Walter Simson
Walter Simson
compilation is currently working but without openmp which makes it unbearably slow. Need to work on static compilation with openmp before merging this.
Currently, [this branch of the k-wave binaries](https://github.com/waltsims/k-Wave-CPUGPU-macOS/tree/m1) compiles and runs but returns Nans. This will need to be corrected before support for MacOS M1 can be merged in.
Replaced with #410
I think instead of an argument in the signature, the best strategy would be to cache precalculated weights in a temp directory and reload them if available. This does not...
If you want the caching to persist only for the lifetime of your simulation program, ```func_tools.cache()``` would be a suitable decorator. For persistent caching to disk, a decorator that saves...
A little further digging allowed me to [find this example](https://gist.github.com/dsevero/252a5f280600c6b1118ed42826d188a9). I tried this on a separate piece of code, and it worked well: ```python3 from joblib import Memory from functools...
Hey hhapk, We test against nearly all of the examples from k-wave. Though we do not have running examples to mirror all of the k-wave examples in MATLAB at this...
Hey @Jelly954, The time reversal example is still not supported. Please join the discord (link on the main readme) if you'd like to discuss further.
Since it is a benchmark a description of the benchmark with a link to the paper, one figure and a table with the performance results would be a good place...
Perhaps now mlx would be a good tool for this in python.