itamarst
Follow
Create your own GitHub profile
Sign up for your own profile on GitHub, the best place to host code, manage projects, and build software alongside 50 million developers.
Sign up
Pinned
1,641 contributions in the last year
Contribution activity
July 2020
- itamarst/im-rs Rust
Created a pull request in tahoe-lafs/tahoe-lafs that received 16 comments
Port humanreadable.py to Python 3
Fixes ticket:3324
I did the port by using futurize --unicode-literals --both-stages and then deleting the bits that were wrong or unnecessary.
Basi…
+224
−109
•
16
comments
- Port version_checks to Python 3
- Port hashtree to Python 3
- Port observer and pipeline to Python 3
- Port allmydata.util.spans to Python 3
- Port abbreviate and time_format to Python 3
- Port netstring and hashutil to python 3
- Port base32 and base62 to Python 3
- More tests for base32 and base62
- Port deferredutil to Python 3
- Delete some dead code.
- Port assertutil and mathutil to Python 3 (by switching to already-ported pyutil versions)
- Switch to flake8 as a better way of running pyflakes.
- Use jemalloc for Rust on Linux
- Compress allocation sizes, reducing tracking overhead
- Reduce report generation memory use
- Reduce tracking memory overhead
- Run tests that ensure Python objects are tracked.
- Load directly from libc, so we don't get some other malloc by mistake.
- Don't load aligned_alloc on macOS
- Include bug reporting link in report.
- Fortran and C++ testing
Created an issue in pythonspeed/filprofiler that received 12 comments
Fil missing massive chunks of memory usage in a Python program that creates mostly just Python objects
Hoping for reproducer, but may have to make my own: Python 3.6 (running on Debian) loads 200MB CSV. Result uses 1GB of RAM, this shows up fine in …
12
comments
- Figure out if `ld --wrap` can remove need for initialization code on Linux
- Windows support
- Consider using jemalloc for Rust code on macOS as well
- Try to reduce memory usage by using nested data structures
- Play around with Sanakirja
- Reduce tracking memory usage with compressed 32bit length
- Fil has a huge amount of memory overhead per allocation due to use of HashMap
- Fil has a huge amount of memory overhead per allocation
- Use jemalloc for allocation
- Creating the report can use a huge amount of memory
- Make sure via test that we're not leaking memory due to using dynamic linker's malloc()
- More robust fallback malloc() and friends on Linux
9
contributions
in private repositories
Jul 1 – Jul 24