Anton

This is not working properly yet. I will go over it later and resubmit.

I checked to script locally and my problem was from no connecting to the database correctly while checking whether it worked. This could be easier to use when the script...

We are seeing cases with RTX 5090 Laptop, RTX 5070 Laptop and RTX 5060 Ti. Due to #1323 we can not use 1.5.2 to get more information.

@MoFtZ I asked you about this problem in discord. I think it would be better to continue here. I will reproduce your extensive explanation here: > Unfortunately, ILGPU normally stops...

Thanks for the quick response. We make heavy use of MathF.* related functions. But I don't see exactly how that could _suddenly_ case resource/memory issues. [This commit](https://github.com/m4rs-mt/ILGPU/commit/b21fa6451d1dcb5607caea7fb4c3d9394adf1917) looks like other...

> The "too many resources requested" likely comes from the ILGPU implementation of the math functions. You will likely get a better result from enabling LibDevice support when creating your...

I looked into the problem again. The first thing I have to solve is: After upgrading to 1.5.2 from 1.5.1, logic that just worked (same data, same setup, same buffer...

Following the beginning of the compiled kernel in both versions and related csharp logic. If this is not enough, I can start to reduce this to a minimal sample. I...

> The kernel parameters are marshaled as a single buffer. > > From the v1.5.1 PTX, `.param .align 8 .b8 _p_49893[136]` says that a parameter of 136 bytes should be...

> [@afmg-aherzog](https://github.com/afmg-aherzog), as discussed with [@MoFtZ](https://github.com/MoFtZ) during our recent community meeting, we've confirmed that the invalid data alignment issue is present in ILGPU v1.5.2. We'll share a patch within the...