This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision Last revision Both sides next revision | ||
cluster:133 [2014/08/11 14:21] hmeij [Ideas] |
cluster:133 [2015/03/18 18:21] hmeij [High Core Count - Low Memory Footprint] |
||
---|---|---|---|
Line 10: | Line 10: | ||
We're on the cusp of a new era! | We're on the cusp of a new era! | ||
+ | |||
+ | Other solutions than the one described below | ||
+ | |||
+ | * 4U/288 cores [[http:// | ||
==== Ideas ==== | ==== Ideas ==== | ||
Line 25: | Line 29: | ||
[[http:// | [[http:// | ||
- | * 28 blades, 4 nodes per blade, each node with | + | * 28 blades, 112 nodes, 4 nodes per blade, each node with |
* 1x Atom C2750 8 core 2.4 Ghz chip | * 1x Atom C2750 8 core 2.4 Ghz chip | ||
- | * up 32 GB ram (4 GB per core, way above what's needed | + | * up 32 GB ram (4 GB per core, way above what's needed) |
* 1x 2.5" disk | * 1x 2.5" disk | ||
* Virtual Media Over LAN (Virtual USB Floppy / CD and Drive Redirection) | * Virtual Media Over LAN (Virtual USB Floppy / CD and Drive Redirection) | ||
* Do these PXE boot? How to get OS on drives? | * Do these PXE boot? How to get OS on drives? | ||
+ | |||
+ | * Other thoughts | ||
+ | * With that many nodes, /home would probably not be mounted | ||
+ | * So users would have to stage job data in / | ||
+ | * ... via scp from a target host | ||
+ | |||
+ | |||
==== Slurm ==== | ==== Slurm ==== | ||
Line 36: | Line 47: | ||
And then we need something that can handle ten of thousand of jobs if we acquire such a dense core platform. | And then we need something that can handle ten of thousand of jobs if we acquire such a dense core platform. | ||
- | Enter [[https:// | + | Enter [[https:// |
Now we're talking. | Now we're talking. | ||
+ | |||
+ | Notes on Slurm are [[cluster: | ||
==== Problem ==== | ==== Problem ==== |