit cores are usable (the fifth core of the TK1 doesn't show up on Ubuntu), and the larger, newer GPU, I'd expect the new version to probably double that.
8) Message boards : Number crunching : Panic Mode On (101) Server Problems? (Message 1736271)
Posted 22 Oct 2015 by Profile ivan
Of course when a tape is "slit"[...]
Careful with that axe, Eugene!
9) Message boards : Number crunching : Updating GPU drivers in Linux. (Message 1731575)
Posted 3 Oct 2015 by Profile ivan
My win install is running fine again.
So i tried my Mint 17.1 partition.
It does nothing, just a black screen.
Anyone complains about windows ?
Tried the recovery mode without luck.

How do i update the drivers when i can`t get into the system ?

Any help is much appreciated.

What happen if you type Ctrl+Alt+F1 or Ctrl+Alt+F2 simultaneously? That should drop you into a console session. Beyond that it gets hairy, so let's see if you get that far first of all.
10) Message boards : Number crunching : CMS@Home needs about fifty good volunteers... (Message 1731573)
Posted 3 Oct 2015 by Profile ivan
Will be any testing status updates here or SETI-originated participants need to look CMS-project site for any news now?

Well, there's a fairly good dialogue going on at http://boincai05.cern.ch/CMS-dev/forum_thread.php?id=79 so I won't repeat myself here. Bottom line, I think we're almost there but more work than expected has fallen on my shoulders. At present we don't need more testers, but we're not going to turn anyone away (numbers may have some impact on a presentation I'm to give at CERN on the 15th, I will admit).
11) Message boards : Number crunching : Panic Mode On (100) Server Problems? (Message 1730387)
Posted 30 Sep 2015 by Profile ivan
And it's getting worse. Now it fails every time. An hour or so ago it worked about every third attempt.

I blame the Pope.

Edit: And now the Replica DB went on a vacation too :-(

I'm not seeing any problems here, except one download failure (which is not that unusual inside our firewall/proxy). Ready-to-send buffer had taken a nosedive on the SSP earlier but seems to be recovering.
12) Message boards : Number crunching : SETI/BOINC Milestones [ v2.0 ] - XXIX (Message 1727572)
Posted 21 Sep 2015 by Profile ivan
40 mil SETI

and 100 mil all projects.

Well done, Kev!
13) Message boards : Number crunching : Panic Mode On (100) Server Problems? (Message 1726975)
Posted 18 Sep 2015 by Profile ivan
If those are the Opencl_nvidia_sah, that is normal for them.

Here's a little info for you on that. If you run 2 at a time of those, the usage will actually go down to about 85% of a core for each work unit.

I would not recommend going over 2 OpenCl_nviidia_sah per card.

On the plus side, you will help to get rid of all those VLARS ;)

I think the only machine I run >2 GPU jobs on is my work Linux machine, which is running anonymous. Not got the feedback yet (because the replica DB was still catching up last I looked) to see how this will affect my RAC -- I don't expect it to go down! What happens when the next batch of AP WUs are available is another matter.
14) Message boards : Number crunching : Panic Mode On (100) Server Problems? (Message 1726959)
Posted 18 Sep 2015 by Profile ivan
We-e-ll, I was just watching Great British Menu on iPlayer (my broadband is currently at 2268 kbps, just enough to watch 1280x720 HD) when it all stopped, even though gkrellm was showing the net flatlining at 248 kBps. Looked at BOINC and it was downloading shedloads of new jobs and a new executable. When the dust settled, there had been 91 new jobs and setiathome_7.08_x86_64-pc-linux-gnu__opencl_nvidia_sah -- so I guess that's out of beta. My RAC should go up, as I'd never got around to setting up the anonymous mechanism on this PC, so I've only had GPU work when there are AP jobs to be had. We shall see...

Might need some tuning -- the GPU jobs are hogging a CPU each!
top - 22:01:36 up 5 days, 19:07, 6 users, load average: 15.07, 14.97, 14.72 Tasks: 274 total, 13 running, 261 sleeping, 0 stopped, 0 zombie %Cpu(s): 2.0 us, 7.1 sy, 77.1 ni, 13.7 id, 0.1 wa, 0.0 hi, 0.0 si, 0.0 st KiB Mem: 8129424 total, 6893796 used, 1235628 free, 396776 buffers KiB Swap: 5119996 total, 36 used, 5119960 free. 3226776 cached Mem PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 8112 ivan 30 10 28.277g 80788 32484 R 99.7 1.0 8:10.46 setiathome_7.08 8610 ivan 30 10 28.279g 104412 48080 R 99.7 1.3 5:58.53 setiathome_7.08 4032 ivan 39 19 108976 39884 12 R 81.8 0.5 26:53.10 setiathome_7.01 4791 ivan 39 19 107752 38664 12 R 74.5 0.5 21:07.65 setiathome_7.01 30201 ivan 39 19 109012 39972 12 R 72.1 0.5 81:36.75 setiathome_7.01 6779 ivan 39 19 108060 39252 12 R 52.5 0.5 11:25.18 setiathome_7.01 6643 ivan 39 19 107548 38568 12 R 49.5 0.5 11:20.55 setiathome_7.01 20264 ivan 39 19 110816 42320 12 R 47.9 0.5 165:22.65 setiathome_7.01 32684 ivan 39 19 108824 39732 12 R 47.2 0.5 58:39.61 setiathome_7.01 7160 ivan 39 19 109356 40016 12 R 42.9 0.5 8:52.95 setiathome_7.01


[homepc01:BOINC] > nvidia-smi Fri Sep 18 22:03:55 2015 +------------------------------------------------------+ | NVIDIA-SMI 352.30 Driver Version: 352.30 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GTX 660 Ti Off | 0000:01:00.0 N/A | N/A | | 30% 65C P0 N/A / N/A | 644MiB / 2043MiB | N/A Default | +-------------------------------+----------------------+----------------------+
15) Message boards : Number crunching : Panic Mode On (100) Server Problems? (Message 1726955)
Posted 18 Sep 2015 by Profile ivan
setiathome_7.08_x86_64-pc-linux-gnu__opencl_nvidia_sah --

Wonder if there is an equivalent Windows release as well?

