Difference between revisions of "Resources"

From DeepSense Docs
Jump to: navigation, search
(Update cloud resources and delete cluster information)
Line 1: Line 1:
{|class="wikitable" style="text-align: center; color: black; font-style:bold"
+
 
|+'''<span style="font-size:100%>Compute Resources</span>'''
+
'''<span style="font-size:120%>AWS EC2 Cloud Compute Resources</span>'''
 +
{|class="wikitable" style="text-align: left; color: black; font-style:bold"
 +
|+'''<span style="font-size:100%> Available accelerated computing instances</span>'''
 
|-
 
|-
|'''Compute Hosts'''
+
|'''Instance'''
|'''#'''
+
|'''GPUs'''
|'''Memory'''
+
|'''vCPU'''
|'''CPU'''
+
|'''Memory (GiB)'''
|'''GPU'''
+
|'''GPU Memory (GiB)'''
|'''Interconnect'''
+
|'''Instance Storage (GB)'''
|'''Operation System'''
+
|'''Network Performance'''
 +
|'''EBS Bandwidth (Gbps)'''
 
|-
 
|-
|Large Memory
+
|g4dn.xlarge
|20
+
|1
|512GB
+
|4
|style="text-align:left" | 20 Core IBM Power8 2.8GHz (ds-cmlm-01,...,ds-cmlm-20)
+
|16
|N/A
+
|16
|Mellanox SN 2700 –100GB Cluster Interconnect Switch
+
|1 x 125 NVMe SSD
|Redhat Enterprise 7.7
+
|Up to 25
 +
|Up to 3.5
 
|-
 
|-
|Huge Memory
+
|g4dn.2xlarge
|2
+
|1
|1TB
+
|8
|style="text-align:left" |20 Core IBM Power8 3.4GHz (ds-cmhm-03,ds-cmhm-04)
+
|32
|N/A
+
|16
|Mellanox SN 2700 –100GB Cluster Interconnect Switch
+
|1 x 225 NVMe SSD
|Redhat Enterprise 7.7
+
|Up to 25
 +
|Up to 3.5
 +
|}
 +
 
 +
 
 +
{|class="wikitable" style="text-align: left; color: black; font-style:bold"
 +
|+'''<span style="font-size:100%> Available general purpose computing instances</span>'''
 
|-
 
|-
|GPU
+
|'''Instance'''
|10
+
|'''vCPU'''
|512GB
+
|'''Memory (GiB)'''
|style="text-align:left" |20 Core IBM Power8NVL 4.0GHz (ds-cmgpu-01,...,ds-cmgpu-10)
+
|'''Instance Storage (GB)'''
|2xP100
+
|'''Network Performance'''
|Mellanox SN 2700 –100GB Cluster Interconnect Switch
+
|-
|Redhat Enterprise 7.7
+
| m5.xlarge
 +
| 4
 +
| 16
 +
| EBS-Only
 +
| Up to 10
 +
|-
 +
| m5.2xlarge
 +
|8
 +
|32
 +
|EBS-Only
 +
|Up to 10
 
|}
 
|}
  
Largest LSF job that can be hosted on a "512GB" node is 450GB,
 
''i.e.'' <code>bsub -R "rusage[mem=450GB]"</code>.
 
Largest job that can be hosted on a "1TB" node is 960GB,
 
''i.e.'' <code>bsub -R "rusage[mem=960GB]"</code>.
 
  
{|class="wikitable" style="text-align: center; color: black; font-style:bold"
+
 
|+'''<span style="font-size:100%>Storage Resources</span>'''
+
 
 +
'''<span style="font-size:120%>AWS SageMaker Notebook Instance Resources</span>'''
 +
{|class="wikitable" style="text-align: left; color: black; font-style:bold"
 +
|+'''<span style="font-size:100%> Available accelerated computing instances</span>'''
 
|-
 
|-
|'''Name'''
+
|'''Instance'''
|'''Location'''
+
|'''GPUs'''
|'''Quota'''
+
|'''vCPU'''
|'''Function'''
+
|'''Memory (GiB)'''
 +
|'''GPU Memory (GiB)'''
 +
|'''Instance Storage (GB)'''
 +
|'''Network Performance'''
 +
|'''EBS Bandwidth (Gbps)'''
 
|-
 
|-
|Home
+
|ml.g4dn.xlarge (fast launch)
| /dshome
+
|1
| 1TB
+
|4
| Personal scripts/code/test data
+
|16
 +
|16
 +
|Elastic EBS
 +
|Up to 25
 +
|Up to 3.5
 
|-
 
|-
|Scratch
+
|ml.g4dn.2xlarge
| /scratch
+
|1
| 2TB
+
|8
| Temporary space
+
|32
 +
|16
 +
|Elastic EBS
 +
|Up to 25
 +
|Up to 3.5
 
|-
 
|-
|Data
+
| m5.xlarge (fast launch)
| /data
+
|No GPU
| 2TB
+
| 4
| Project group data
+
| 16
 +
|No GPU
 +
| EBS-Only
 +
| Up to 10
 +
| Up to 4,750
 
|}
 
|}
See [[Storage policies]] for more information about the available shared file systems, storage policies, and backup policies.
+
 
 +
Contact support for more options.

Revision as of 21:04, 24 February 2023

AWS EC2 Cloud Compute Resources

Available accelerated computing instances
Instance GPUs vCPU Memory (GiB) GPU Memory (GiB) Instance Storage (GB) Network Performance EBS Bandwidth (Gbps)
g4dn.xlarge 1 4 16 16 1 x 125 NVMe SSD Up to 25 Up to 3.5
g4dn.2xlarge 1 8 32 16 1 x 225 NVMe SSD Up to 25 Up to 3.5


Available general purpose computing instances
Instance vCPU Memory (GiB) Instance Storage (GB) Network Performance
m5.xlarge 4 16 EBS-Only Up to 10
m5.2xlarge 8 32 EBS-Only Up to 10



AWS SageMaker Notebook Instance Resources

Available accelerated computing instances
Instance GPUs vCPU Memory (GiB) GPU Memory (GiB) Instance Storage (GB) Network Performance EBS Bandwidth (Gbps)
ml.g4dn.xlarge (fast launch) 1 4 16 16 Elastic EBS Up to 25 Up to 3.5
ml.g4dn.2xlarge 1 8 32 16 Elastic EBS Up to 25 Up to 3.5
m5.xlarge (fast launch) No GPU 4 16 No GPU EBS-Only Up to 10 Up to 4,750

Contact support for more options.