All SYMMIC users may take advantage of the cloud version of the product, by subscribing to
xSYMMIC in the Cloud
from their Amazon Web Services account.
The following guide explains how to get started using the cloud product on Amazon Web Services,
in both video and written format.
Click the Sign Up button on the AWS landing page
and register for an account with your email address. Then enter your contact and payment information
to verify your identity. All AWS usage fees will be charged to the credit card you provide at sign up.
Finally, select the Basic Plan for free support with AWS.
The AWS console provides access to all of the services available on the Amazon cloud, plus links to account
billing and support. The console is located at console.aws.amazon.com.
Computational resources for running xSYMMIC are accessed from the Amazon Elastic Compute Cloud (EC2).
The EC2 dashboard is used to select and launch instances of computing resources for running simulations
in the cloud. The EC2 dashboard can be reached by Services from the top menu bar,
and then selecting EC2 under the Compute category.
Go to Instances to see what machines are running under your account, and to
launch new instances or terminate existing ones. You are billed hourly for instances that are running, or
that are stopped but not yet terminated. All billing is handled through AWS.
Use the Launch Instance button to start the Launch Wizard.
Amazon classifies instance types as general purpose (e.g. m4 and m5),
memory-optimized (e.g. r4 and x1) and compute-optimized (e.g. c4 and c5).
The correct choice of EC2 instance type may significantly speed up simulations in the cloud
compared to desktop simulations.
In the following table, cloud instances are compared to a desktop workstation for
steady-state solution of the FETbig.xml device with 1.8 million nodes.
All cases were run from the SYMMIC GUI using Remote Run at random times during the workday.
Solve time (in seconds) is the duration reported by the xSYMMIC solver on the cloud instance.
Wall time (in seconds) is from Remote Run on the desktop until the dialog
reports the solution has been downloaded. The solution file was 106 Mbytes in size.
Performance of xSYMMIC in the Cloud Version 3.0.0
Compute Location
Instance Type
Processor
Cores
RAM(Gb)
Spot($/hr)
Solve(s)
Wall Time(s)
AWS us-east (N. VA)
r4.large
Xeon E5-2686v4 2.3 GHz
1
15
0.04
167
180
AWS us-east (N. VA)
r4.xlarge
Xeon E5-2686v4 2.3 GHz
2
30
0.08
94.2
106
Desktop (Natick, MA)
--
Xeon E5-1620 3.6 GHz
4
32
--
92.4
95
AWS us-east (N. VA)
m4.2xlarge
Xeon E5-2686v4 2.3 GHz
4
32
0.13
56
68
AWS us-east (N. VA)
m5.2xlarge
Xeon Platinum 8175 2.3 GHz
4
32
0.14
42.8
58
AWS us-east (N. VA)
r4.2xlarge
Xeon E5-2686v4 2.3 GHz
4
64
0.15
56.7
72
AWS us-east (N. VA)
c4.2xlarge
Xeon E5-2666v3 2.9 GHz
4
15
0.13
48.7
62
AWS us-east (N. VA)
c5.2xlarge
Xeon Platinum 8124 3 GHz
4
16
0.16
48
58
AWS us-east (N. VA)
r4.8xlarge
Xeon E5-2686v4 2.3 GHz
16
244
0.65
28.3
42
AWS us-east (N. VA)
c4.8xlarge
Xeon E5-2666v3 2.9 GHz
18
60
0.55
26.3
40
AWS us-east (N. VA)
c5.9xlarge
Xeon Platinum 8124 3 GHz
18
72
0.55
23.7
36
AWS us-east (N. VA)
m4.10xlarge
Xeon E5-2686v4 2.3 GHz
20
160
'2
30.1
43
AWS us-east (N. VA)
m4.16xlarge
Xeon E5-2686v4 2.3 GHz
32
256
'3.2
28.6
43
AWS us-east (N. VA)
r4.16xlarge
Xeon E5-2686v4 2.3 GHz
32
488
1
27.3
42
AWS us-east (N. VA)
x1.16xlarge
Xeon E7-8880v3 2.3 GHz
32
976
2
31.5
42
AWS us-east (N. VA)
m5.24xlarge
Xeon Platinum 8175 2.3 GHz
48
384
'4.6
20.5
33
'On demand pricing was used because spot pricing was not available at the time for these instances.
To get the AMI for
xSYMMIC in the Cloud in the Launch Wizard, select the
AWS Marketplace from the left-hand menu. In the search box type "xSYMMIC"
to find the AMI, then press the Select button next to it.
This machine image comes with xSYMMIC
already installed and licensed for cloud use.
An additional ten cents per hour will be added to the price of your
instance for the use of this AMI.
Select the desired machine-type, paying particular attention to the amount of RAM needed for
the simulation and the number of CPUs. (The listed value is twice the number of physical cores.)
In general, more CPUs will give better performance and incur higher costs. For pricing information
see the
Pricing link on the right side of
the EC2 dashboard. For more performance information see the discussion of Instance Types above.
For lower pricing, try checking the Request Spot instances Purchasing option. If surplus
capacity is available, the page will then show availability zones and spot rates. Otherwise,
uncheck the box to receive on-demand pricing. Make sure the
Auto-assign Public IP option is enabled, since the public IP address is needed to communicate
with the instance from Remote Run in SYMMIC.
Create a new security group with TCP input port 22 open for SSH administration.
For the source, chose My IP to allow only the desktop to access the instance.
Leave output ports open (the default) to allow the instance to access the license server.
The final step is to create a key pair for secure commincations with the instance.
Save the private key file. The instance will be booted with the matching
public key already installed. Use the private key file in the Remote Run dialog
to authenticate SSH communications with the instance once it is running.
From the SYMMIC application on the desktop, open the Remote Run dialog. Enter the public IP address
of the instance in the dialog. The Public IP is shown on the Instance page in the EC2 Dashboard.
If no Public IP is displayed when the instance is running, then it was not launched with the
Auto-assign Public IP option enabled, so terminate it and start over.
In the Remote Run dialog, enter ec2-user for the username and leave the working directory blank.
In the Private key file box, enter the
filename of the private key from the last step in the AWS Launch Wizard. Click the
Connect button and accept the fingerprint. (Copy the fingerprint to the Fingerprint
edit box to prevent this challenge in the future.) Then use the Launch button to upload
the files, run the simulation, and download the results. The progress will be reported
at the bottom of the Remote Run dialog.