-
Couple of Python test failures (debug/windows) whilst investigating the SWIG 4.2.1 issue, note this is with the Python wheel built with SWIG 4.2.1. Although, I've built it with 4.0.2 and received the …
-
WP 2.4 of the BBR grant is to "Enhance GPGPU functionality using FLAME GPU":
> We will also significantly enhance Chaste GPGPU functionality by utilising [FLAME GPU](https://github.com/FLAMEGPU/FLA…
-
Encountered a bug when working on a non-trivial model, where when using the DAG api (`dependsOn` etc) errors would occur when abstracting the definition of agent function behaviours and inclusion in t…
-
In the past few months Google Collab has updated the Python version twice, in both cases breaking the pyflamegpu install.
Would be nice if wheel selection could be automated.
https://github.com/…
-
Using doubple precision (64 bit) floating point values for X/Y/Z (and presumably glm position) types results in an error at runtime in `ModelVis.cpp`.
```
Agent 'agent' has not had x, y or z vari…
-
Right now the main() code is ignored, because Flame/FlameGPU don't have a good place where it can be run. Because of this it's currently necessary to generate the initial agent states separately.
T…
nikic updated
7 years ago
-
`PrettyWritter` extends the `Writer` class but the base class methods are not virtual and as such if a pointer to a `PrettyWriter` is cast to a `Writer` the default behaviour is that output is not wri…
-
I want to call a python function when FLAMEGPU is running and pass some Environment variables to it, and then pass the return value to FLAMEGPU to update these environment variables. Is this possible …
-
Current telemetry data approximates unique users by the build ID. For Python distributions all users have the same ID so the true number of unique (but anonymous) users is unknown. To better understan…
-
Cmake uses the appropriate compiler based on the extension of teh file. i.e `.cpp` files will use `g++` while `.cu` files will use `nvcc`.
Visual studio with cuda enabled will pass all files through…