JupyterLab Interface

① Main screen
② Left side screen
③ Switching Side Bar
④ LAUNCHER
⑤ Countdown to 60 days of continuous activation
- See Matlantis shutdown conditions for details
⑥ PFP Load Status / GPU load status for PFP inference (for each tenant)
⑦ Translate to Dashboard screen
- Dashboard screen introduced on another page
⑧ CPU and RAM usage (for each user)
⑨ Disk usage (for each user)

Switching Left Side screen

You can select the following (1) ~ (3) settings.

① File Browser
LAUNCHER
- Basic Launcher
- Package Launcher
- Example Launcher

A list of files in an exclusive environment for each user. It can be uploaded using drag & drop.

Trash bin is located at the root “/” directory.
Files deleted by right-click will be moved to Trash.

Note:

Files removed by “rm” on CUI will not be moved to Trash
Files in Trash will be counted as a part of disk space (this is also described later)
There is no automatic removal of old files; User needs to empty Trash by hand to free the disk space

Matlantis Guidebook > Other Notebook features > How to use Trash

"OPEN TABS" : Files that are open.
"KERNELS": Files with active python kernel.
# If a job is no longer needed and still listed under the active kernels, you should shut it down to free up the RAM (see also Memory usage)
"TERMINALS": Terminal that are open.

Shows outline of the file written by Markdown system Recommended to use Markdown to document notebook eg. : Matlantis Examples > Basic> Welcome

It can visualize usage of tokens consumed in PFP inference (on upper part of JupyterLab interface(6))

# See also here for more information about tokens.

Token Consumption Rate
- This is the consumption rate ratio to the token recovery rate. (consumption/token recovery rate [%])
- Displayed in a stacked graph for each user.
- Less than 100%:
  Tokens are not consumed because the consumption rate is less than the recovery rate.
- More than 100%:
  Tokens are consumed and Token Availability is reduced until 0% is reached (burst state).
Token Availability
- The percentage of tokens available out of the token capacity per tenant.
- At 100%: Tokens are fully stored and will not recover any more.
- At 0%: The tokens have been used up and only the token recovery rate can be calculated.

Matlantis Guidebook > About pfp-api-client > About Token System and Priority Function

Click "Open group drive" to open the Group Drive browser.
Drag and drop to copy and paste files between Notebook user spece and Group Drive.

GroupDrive also opens from the Basic Launcher.

These are displayed in the lower left corner of the Notebook env. screen. (Lower left of JupyterLab main screen)
- (1) CPU usage
- (2) Memory usage
- (3) Disk usage
All these usages will be updated at real-time. See also here for your information, on how to save on memory and disk usage.
Upgrade may be possible if specifications are insufficient. Please contact us.

If the memory usage and disk usage reach their maximum values, it may cause problems such as not being able to open the Notebook environment. Please check the amount of usage during running calculation.
How to save on Memory usage:
- Close unused Notebook (tabs).
- Shut Down any unused kernels in the Running Terminals and Kernels view
How to save on Disk usage:
- Empty the trash regularly.
- Move data to GroupDrive.
What to do if Notebook env. cannot be opened:
- If the Dashboard screen opens, use “Stop Application” to stop the Matlantis environment. If not, try “Force Stop Application” (Dashboard screen> Advanced)
- Please try to see if you can start the app from “Launch Application with Initial Layout” (Dashboard screen> Advanced) in the center.
- If the problem persists, please contact us via contact.

It is consumed by performing calculations using PFP (using computational resource for inference)
- Token consumption is determined by a number of atoms and neighbors of the system to calculate.
The amount of tokens recovered (token recovery rate) is set by the tenant per contracted plan.
Tokens are stored up to token capacity, and token capacity is set according to the token recovery rate.

Matlantis Guidebook > About pfp-api-client > About Token System and Priority Function

Illustration:

Tokens accumulate until the capacity (bucket) is full.
The remaining percentage of accumulated tokens is displayed as Token Availability, and the PFP calculation consumes tokens from the bucket (burst).
When the token consumption rate exceeds the recovery rate, tokens are depleted from the bucket and the Token Consumption Rate builds up.
If tokens are plentiful for the calculation system and type, the performance of the calculation is guaranteed.
- The performance is defined by token recovery and capacity, and varies with each subscription plan.
If there are no tokens accumulated (i.e. Token Availability is 0%), only the token recovery rate can be calculated.
The remaining percentage of tokens (Token Availavility) can be used to set priorities. Please see also the another post about using PFP.