| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Android Support Forum
Thread: Performance compared to PC? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 46
|
|
| Author |
|
|
noderaser
Senior Cruncher United States Joined: Jun 6, 2006 Post Count: 297 Status: Offline Project Badges: 
|
Now I got it Rob. You mean linux on my computer. Technically, Android is already a flavor of linux... |
||
|
|
Eric_Kaiser
Veteran Cruncher Germany (Hessen) Joined: May 7, 2013 Post Count: 1047 Status: Offline Project Badges: 
|
@noderaser: Yes, I know but I understood (my mistake): run an origin linux on the android device which is possible on some android devices.
----------------------------------------@Rob: Yes, I've seen this feature in the new boinc version. I'm not sure whether the output will increase if boinc runs in a virtual linux machine hosted on windows or not.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
What I'm hoping to gain most from this is running VINA *AND* CEP2 in there and the VM taking care of snapshotting, so there will be a fairly lossless bootcycle and or crash recovery. Will get to testing it when ready.
|
||
|
|
Eric_Kaiser
Veteran Cruncher Germany (Hessen) Joined: May 7, 2013 Post Count: 1047 Status: Offline Project Badges:
|
That makes sense. Especially for CEP2.
---------------------------------------- |
||
|
|
captainjack
Advanced Cruncher Joined: Apr 14, 2008 Post Count: 147 Status: Offline Project Badges: 
|
On Nov. 26, 2013, SekeRob piqued my curiosity by saying
There's now a BOINC with VBOX included for Windows http://boinc.berkeley.edu/dl/boinc_7.2.33_windows_x86_64_vbox.exe [84MB package]. Not read up, no experience at all, but this could set up a VM with Linux running in it. I'm considering for my Octo "must run W7-64+Office13". The idea to keep 2 cores for W7 and 6 set aside to a Linux-VM has substantial appeal to me. AFAIK, this setup is primarily intended for projects like LHC Test4theory@home where all of the scientific computing is done inside a VM. By setting up the project this way, the programmers can install the O/S of their choice inside the VM then they only have to set up the application software to run under one O/S. Test4theory (CERN) has their own scientific flavor of Linux that they download to run under the VM. The bad news is that it is very tricky to get everything set up and working together properly so participation in the project seems to be low. I was a beta tester over at Test4theory so I have some experience with the setup and have been running it successfully on Win7 and Ubuntu. Back to the main topic of my post, I was curious about Rob's idea of running a VM with Ubuntu on a Windows 7 host and how the performance would be inside the VM. Especially knowing that some types of research performs much better on Linux. So yesterday, I created a VM on my Windows 7 box, gave it 4 CPU's, installed Ubuntu 13.10, installed BOINC 7.2.33 and loaded up 2 MCM and 2 CEP2 jobs. Based on my very non-scientific observation, the MCM tasks seemed to run a bit slower inside the VM (maybe 10-20% longer) and the CEP2 tasks seemed to run much slower (maybe 50-75% longer) than they ran on the Windows host. Fun to play with, but no gain for "set it and forget it" crunching. And here's the disclaimer, your mileage might vary if you have different hardware (and everybody has different hardware). The other advantage of virtualbox is that it lets you install new/different versions of operating systems and try them out without having to set up dual/triple boot or without overlaying your current operating system. This is ideal if you want to take the latest beta version of Ubuntu for a test spin, but don't want to damage your production setup. You could also test different versions of BOINC in a controlled environment without affecting production processing. If you haven't started testing yet and want some help, let me know. If you have started testing, be sure to share your results with us. |
||
|
|
Coleslaw
Veteran Cruncher USA Joined: Mar 29, 2007 Post Count: 1343 Status: Offline Project Badges:
|
captainjack, were you using the 64bit versions of Linux or the 32bit version? 64bit is where the real gain is. Also, the important thing to note about the ARM devices is that most crunchers aren't buying entire farms of these devices for crunching. Most are recycling or re-purposing devices they already have. When you look at it that way, they have much better efficiencies. However, for the professional cruncher, there just aren't viable ARM options yet to replace traditional x86/x64 processors. Time will tell since there are still a lot of entrepreneurs out there coming up with new ideas. If you have them, use them. I just wouldn't recommend buying the ARM just for crunching just yet. I have a couple that were basically phones that either I or my wife have since replaced. I periodically ask if anyone has any Android phones collecting dust, but there are still a lot of people that don't have smart phones. Getting people to give them away is still tough. ;)
----------------------------------------    |
||
|
|
captainjack
Advanced Cruncher Joined: Apr 14, 2008 Post Count: 147 Status: Offline Project Badges:
|
Coleslaw,
The host is Windows 7 Ultimate 64-bit and the guest is Ubuntu 13.10 64-bit. |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges: 
|
alver to answer your question
----------------------------------------an ARMv7 core is roughly the same clock for clock as that found on the Atpm with HT switched off So if all my WU's were valid I'd be getting twice the results of a Netbook out of my Moby for even less power It's no race horse but the combined power of every android powered mobile on the planet make this a huge untapped computing resource for good. Now if we could only get a reasonable level of Valid's,,,, Android's contribution to the research being done here at WCG would be awesome. Note: I just had 2 on the trot  which makes FAHV on Android's success rate currently 33.33%. We need that other wasted 66.66% Dave  |
||
|
|
enels
Senior Cruncher Joined: Apr 25, 2008 Post Count: 286 Status: Offline Project Badges: 
|
One thing to consider when running on battery is the efficiency of the charging:
http://evbatterymonitoring.com/WebHelp/Section_3.htm "Since no chemical or physical process can ever be 100% efficient, more energy is always used to charge the battery than can be recovered from it. Thus, determining the optimum conditions for battery charging grows in importance as the cost of energy increases." "Used correctly, under proper conditions, a modern battery charger will routinely provide overall efficiencies on the order of 85% with a battery of 18-24 cells; 80% with 12 cells and 75% with a 6-cell battery." |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
Maybe these would be a good crunching chip...
----------------------------------------http://www.anandtech.com/show/7724/it-begins-...-64bit8core-opteron-a1100
Crunching for humanity since 2007!
 |
||
|
|
|