Spiedie Partitions and Features


Table of Contents

  1. Partitions
  2. Features

Partitions

The nodes on Spiedie are partitioned into 4 distinct sets depending on the resources available.

Quick

For programs with short run times. Ideal for testing out your code as queue remains small due to the 10 min time restriction of the partition. Use this partition for jobs that will finish immediately and you need the results right away.

Standard / DefQ

Standard and Default queue are the standard compute nodes used by default. The memory and cpu allocation for this partition are restricted, as some nodes are hardware limited. If you do not need specialized hardware (i.e. GPU) or massive multi-core support and high memory( > 16 cores), these partitions are ideal.

Knights Landing

The Knights Landing partition contains the nodes with 64 cores per socket and 4 threads per core Xeon phi cpus. The partition should be used for extremely multithreaded tasks that really on parallel computation (like openMP) and not message parsing (MPI.

GPU

THe GPU partition nodes contain NVidiaP100 GPU’s for CUDA-accelerated programs. These nodes should only be primarily used for GPU-accelerated programming. Each node contains a 12 GB P100 GPU with 28 available cores and 128 or 256 GB of memory.

Although you are able to run non-GPU accelerated code on these nodes, it is best practice to not do so in order to not bottleneck users trying to use the GPU’s


Features

Spiedie also has specific feature flags that users can use to quickly choose nodes that may fit their needs. Below are all the available features combinations users can request.

You can select nodes by memory size, cpu architectures and GPU.

Features Available Memory(MB) Nodelist Sockets     CPUs per Socket     Threads per CPU  
X5690,SmallMem 48000, compute[021-031] 2 6 1
E5v4,SmallMem 64000+ compute[109-129,146-152,154-160] 2 8 1
X5690,LargeMem 144000+ compute[000-001,016-020] 2 6 1
X5690,MedMem 96000, compute[002-015] 2 6 1
E5v2,MedMem 96000, compute[032-059] 2 8 1
E5v2,LargeMem 128000, compute060 2 8 1
E5v3,MedMem 128000+ compute[061-079] 2 8+ 1
E5v4,MedMem 128000, compute[080-107] 2 12 1
LargeMem,Scalable 256000, compute[143-144] 2 20 1
E5v4,LargeMem 128000, compute[145,153] 2 10 1
Scalable,LargeMem 256000+ compute[161-165] 2 14+ 1
E5v4,P100,MedMem 128000, compute[130-131] 2 14 1
E5v4,P100,LargeMem 256000, compute[132-133] 2 14 1
E5v4,P100 128000, compute142 2 14 1
knl 96000, compute[134-141] 1 64 4

Legend:

SmallMem: Default

MedMem: For jobs requiring > 48 GB of RAM

LargeMem: For jobs requiring > 96 GB of Ram

P100: GPU

knl: Knights Landing CPU’s