Catalyst In Situ Analysis - #599
Conversation
…stration orders, regarding first attribute will be used for colouring extracted pngs
…to catalyst-viz-v2.0.0
|
now resolves #590 completely by replacing third party json (nlohmann/json) parser with catalyst_conduit based json parser (which we have already due to the catalyst dependency). Will supersede (resolves #210 PR) prio catalyst insitu code which then can be closed. Further PR #350 (ascent based insitu; if further pursued) should replicate this PR approach for general insitu. |
|
cscs-ci run cscs-ci-gh200, cscs-ci-mi300, cscs-ci-openmp |
|
cscs-ci run cscs-ci-openmp |
On Linux x86_64 Conduit implements conduit_int64 as signed long, so the Catalyst Conduit C++ wrapper exposes no set_external(long long*) overload. Add ippl::detail::conduitCompatiblePtr in ParticleAttrib.hpp that reinterpret-casts long long* / unsigned long long* to the matching Conduit bitwidth pointer when necessary, guarded by static_asserts on size and alignment. Use it for scalar attributes and vector components in signConduitBlueprintNode. This fixes the ParticleUpdate build error caused by ParticleAttrib<long long>.
|
I tried building on my machine and had a compilation error - I have a made a PR onto this branch in @klappi-s 's repo - I can't make the PR directly into this branch it seems. |
VIZ: map long long particle attributes to Conduit bitwidth types
|
thanks @biddisco I accepted the PR in my fork |
biddisco
left a comment
biddisco
left a comment
There was a problem hiding this comment.
This basically works, there are some issues with the caching that need to be addressed, but these can be done in a separate PR since we need this basic infrastructure merged in order to make progress.
I will submit a follow UP PR with fixes and additional functionality
3 participants
resolve #598
aand maybe #590
This PR...