October Changes from Julius, mainly time statistics #71
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- read in old time statistics values - set how many time has passed for weighting the statistics correcltly - start time statistics in the future - yes!
- first in-depth performance debug for balanceLoad, saves work done for every processor - more timings to deconstruct createMask and look why it's so damn expensive - reworked post_mean with modern functions - corrected relative tolerance for poisson solver - module_operator now has the funky Laplace operators, enabled by default for skew-symmetric discretizations
- new debug switches that enable various development information to debug the loadBalance for large parallel runs - thank you little helper for going through the makefile and adding missing dependencies
- skip wing drawing if there are no wings (for fractal_tree) - skip sponge layer loop for blocks not in sponge layer
- reduced program size by 60% by changing stack allocations to heap allocations in module_insect - moved multigrid_vcycle to poisson module
- compute covariance for velocity or velocity-pressure, needs to be followed by both mean values at some point - before, you had to strictly place the mean computation after the variance computation, now you can place it anywhere, as long as it is behind the variable
Update my working branch
- helabs as timestatistics variables - reduce number of equations for adaption of timestatistics, but only if we do not want to threshold to any - some more stencils because why not? Includes 3rd, 5th, 7th and 8th order composite stencils - now all the stencils are taken from module_operators - for second order stencils we still have to decide how to continue in order to chooce the diffusion or poisson stencil, for now i have parity with previous versions
- tsave_stats was broken and now can be used again - we will now save statistics at every last step of a simulation, elsewise plots look weird - little bugfix for high order poisson stencil sizes - performance optimizations for GS iterations, I hardcoded the stencil operations (yes, it does make a difference)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What I changed
An overview of what I worked on, so that the code makes more sense.
Commit Minor changes from October 6th:
Commit More time statistics option from October 7th:
Commit More timing and two bugfix-commits from October 8th:
Commit Debugging workload from October 9th:
Commit Mask optimizations from October 10th:
Commit Reduced program size by 60% on October 16th:
Commit Covariance and more flexible timestatistic placement and three very minor commits on October 16th:
Commit Correct velocity saving for ConvDiff module on October 20th: