introduction to flash jeremy johnson & frank witmer computing and research services 8 jul 2014

13
Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

Upload: nathaniel-carroll

Post on 22-Dec-2015

215 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

Introduction to Flash

Jeremy Johnson & Frank WitmerComputing and Research Services

8 Jul 2014

Page 2: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

2

Outline

• Flash specifications• What’s New/Changed• How to log on• File movement and policies• Server etiquette• Multi-threaded analysis in Stata & R

Page 3: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

3

Flash Specifications

• Processor: – Intel® Xeon® Processor E5-2667 v2 (25M Cache, 3.30 GHz)

• Max Turbo Frequency 4GHz• 1 Physical CPU – 8 Cores – 16 Threads

• Memory:– 96GB RAM

• 4 x 16GB 1866MHz DDR3 ECC RAM

• Disks– 1.8TB RAID10 Storage

• 4 x 900GB 2.5” 10K RPM SAS Drive

• Video Card– NVIDIA Quadro K2000

• CUDA Parallel-Processing Cores 384• Frame Buffer Memory 2 GB GDDR5

Page 4: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

4

What’s New/Changed• Usernames and passwords

– Flash will use identikeys– External research colleagues now only need one username/password for both VPN and

Flash• Contact Jeremy and Steve to set up a new Person of Interest (POI)

• User profile size limits– Minimize use of My Documents, Desktop, etc.

• 2GB HARD limit– Work from F:\ instead

• 10GB initial limit• If you need more space contact Jeremy

– Flash is not a file/backup server, so please try and keep storage space down to what is needed and remove what you no longer are using as a courtesy to others

• No more mapped drives– Wilbur had a few mapped drives, this is no longer possible– Use the departmental “P” and personal “U” drives

• If you don’t already have access to these, contact Jeremy

Page 5: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

5

What’s New/Changed

• Many new options for connecting– New Remote Desktop App

• iOS (iPad, iPhone), Android (tablets, phones), Win 8, Mac 10.7+

– RemoteApp• Allows you to open Flash applications directly as if installed on your local

device– Network bandwidth minimized since you’re not bringing over full desktop session

– Website• IE works best and opens apps directly• Other browsers work, but will download a .RDP file you have to run for every

app you open

• Mac users need to move to new Remote Desktop App if using 10.7 or above– Download from App Store

Page 6: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

6

What’s New/Changed

• Auto logoff after 14 days on Remote Desktop– Timer resets when you log in

• Auto logoff after 5 days on RemoteApp– Timer resets when you open a RemoteApp

• Auto delete User profile of inactive accounts after 90 Days (Back your stuff up!)– Includes F:\Private\username– Does not include F:\Shared– Does not deny access to Flash, only removes your files

• Next time you log in, new (empty) folders will be recreated automatically

Page 7: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

7

Connecting

• Remote Desktop– http://ibs.colorado.edu/crs/helpdocs/mainremote

• RemoteApp– http://ibs.colorado.edu/crs/helpdocs/remoteapp

• Website– https://flash.colorado.edu

• CU IdentiKey– contact [email protected] to activate

yours for Flash

Page 8: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

8

File System• As with Wilbur, the file system is not backed up• No files will be migrated from Wilbur, so you will need to copy any that

you need manually– If you need to transfer more than 10GB contact Jeremy

• C:\Users\”identikey”\Documents– user read/write permission, but not meant for use– space limited

• F:\Private\”identikey”– user read/write permission

• F:\Shared– user read/write permission, user can add permissions for others

• F:\Shared\Public– read/write permission for all– anyone can delete, so best to use for short-term sharing of files with other

users

Page 9: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

9

Transferring Files To/From Flash

• Copy & paste files– Windows clipboard is shared

• Your local computer shows up in the Windows Explorer– Must be set up to do so through the Remote

Desktop application• P: and U: drives– Automatically mounted, if you don’t see contact

Jeremy

Page 10: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

10

Software

• You may install Stata and R packages to a personal folder– R will ask to use a personal library automatically – Stata requires a few extra steps: https://kb.iu.edu/d/azwf – System wide installs should not be needed

• (These had to be done by Jeremy and required all users to finish and quit the application)

• You are not able to install new software• Installation requests may be made to

[email protected]– New software requests will be reviewed on a case-by-case

basis

Page 11: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

11

Server Etiquette

• Close all unnecessary applications– especially Stata, since IBS is limited to 10 concurrent users

• Log off when not using the server– Click the Start Button then logoff– just closing the Remote Desktop window leaves you

logged in• Monitor your (& other’s) usage using the Task

Manager– right-click the toolbar at the bottom and select Task

Manager then “more details” on the bottom left

Page 12: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

12

Multi-Threaded Analysis in Stata• Stata defaults to using all 16 processors

– if 8 processors are enough, just request 8 (there are often diminishing returns to using more processors)

– `set processors 8`• Check the list of multi-threaded commands supported by Stata

– http://www.stata.com/statamp/statamp.pdf• For advanced users, can use the ‘parallel’ module• Can also use batch mode for do-files

– for a file names bigjob.do, type– stata –b do bigjob– http://www.stata.com/manuals13/gsub.pdf

• Please automatically ‘exit’ do-files to close Stata– See F:\Shared\FlashIntro\stataExample.do

Page 13: Introduction to Flash Jeremy Johnson & Frank Witmer Computing and Research Services 8 Jul 2014

13

Multi-Threaded Analysis in R• No ‘easy’ multi-threaded commands• Use libraries that support parallel processing– e.g. parallel, snow/snowfall, multicore

• See example inF:\Shared\FlashIntro\parallelExample.R

• Additional info on parallel computing packages– http://cran.r-project.org/web/views/HighPerformanceComputing.html