-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
Got Mosix working!
- Thread starter JchrisW
- Start date
Orange Kid
Elite Member
What kinds of chips and speeds?
What kind of key rate are you getting using mosix?
This sounds interesting--could Seti be run this way?
😎 😎
What kind of key rate are you getting using mosix?
This sounds interesting--could Seti be run this way?
😎 😎
Yes, Linux. Using RH6.2 on all machines.
Machine 1 is a Celeron 366 @ 550
Machine 2 is a P200, and 3 is a P133.
Once I iron out the few bugs that exist, I'm going to move the rest of my cluster over.
All three machines are running on a 100Mbit switched backbone, but bandwidth isn't an issue with RC5.
Orange Kid : I'll let you know exact keyrates in the morning after a good night of cracking.
I'll also post pics in the morning of the setup.
All in all, Mosix seems like a very cool way of setting up distributed computing. There are a few setup issue to get around, but once you're setup, it's all foolproof. (Well, so far 😛).
Russ, or anyone else for that matter, let me know if you ever want help setting something like this up.
-Jordan
Machine 1 is a Celeron 366 @ 550
Machine 2 is a P200, and 3 is a P133.
Once I iron out the few bugs that exist, I'm going to move the rest of my cluster over.
All three machines are running on a 100Mbit switched backbone, but bandwidth isn't an issue with RC5.
Orange Kid : I'll let you know exact keyrates in the morning after a good night of cracking.
I'll also post pics in the morning of the setup.
All in all, Mosix seems like a very cool way of setting up distributed computing. There are a few setup issue to get around, but once you're setup, it's all foolproof. (Well, so far 😛).
Russ, or anyone else for that matter, let me know if you ever want help setting something like this up.
-Jordan
SilverBack
Golden Member
Hmmm
I wonder what 4 TBird GHZ machines would be capable of?
Hmmm
🙂
I wonder what 4 TBird GHZ machines would be capable of?
Hmmm
🙂
This is exactly the kind of thing that makes RC5 cracking a worthwhile project. 🙂 Individually and as a group we all learn a little bit more about computing. The gifted ones amongst actually advance the science a little.
I think that there are quite a few of us that would like to have all the details of this published on one of the team's web sites. 😀
Good day to you and thankyou 😀
I think that there are quite a few of us that would like to have all the details of this published on one of the team's web sites. 😀
Good day to you and thankyou 😀
I know you did not do this just for the joy of RC5 stats but...
Would this set up not be slower then the output of the 3 machines just crunching by them selves? Install klinux on the boxes for a stripped down, quick machine and let it go.
Like I said I know there is more to life then RC5 and that playing around with the cluster is a lot better then watching a dot move across the screen. I guess I am interested in just how much overhead is used with setting up the cluster over running the machines by themselves.
S.
Would this set up not be slower then the output of the 3 machines just crunching by them selves? Install klinux on the boxes for a stripped down, quick machine and let it go.
Like I said I know there is more to life then RC5 and that playing around with the cluster is a lot better then watching a dot move across the screen. I guess I am interested in just how much overhead is used with setting up the cluster over running the machines by themselves.
S.
Dantoo :
I'm working on documentation of how I did it as we speak, I'll keep you updated on how it goes.
Sloth :
I can understand your concern regarding processor overhead and reduced keyrate, but believe me, the difference is minimal. On a three machine cluster, Cel 550, P200, and a P133, one could expect a keyrate of approximately 1.95Mkeys/sec (1.5Mkeys + 180Kkeys + 275Kkeys). Using my current setup, I'm getting approximately 1.81 to 1.85 Mkeys/sec. So there is some processor overhead, but it's fairly minimal.
murdmath :
The only thing I needed to modify with the RC5 client was to disable multithreaded operation. (Menu 4 --> 2 I believe). You then start running several RC5 processes, and "migrate" them over to the other systems in the cluster. Mosix handles all of the rest, and gives me one nice keyrate to work with.
Hope this helps for now,
Jordan
I'm working on documentation of how I did it as we speak, I'll keep you updated on how it goes.
