If I get a WU that uses the a1 core via the Linux SMP beta client, can I switch to the Windows SMP client to work on the same WU, or must it be completed with the Linux client?
Background:
Since the Windows SMP client can't work on any of the a2 WUs, I was hoping that I would be able to make it get only a2 WUs. My system that's currently running the Windows SMP client runs faster than my Linux version because it's on a different machine. (~ 1800 PPD vs. ~1200 PPD). So I was thinking that if I moved the WUs that the Linux SMP client gets that only uses the a1 core that I would move it over to my other system that will be able to do the a1 WU faster, in hopes that I would be able to only load up the Linux system with a2 WUs.
Can I/am I allowed to make such a move (provided that I am still within the WU deadlines)?
Or must it be that if the Linux SMP client gets an a1 WU, that ONLY the Linux client can run that WU?
I might end up with a bit of stack up, but if I am near completion of one, maybe I can move it over to my other system that does the a1 WUs faster; while my Linux machine picks up a new WU.
Just thinking/wondering out loud. That's all.
I'm otherwise not changing the program or anything, just moving the WUs around to take advantage of the speed/architectural differences between systems.
general quesiton about WUs
Moderators: Site Moderators, FAHC Science Team
-
anandhanju
- Posts: 522
- Joined: Mon Dec 03, 2007 4:33 am
- Location: Australia
Re: general quesiton about WUs
Until someone answers, here's a related topic: viewtopic.php?f=16&t=1804&start=0
-
alpha754293
- Posts: 383
- Joined: Sun Jan 18, 2009 1:13 am
Re: general quesiton about WUs
Would that affect the underlying science of it though even though from what I can tell, both Linux and Windows cores are 1.74? by making such a switch?
-
7im
- Posts: 10179
- Joined: Thu Nov 29, 2007 4:30 pm
- Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
- Location: Arizona
- Contact:
Re: general quesiton about WUs
The wine answer from the other thread seemed appropriate.
This is also the equivalent of sneakernetting a work unit. Please read up on the topic in the fah wiki, paying attention to the Machine ID info. If you don't send a work unit back from the same ID that received it, the servers might assume the WU was lost, and resend the same work unit again, but you only get credit once. Also read up on how to convert a Windows to Linux Machine ID elsewhere in the wiki, something about Little Indians.
This is also the equivalent of sneakernetting a work unit. Please read up on the topic in the fah wiki, paying attention to the Machine ID info. If you don't send a work unit back from the same ID that received it, the servers might assume the WU was lost, and resend the same work unit again, but you only get credit once. Also read up on how to convert a Windows to Linux Machine ID elsewhere in the wiki, something about Little Indians.
How to provide enough information to get helpful support
Tell me and I forget. Teach me and I remember. Involve me and I learn.
Tell me and I forget. Teach me and I remember. Involve me and I learn.
-
alpha754293
- Posts: 383
- Joined: Sun Jan 18, 2009 1:13 am
Re: general quesiton about WUs
Are the WUs really keyed to machine IDs though?
You might be able to run the Windows client in Linux via WINE, but I can't run everything else that I've got the same way. (Which is why it isn't a Linux system to begin with).
I think that people are misunderstanding the question -- I've got two systems -- one's quad AMD Opteron 880 (2.4 GHz, dual-core) that's running Linux with the Linux SMP client, and the other is my desktop Q9550 (2.83 GHz, quad-core) running Windows XP with the Windows SMP client.
the Q9550 gets about 1800 PPD while the quad Opteron 880 gets about 1200 PPD when the WU uses the a1 core.
So, the idea is if my linux machine gets a WU that uses the a1 core; I would make it the next WU for the Q9550 by putting the WU data into the Window SMP client directory. (I'd interrupt the current run by changing the flag so that it'll run with "-smp -pause" so that when the current WU is complete, it'd pause the system, pull the WU that the Linux client picked up, and hopefully continue on with that.)
That way, hopefully the Linux client would be able to get a new WU that uses the better suited a2 core on the Linux machine.
It isn't that I'm running the same WU twice. I might start working on it while waiting for the Windows WU to finish or be close to finishing, but I wouldn't be submitting the results twice.
TECHNICALLY the WU hasn't changed, I've just moved it from one machine to another, and used to the Windows client to finish it off.
Can I do that? Would there be any adverse effects in terms of the science? (I'm doing this so that I can take advantage of the faster a1 turnaround (~50% faster))
You might be able to run the Windows client in Linux via WINE, but I can't run everything else that I've got the same way. (Which is why it isn't a Linux system to begin with).
I think that people are misunderstanding the question -- I've got two systems -- one's quad AMD Opteron 880 (2.4 GHz, dual-core) that's running Linux with the Linux SMP client, and the other is my desktop Q9550 (2.83 GHz, quad-core) running Windows XP with the Windows SMP client.
the Q9550 gets about 1800 PPD while the quad Opteron 880 gets about 1200 PPD when the WU uses the a1 core.
So, the idea is if my linux machine gets a WU that uses the a1 core; I would make it the next WU for the Q9550 by putting the WU data into the Window SMP client directory. (I'd interrupt the current run by changing the flag so that it'll run with "-smp -pause" so that when the current WU is complete, it'd pause the system, pull the WU that the Linux client picked up, and hopefully continue on with that.)
That way, hopefully the Linux client would be able to get a new WU that uses the better suited a2 core on the Linux machine.
It isn't that I'm running the same WU twice. I might start working on it while waiting for the Windows WU to finish or be close to finishing, but I wouldn't be submitting the results twice.
TECHNICALLY the WU hasn't changed, I've just moved it from one machine to another, and used to the Windows client to finish it off.
Can I do that? Would there be any adverse effects in terms of the science? (I'm doing this so that I can take advantage of the faster a1 turnaround (~50% faster))
-
alpha754293
- Posts: 383
- Joined: Sun Jan 18, 2009 1:13 am
Re: general quesiton about WUs
according to the link -- it's Windows in Wine in Linux.7im wrote:The wine answer from the other thread seemed appropriate.
This is also the equivalent of sneakernetting a work unit. Please read up on the topic in the fah wiki, paying attention to the Machine ID info. If you don't send a work unit back from the same ID that received it, the servers might assume the WU was lost, and resend the same work unit again, but you only get credit once. Also read up on how to convert a Windows to Linux Machine ID elsewhere in the wiki, something about Little Indians.
Here it's Linux in Windows (sans cygwin).
-
codysluder
- Posts: 1024
- Joined: Sun Dec 02, 2007 12:43 pm
Re: general quesiton about WUs
I'm assuming that you're not talking about wine, but you're talking about sneakernetting a WU from a Linux client to a Windows client. I doubt that anybody really knows the answer.
I think there are a lot of risks but if you're ok with them, try it. There's a chance it might work and there's a chance you might corrupt the science. Since you'd also have to finish the current WU (with the -pause) there's a chance that you'd miss a deadline. If it does work, there's also the potential problem that all sneakernetting faces: possible duplicate assignments.
I think there are a lot of risks but if you're ok with them, try it. There's a chance it might work and there's a chance you might corrupt the science. Since you'd also have to finish the current WU (with the -pause) there's a chance that you'd miss a deadline. If it does work, there's also the potential problem that all sneakernetting faces: possible duplicate assignments.