itamarst

11 ms·

61.
by itamarst 4y ago
https://groups.io is good for mailing lists, yes.
62.
by itamarst 4y ago
Unfortunately NumPy's thread pool is only for BLAS, the underlying library for numpy.linalg functions, mostly. Other operations are single-threaded. So you need your own thread pool (or process pool) if you want to parallelize anything
63.
by itamarst 4y ago
Yes, but... it's harder than one might think to avoid threads. Merely importing NumPy starts threads, for example.
64.
by itamarst 4y ago
If I add a `if __name__ == '__main__':` I get the same result. Here's the gdb stack trace: #0 __futex_abstimed_wait_common64 (private=<optimized out>, cancel=true, abstime=0x0, op=393, expected=0, futex_word=0x17bc
65.
by itamarst 4y ago
See the article for an explanation of why this is a bad idea. Note that on macOS Python disabled fork()-based multiprocessing as the default long ago (Python 3.8?) because it's so broken, and it will stop being the default for multipro
66.
by itamarst 4y ago
From memory, the reason it deadlocks is: 1. Writing to stderr grabs a lock. 2. Part of the multiprocessing code (perhaps not present in Python 2) also grabs this lock. 3. If you fork at the right moment (which is quite likely with the loop)
67.
Python’s Multiprocessing Performance Problem(pythonspeed.com)
4 points by itamarst 4y ago|0 comments
68.
by itamarst 4y ago
I fundamentally don't trust the ADL, so not going to waste time reading their stuff. Based on my political opinions they would claim I'm an anti-semite, and again, I'm Jewish, grew up in Israel, still have family there. For a
69.
by itamarst 4y ago
The ADL tries to redefine any support for Palestinians as anti-semitism... they're not a credible source of information on the topic. Nor is the Jerusalem Post on this topic, nor is the NY Post on... anything? Looking at the Wikipedia
70.
by itamarst 4y ago
These AI systems are like a spell checker that hallucinates new words: did you mean to type "gnorkler"? At least Google (when not using the summarization "feature") doesn't invent new stuff on its own.
71.
by itamarst 4y ago
This is an excellent summary, but it glosses over part of the problem (perhaps because the author has an obvious, and often quite good solution, namely DuckDB). The implicit problem is that even if the dataset fits in memory, the software p
72.
Don’t bother trying to estimate Pandas memory usage(pythonspeed.com)
3 points by itamarst 4y ago|0 comments
73.
The problem with float32: you only get 16M values(pythonspeed.com)
1 points by itamarst 4y ago|1 comments
74.
Float64 to float32: Saving memory without losing precision(pythonspeed.com)
3 points by itamarst 4y ago|0 comments
75.
Some Reasons to Avoid Cython(pythonspeed.com)
3 points by itamarst 4y ago|0 comments
76.
Why Polars uses less memory than Pandas(pythonspeed.com)
6 points by itamarst 4y ago|0 comments
77.
by itamarst 4y ago
If you're writing a compiled extension, you can use Rust. Or Cython. Or C. Or C++. Which should you use? TL;DR: * If you're wrapping existing C library, I'd use Cython. * If you're wrapping existing C++ library, I'd
78.
by itamarst 4y ago
Quick overview of the design space: * PyPy JITs everything, so it can do _normal_ Python numerical code that is quite fast and regular Python code that is fast. However, its interactions with libraries like NumPy add overhead, and it seems
79.
It’s time to stop using Python 3.7(pythonspeed.com)
6 points by itamarst 4y ago|1 comments
80.
Hardcore Aesthetic(discoelysium.fandom.com)
1 points by itamarst 4y ago|0 comments
81.
Amazon has as much Seattle office space as next 40 biggest employers (2017)(seattletimes.com)
4 points by itamarst 4y ago|0 comments
82.
The cold incubator: the VC dream of workerless wealth(codewithoutrules.com)
1 points by itamarst 4y ago|0 comments
83.
Who controls parallelism? A disagreement that leads to slower code(pythonspeed.com)
5 points by itamarst 4y ago|0 comments
84.
Recommendations for publishing End-of-life dates and support timelines(endoflife.date)
1 points by itamarst 4y ago|0 comments
85.
by itamarst 4y ago
You can buy chest freezers that have conversion to refrigerator as a built-in feature, e.g. https://www.bestbuy.com/site/amana-16-cu-ft-chest-freezer-wi... (Appears common enough that BestBuy has it as a search flag).
86.
Good old-fashioned code optimization never goes out of style(pythonspeed.com)
4 points by itamarst 4y ago|1 comments
87.
by itamarst 4y ago
Yeah this was a copy/paste mistake, I have fixed it.
88.
Early speed optimizations aren’t premature(pythonspeed.com)
7 points by itamarst 4y ago|6 comments
89.
How to speed up the Rust compiler in October 2022(nnethercote.github.io)
75 points by itamarst 4y ago|10 comments
90.
Faster hardware is a bad first solution to slow software(pythonspeed.com)
1 points by itamarst 4y ago|0 comments

More ›