perf now available also to Python programs. This definitely can be useful for proper profiling.
Interesting position from AMD regarding the race on the next super computers. They're all being caught up by energy efficiency so it'll need to be addressed both at the processor architecture level but also at the software architecture level. How we design our computing tasks will matter more and more.
A bit of a sarcastic tone but a few good point in there. Also shows interesting alternatives to C++ to squeeze every ounce of performance out of your code whatever the platform it runs on. Of the three options explored I knew only about Numba really.
Excellent analysis and explanation of the stutter problem people experience with game engines. It's an artifact of the graphics pipeline becoming more asynchronous with no way to know when something is really displayed. Extra graphics APIs will be needed to solve this for real.
Time to look a bit at the maze of WebAssembly runtimes. Good overview on how they currently perform and how well they are documented or easy to use.
Python is getting faster but is still far from what you can get with C++ of course. That said, for simulations you likely don't want everything in Python or in C++. Part of the challenge is to split the subsystems properly and use C++ where it matters.
Don't underestimate performance of the generated code when a JIT is in the picture. Very good example with the JVM just there.
Don't bank it all on faster hardware, make sure your software isn't slow first. Otherwise it'll bring quite some hidden costs.
Little simple benchmark of WebAssembly performances for the most common languages found there. Careful to the payload size though.
Definitely this, we have to stop pointing disk I/O so much for performance issues. This is just not really slow anymore. Obviously network is a different story.
Nice summary on the false sharing problem with caches and how it can impact your performances in multithreaded contexts.
Interesting deep dive on how sets and dicts are implemented in CPython. There are a couple of interesting tricks in there.
There are indeed a few architectural problems with the Fediverse as it is. Can this be solved? Hopefully yes.
Interesting take about how performance optimizations can sometimes leverage even more performance gains than you would expect.
Good reminder that "premature" doesn't mean "early". Poor Knuth is so often badly quoted in the context of optimization that it's really sad. The number of times I see "early pessimisation" on the pretense of avoiding "premature optimization". Such a waste...
This is good news, this provide more venues for improving performances in Python modules next to switching to compiled Rust with something like PyO3. There's clearly a case to be more for not having to rewrite when the codebase was already mostly Python.
This has some interesting promises in terms of performance using Python. Looks a bit like a CUDA for Python... to be seen how it fares in practice.
Let's put this quote back in its context, shall we?
One of the best developer tools around for analysis and profiling. I'm glad it exists, saved me a few times.
Wow, this is a very good exploration of the performances of several common languages and runtimes. This is one of the most thorough I've seen. A good resource for deciding what to pick.