Argo Overview, Table of available queues/partitions
Argo overview
Argo is ICTP HPC cluster, comprising of 112 hosts/nodes, with total count of 2100 CPUs, nearly 17 TB of memory, 100Gbps Omnipath or Infiniband interconnects, 1Gbps network and several houndreds of TB of dedicated NFS storage.
The available worker/compute nodes are organised in queues(partitions).
There are three more special cluster nodes: a master node that controls job execution and login nodes argo-login1 and argo-login2, where users login, submit jobs, and compile code.
Jobs can be submitted from argo (argo-login2), argo-login1 and argo-login2.
ssh argo.ictp.it
or
ssh argo-login1.ictp.it
List of available queues/partitions
Queue infomation can be listed with the sinfo command:
$ sinfo -s PARTITION AVAIL TIMELIMIT NODES(A/I/O/T) NODELIST cmsp up 1-00:00:00 30/6/4/40 node[01-16,161-184] esp up 1-00:00:00 16/12/8/36 node[21-56] long* up 1-00:00:00 12/13/7/32 node[61-92] gpu up 1-00:00:00 0/1/1/2 gpu[01-02] serial up 7-00:00:00 0/2/0/2 serial[01-02] testing up 6:00:00 42/19/11/72 node[01-16,61-92,161-184] esp_guest up 1-00:00:00 2/0/0/2 node[28-29]
The principal queue for all users is the long queue, with 32 nodes and 24 hours time limit. It is the default queue, if none is specified in the job.
Dedicated queues cmsp, esp, esp_guest and gpu are available to specific Argo users, upon authorization.
testing queue is special, comprising of several nodes from both long and cmsp queues, but with a short time limit of 6h.
serial queue is specific for serial jobs, while it has a very long time limit of 7 days. Two nodes are in the serial queue.
Generally for all queues the nodes are NOT shared among jobs. The exceptions are queues serial and gpu.
Node features
Overall, Argo is a heterogeneous cluster, with nodes belonging to various generations of Intel CPU microarchitectures. Numerous are nodes of the broadwell and skylake-cascade architecture, followed by Cascade-Lake. Memory size also varies.
For each node we list it's microarchitecture, memory size, and other features in sinfo output:
$ sinfo -N -o "%.20N %.15C %.15P %.40b" NODELIST CPUS(A/I/O/T) PARTITION ACTIVE_FEATURES gpu01 0/40/0/40 gpu 128gb,broadwell-ep,e5-2640v4 gpu02 0/0/16/16 gpu 32gb,sandybridge-ep,e5-2665
node01 40/0/0/40 cmsp omnipart,128gb,broadwell-ep,e5-2640v4 node02 40/0/0/40 cmsp omnipart,128gb,broadwell-ep,e5-2640v4 ... node15 40/0/0/40 cmsp omnipart,128gb,broadwell-ep,e5-2640v4 node16 40/0/0/40 cmsp omnipart,128gb,broadwell-ep,e5-2640v4
node21 40/0/0/40 esp omnipart,128gb,skylake-cascade,silver node22 40/0/0/40 esp omnipart,128gb,skylake-cascade,silver
... node56 0/40/0/40 esp omnipart,128gb,skylake-cascade,silver
node61 64/0/0/64 long* infiniband,187gb,Cascade-Lake,silver-421 node62 0/0/64/64 long* infiniband,187gb,Cascade-Lake,silver-421 ...
node91 64/0/0/64 long* infiniband,187gb,Cascade-Lake,silver-421 node92 64/0/0/64 long* infiniband,187gb,Cascade-Lake,silver-421
node161 0/0/40/40 cmsp omnipart,192,broadwell-ep,e5-2640v4 node162 0/40/0/40 cmsp omnipart,192,broadwell-ep,e5-2640v4 ...
node183 40/0/0/40 cmsp omnipart,192,broadwell-ep,e5-2640v4 node184 0/40/0/40 cmsp omnipart,192,broadwell-ep,e5-2640v4
serial01 0/16/0/16 serial 32gb,sandybridge-ep,e5-2650 serial02 0/16/0/16 serial 32gb,sandybridge-ep,e5-2650
Within each queue, nodes are homogeneous in terms of all of their features. .
All nodes are networked together with 1 Gbps ethernet links spanning multiple switches.
Access to storage is also done through the Gigabit ethernet network.
Nodes within each queue are also networked together in a low-latency fabric for MPI communication, thanks to Infiniband or Omni-Path technology, performing at 100 Gbps (HDR).
Table of available queues and nodes
Table below summarises queue, nodes and their characteristics
Queue/Partition |
Access Policy | Notes |
---|---|---|
long | All users |
- Allows allocations of a maximum of 10 nodes for running parallel jobs. |
testing | All users | |
serial | All users |
- ONLY FOR single core, cpu (or task) jobs; Parallel or MPI jobs will NOT WORK. - Up-to a maximum of 7 running independent serial jobs are allowed. - Resources are over-subscribed (the nodes are shared among jobs and users). |
OTHER Dedicated Queues |
||
cmsp | Special authorization needed | |
esp |
Special authorization needed | |
gpu
|
Special authorization needed |
- several GPU Accelerators are available of the type:
Nvidia Tesla K40
Nvidia Tesla P100
- resources are over-subscribed (the nodes are shared among jobs.).
|
Table with technical details
Queue/ Partition |
Max walltime (h) | Node range | Micro-architecture | Cores |
Ram per core (GB/c) |
Total nodes |
Total cores |
Ram per node (GB) |
---|---|---|---|---|---|---|---|---|
long | 24:00 |
node[61...92] |
Cascade-Lake |
16 |
11.69 |
32 |
512 |
187 |
testing | 6:00 | node[01...16] node[161...184] node[61...92] |
see cmsp queue see cmsp queue see long queue |
- |
- - - |
- - - |
- - - |
- - - |
cmsp | 24:00 |
node[01...16] |
Broadwell Broadwell |
20 |
6.4 |
16 |
320 480 |
128 |
serial | 168:00 | serial[01...02] | Sandybridge | 16 | 2 | 2 | 32 | 32 |
esp | 24:00 | node[21...56] | skylake-cascade |
20 |
6.4 | 36 | 720 | 128 |
gpu | 24:00 |
gpu01 |
Broadwell+ 2*gp100 |
20 |
6.4 |
1 |
16 |
128 |