filprofiler
filprofiler copied to clipboard
A Python memory profiler for data processing and scientific computing applications
E.g. if you try to allocate huge table explicitly, it might fail due to fragmentation, and user code might fall back to normal?
Per `ld.so` man page: " The items of the [LD_PRELOAD] list can be separated by spaces or colons, and there is no support for escaping either separator."
We can read just the files we need with https://github.com/kata-containers/cgroups-rs/blob/master/src/lib.rs#L244, probably.
Right now there is a mapping from memory address to callstack ID + allocation size, for every single allocation. This uses a bunch of memory. Some ways to shrink the...
systemd-run is _probably_ the same? But worth testing with some closer to what people use in reality.
New OOM detection from #110 needs to be tested with cgroups v1 and v2. Ubuntu 20.04 is cgroups v1, so that's easy, but not clear how to test cgroups v2....
Pinning dependencies will reduce random breakage, dependabot doing upgrade PRs will catch problems.
`execve()`ing to a new process would allow outputing actual HTML, because all the application memory would be freed up and one could get a new Python process in good working...
Should be possible to customize `RUST_BACKTRACE` output, and maybe even C segfault output? If it included GitHub URL, users would be more likely to report the bug.
* [ ] Move tracer into Rust? * [ ] Do same line number caching logic for the bytecodeindex->line number logic. * [ ] Pre-allocate data structures that will likely...