[ENH]bedpostx and probtrackx gpu and multithread support

[mauriliogenovese]

This will enable the gpu version of probtrackx2 and add support to specify the number of thread used by bedpostx (by default it uses a thread per available core)

[codecov]

Codecov Report

❌ Patch coverage is 45.45455% with 12 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.88%. Comparing base (114c73d) to head (3bea7ef).
⚠️ Report is 31 commits behind head on master.

Files with missing lines	Patch %	Lines
nipype/interfaces/fsl/dti.py	45.45%	12 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3722      +/-   ##
==========================================
+ Coverage   72.85%   72.88%   +0.03%     
==========================================
  Files        1279     1279              
  Lines       59321    59387      +66     
==========================================
+ Hits        43219    43287      +68     
+ Misses      16102    16100       -2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[effigies]

IIUC, this will only trigger _cuda_update if use_gpu is set after the interface is initialized, and not respect the parameter in ProbTrackX2(..., use_gpu=True).

What about:

Suggested change

	super().__init__(**inputs)
	self.inputs.on_trait_change(self._cuda_update, "use_gpu")
	super().__init__(**inputs)
	self.inputs.on_trait_change(self._cuda_update, "use_gpu")
	self._cuda_update()

[effigies]

Likewise, does this work if I pass use_gpu or num_threads to __init__()? Or should you call these unconditionally here?