Feature device synchronization

[kris-rowe]

Description

Closes #598.

This could be split into 2 separate PRs: the first implementing stream::finish() and a subsequent one implementing device::finishAll(); however, the need for stream::finish is less clear outside of the current context.

[codecov]

Codecov Report

Merging #609 (3f71526) into development (de6630f) will increase coverage by 0.00%.
The diff coverage is 33.33%.

❗ Current head 3f71526 differs from pull request most recent head 4d932ce. Consider uploading reports for the commit 4d932ce to get more accurate results

@@              Coverage Diff              @@
##           development     #609    +/-   ##
=============================================
  Coverage        77.26%   77.26%            
=============================================
  Files              261      258     -3     
  Lines            19554    19392   -162     
=============================================
- Hits             15109    14984   -125     
+ Misses            4445     4408    -37     

Impacted Files	Coverage Δ
src/c/device.cpp	95.00% <0.00%> (-1.62%)	⬇️
src/c/stream.cpp	0.00% <0.00%> (ø)
src/core/device.cpp	90.03% <0.00%> (-1.09%)	⬇️
src/occa/internal/modes/serial/device.cpp	89.72% <ø> (-0.06%)	⬇️
src/occa/internal/core/device.cpp	72.46% <40.00%> (-2.54%)	⬇️
src/core/stream.cpp	67.24% <100.00%> (+1.16%)	⬆️
src/occa/internal/modes/serial/stream.cpp	100.00% <100.00%> (ø)
src/c/kernel_fortran_interface.cpp	0.00% <0.00%> (-10.00%)	⬇️
src/occa/internal/utils/store.hpp	90.62% <0.00%> (-6.25%)	⬇️
src/occa/internal/utils/string.hpp	92.98% <0.00%> (-4.72%)	⬇️
... and 17 more

[deukhyun-cha]

Currently we have occaFinish()/occa::finish() as c/c++ API. If I didn't miss this doesn't seem include the exposure of stream::finish() and device::finishAll() to that end. I am wondering if/how it would come later. BTW this change would make the implementation (and API) much clear IMO.

[kris-rowe]

Currently we have occaFinish()/occa::finish() as c/c++ API. If I didn't miss this doesn't seem include the exposure of stream::finish() and device::finishAll() to that end. I am wondering if/how it would come later.

Good catch. I was thinking to add this later, however I might as well do it now before it is merged so that it isn't forgotten.