Sloth :
I can understand your concern regarding processor overhead and reduced keyrate, but believe me, the difference is minimal. On a three machine cluster, Cel 550, P200, and a P133, one could expect a keyrate of approximately 1.95Mkeys/sec (1.5Mkeys + 180Kkeys + 275Kkeys). Using my current setup, I'm getting approximately 1.81 to 1.85 Mkeys/sec. So there is some processor overhead, but it's fairly minimal.
murdmath :
The only thing I needed to modify with the RC5 client was to disable multithreaded operation. (Menu 4 --> 2 I believe). You then start running several RC5 processes, and "migrate" them over to the other systems in the cluster. Mosix handles all of the rest, and gives me one nice keyrate to work with.
Hope this helps for now,
Jordan
A co-worker and linux mentor of mine is running Mosix at home on his farm (with about 5 machines with it)... He has seti on his nodes. I have been considering setting it up myself...
I think the benefits come more with the cluster's ability to dynamically handle the different loads (and provide administration of them), moving the processes back and forth depending on what a particular node is doing at the time (unless you choose to run the processes statically on that machine).
So for example, if you have a slow machine that you're not normally cracking or crunching on (for shame! But maybe because it's normally used as a game server or DNS server something... 😉). Assuming you do crack or crunch on your primary machine, you may suddenly need to run some memory-hungry app that'll really push your distributed process back into the background big time dropping your keyrate or flops down to almost nothing. Well your seti or RC5 process will migrate from your machine to the node that's not doing as much (eg., the gamebox or DNS machine that's not normally cracking or crunching). Once you're done with your work, the process will move back to your machine because it's faster.
This way, you won't lose precious processing time when you have to actually do some work on your herd... 😉
I think the benefits come more with the cluster's ability to dynamically handle the different loads (and provide administration of them), moving the processes back and forth depending on what a particular node is doing at the time (unless you choose to run the processes statically on that machine).
So for example, if you have a slow machine that you're not normally cracking or crunching on (for shame! But maybe because it's normally used as a game server or DNS server something... 😉). Assuming you do crack or crunch on your primary machine, you may suddenly need to run some memory-hungry app that'll really push your distributed process back into the background big time dropping your keyrate or flops down to almost nothing. Well your seti or RC5 process will migrate from your machine to the node that's not doing as much (eg., the gamebox or DNS machine that's not normally cracking or crunching). Once you're done with your work, the process will move back to your machine because it's faster.
This way, you won't lose precious processing time when you have to actually do some work on your herd... 😉
Poof :
You hit the nail right on the head. The main advantage of Mosix does seem to be the dynamic handling of processes. I've done some load testing with an Apache server running on one of the machines, and it works beautifully. RC5 immediately backs down and allows needed CPU resources for the needed server, then migrates back over.
Jordan
You hit the nail right on the head. The main advantage of Mosix does seem to be the dynamic handling of processes. I've done some load testing with an Apache server running on one of the machines, and it works beautifully. RC5 immediately backs down and allows needed CPU resources for the needed server, then migrates back over.
Jordan
Oh oH OH OH OhOhohoh oHOH I think it is alive!!!!!!!!!!!!!
If I'm correct that part in Dnet config I need to change is the part about it automatically detecting the number of processors. Change that to 0. Ok. So, I have two systems, do I need to run two instances of Dnet?
Murdmath
If I'm correct that part in Dnet config I need to change is the part about it automatically detecting the number of processors. Change that to 0. Ok. So, I have two systems, do I need to run two instances of Dnet?
Murdmath
murdmath :
Actually, no, that's not the option. It's the one before that I believe, I don't have a client nearby right now, otherwise I'd check. But yes, you do change it to 0. After you have that working, try running 2-3 instances of dnetc (or rc5des depending on client), and then use the migrate tool to migrate the processes over to the other machine(s). (migrate <PID> <machine>).
Good Luck,
Jordan
Actually, no, that's not the option. It's the one before that I believe, I don't have a client nearby right now, otherwise I'd check. But yes, you do change it to 0. After you have that working, try running 2-3 instances of dnetc (or rc5des depending on client), and then use the migrate tool to migrate the processes over to the other machine(s). (migrate <PID> <machine>).
Good Luck,
Jordan
I can't find the option that you are talking about. I have 3 systems running it right now. T-Bird 750@1G Ppro 180@200 and a P75@120. I have 3 instances of dnet running on one machine and while looking at the monitor It looks like it has all three are running at full speed.
murdmath
murdmath
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 24K
-
-