Not according to http://setiathome.berkeley.edu/apps.php , though there are some new Mac CPU apps as well.

I'd like to take a whack at it outside of Beta as well.

Hmm.. But you are right, anonymous platform will mean modification to try and get that app and executable...
16) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1726950)
Posted 18 Sep 2015 by Profile ivan
On my Windows 10 PC I see lots of data going up to CERN programs such as vLHC@home, Atlas@home and, especially, CMS-dev which is a Beta project.

Yeah, I'm still tuning that of course. We had a computing meeting today which was a bit more positive than earlier in the year. Things are probably going to change in the submission method, but I don't think the volunteers will see a difference -- I'm hoping it will make my part easier, though. I'm also starting to get some reaction to my plea for consistent jobs with minimal data return, just don't expect it Real Soon Now. I'm trying to retrieve recent data so that I can compare it with normal GRID jobs and report at a meeting at CERN next month. With a bit of luck I can get a publicly-available report out of it, which should please the volunteers.
(I realise this is more directed at CMS@Home, just reacting to tullio's comment.)
(As a consequence of my broadband woes, I can't run CatH at home myself -- not enough capacity. :-( )
17) Message boards : Number crunching : Panic Mode On (100) Server Problems? (Message 1726945)
Posted 18 Sep 2015 by Profile ivan
We-e-ll, I was just watching Great British Menu on iPlayer (my broadband is currently at 2268 kbps, just enough to watch 1280x720 HD) when it all stopped, even though gkrellm was showing the net flatlining at 248 kBps. Looked at BOINC and it was downloading shedloads of new jobs and a new executable. When the dust settled, there had been 91 new jobs and setiathome_7.08_x86_64-pc-linux-gnu__opencl_nvidia_sah -- so I guess that's out of beta. My RAC should go up, as I'd never got around to setting up the anonymous mechanism on this PC, so I've only had GPU work when there are AP jobs to be had. We shall see...
Can't check my full stats, though, with the replica being behind times.
Oh, and Matt beat Eve for the North-west crown once the judging was finished </spoiler>.
18) Message boards : Number crunching : Windows 10 - Yea or Nay? (Message 1725440)
Posted 13 Sep 2015 by Profile ivan
I have BOINC 7.6.6 on my 4-core Windows 10 64-bit CPU. ATLAS@home sends me a maximum of 4 tasks, vLHC@home 2 tasks, CMS-dev 1 task. They all use Virtual Box, no GPU. When I allow the downloading of SETI@home tasks, I get about 27 of them, both CPU and GPU, so they overfill my CPU. GPU tasks are run one at the time, CPU tasks occupy one or two cores, leaving the other 2 to the CERN programs. but why does SETI@home sends me all those tasks? On my 32-bit Linux laptop I get a maximum of 2. I am running 4 CPU tasks plus one GPU task.
Tullio

That's a per-project setting. When I first started up CMS@Home-dev. Laurence hadn't set that, so for my 8-(hyper)core machine, the server sent me 18 or 19 tasks, presumably enough for 8 CPUs to process in the 2- or 3-day deadline he set for the tasks. Needless to say, trying to run 8 1-GB VMs on a 4 GB machine wasn't a pretty sight! So he introduced the limit of 1 job/machine. So basically, that's what S@H is doing, sending enough jobs to keep your machine busy for the duration you've set for cache -- up to the maximum of 100 CPU jobs per machine and 100 GPU jobs per GPU. The Linux laptop takes a fair while to process a MB (15 hrs) and even longer for an AP -- what cache duration have you set for it?
19) Message boards : Number crunching : Migration of BOINC within Linux Installations (Message 1722510)
Posted 5 Sep 2015 by Profile ivan
A couple of random thoughts, on posts so far:
o Mint is derived from Ubuntu, so there shouldn't be any problems with BOINC
o I learnt long ago to have (at least) three partition on a Linux machine; /, /home and swap. Then when you change/upgrade the OS everything goes into / and you don't touch /home so all your files are still there. You just need to add or modify your user on the new system to have /home/<user> as the home directory.
20) Message boards : News : Graduate student Lauren Weiss is searching for planets (Message 1722263)
Posted 4 Sep 2015 by Profile ivan
Will we ever be able to send satellites to other systems? If yes, can we get an educated guess on when it would be possible?

The nearest star system is 4 light-years away. Assuming we could send matter at light-speed, we couldn't expect a signal back for 8+ years. In practice you need to accelerate to speed, then decelerate to match velocity at destination. We need hyperdrive/wormholes/whatever before it's ever going to be feasible, normal propulsion systems won't cut it.


Next 20

Copyright © 2016 University of California