Bug fix: printing non-distributed data

[ClaudiaComito]

Due Diligence

• General:
  • title of the PR is suitable to appear in the Release Notes
• Implementation:
  • unit tests: all split configurations tested
  • unit tests: multiple dtypes tested
  • benchmarks: created for new functionality does not apply
  • benchmarks: performance improved or maintained
  • documentation updated where needed

Description

I noticed that printing out moderately large DNDarrays in non-distributed mode (i.e. interactive session or split=None) takes a disproportionate amount of time (see below) compared to printing the underlying tensor. Culprit is the Formatter call.

I've changed the printing module to bypass Formatter if the input dndarray is not distributed. Torch takes care of the data formatting, tests pass.

Example:

data =  ht.random.randn(91392, 52, 4)
print(data)

On 1 process:
main branch: 84 seconds
this PR: 0.01 seconds

Issue/s resolved: #

Changes proposed:

• let torch handle formatting of non-distributed data

Type of change

• Bug fix (non-breaking change which fixes an issue)

Memory requirements

NA

Performance

see above

Does this change modify the behaviour of other functions? If so, which?

no

[github-actions]

Thank you for the PR!

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.25%. Comparing base (87f2812) to head (3edaca4).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1756      +/-   ##
==========================================
- Coverage   92.26%   92.25%   -0.01%     
==========================================
  Files          84       84              
  Lines       12445    12447       +2     
==========================================
+ Hits        11482    11483       +1     
- Misses        963      964       +1     

Flag	Coverage Δ
unit	92.25% <100.00%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[github-actions]

Thank you for the PR!