Steffen

OK, please test the git version since not all clang compatibility patches were back ported for gentoo. I'll try to write a live ebuild for that. But I think this...

Okay, I have a live ebuild here: https://gitlab.com/nobounce/gentoo-overlay/-/raw/main/sys-process/btop/btop-9999.ebuild so the files are still managed by portage

As long as it's the git version it's fine

Ok this took some time. Did you experience this with `-O2` as well? You said that you're going to try it with no optimizations at all. This would be interesting...

Okay, I can reproduces this when hitting the system with some load

Nope, it happened 2/4 times when running under high system load with Clang, O3, LTO and libc++. The variable responsible for that is `pread` in the `Proc::collect()` function of `src/linux/btop_collect.cpp`

@aristocratos `max()` doesn't seem to be the issue. ``` #0 0x00007ffff7f3b96b in std::__1::basic_streambuf::uflow ( this=0x7ffff75fe720) at include/c++/v1/streambuf:447 #1 0x00007ffff7f3e1cf in std::__1::basic_streambuf::sbumpc ( this=0x7ffff002e110) at include/c++/v1/streambuf:191 #2 std::__1::basic_istream::ignore (this=0x7ffff75fe710, __n=9223372036854775807, __dlm=32)...

Can you post the last log and the output of `btop --version`? Do you have good output with nvidia-smi?

> The Jetson systems don't use the standard nvidia-smi application to access the GPU information; however, these details are accessible via jetson-stats: Ok I guess this is the problem. btop...

From the [README.md:](https://github.com/aristocratos/btop/blob/main/README.md#compilation-linux) > For x86_64 Linux the flag GPU_SUPPORT is automatically set to true, to manually disable gpu support set the flag to false, like: > > make GPU_SUPPORT=false...