Dr. Moritz Lehmann

Hi @gitcnd, a little hack is needed to make this work. Modify the [`LBM::Graphics::write_frame()`](https://github.com/ProjectPhysX/FluidX3D/blob/master/src/lbm.cpp#L1197) function as follows ```diff void LBM::Graphics::write_frame(const uint x1, const uint y1, const uint x2, const uint...

Hi @sumseq, yes you can override the detection with some hacks: ```diff diff --git a/src/kernel.cpp b/src/kernel.cpp @@ -3,7 +3,7 @@ string opencl_c_container() { return R( // ########################## begin of O...

Hi @sumseq, - The `0 MHz` is just a cosmetic information. The `Intel CPU Runtime for OpenCL` internally uses a lookup-table to report `CL_DEVICE_MAX_CLOCK_FREQUENCY`, and for AMD CPUs there is...

Hi @sumseq, I've tested a 2x EPYC 9754 system today. The Intel CPU Runtime for OpenCL is way faster than PoCL on this system too. Kind regards, Moritz

Hi @Snektron, the good old bandwidth anomaly on GCN/CDNA! I'm very familiar with this. I choose 4-byte loads as those are the most common (scalar FP32). The 16 values was...

Omg I'm jealous on all the hardware you have! MTT claimed OpenCL support on their marketing material but never delivered it on Windows. But there is one report of where...

Hi @IMbackK, so with multi-GPU, each GPU only has its own domain kN memory and can't see the others. For rasterization rendering, each GPU can still independently render its own...

@IMbackK the problem here is that one GPU cannot fit the other domains in memory anymore. An option here would be to copy all domains to CPU and then optionally...

Hi @wuxin008, I've been looking into free-slip boundaries for a long time, and have a beta implementation. However it only works normal to axis-aligned or edge diagonal directions. For space...

Hi @rodionstepanov, here you copy the data over PCIe every time step. This of course is very slow. I recommend writing a small OpenCL kernel that does the force update,...