Enhex (@Enhex) profile picture

Working on new compiler technologies https://teapawt.net/t.co/bbvEdJLpwY

Joined November 2013
Are medical studies being written with ChatGPT? Well, we all know ChatGPT overuses the word "delve". Look below at how often the word 'delve' is used in papers on PubMed (2023 was the first full year of ChatGPT).
354
1,758
553
11,445
2,470,070
removing bound checking on a simple sum loop made it 42 times faster with Clang (unroll and SIMD): quick-bench.com/q/LlEHKowZQA… GCC generated code that's only 3 times faster. #cpp
1
326
generated std::array's size is set based on statically inserted elements. youtube.com/watch?v=idopHR66…
196
I'm designing my programming language's GUI code editor to allow writing code fast by minimizing needed keystrokes. key presses + releases C++ pure text: 50 C++ assisted text: 42 Python assisted text: 36 L3 GUI AST editor: 28 (and there's still room for improvement)
2
3
I wonder if software slowing down (Wirth's law) is caused by devs believing hardware is fast so trading off perf for dev time has insignificant slowdown, which may be true in isolation, but the slowdown compounds with software of other devs that think the same ...
1
... causing significant slowdown. inverse self-fulfilling prophecy.
there are no infinite loops
this is a tough one
295
2,897
282
23,379
The Computer Language Benchmarks Game is deceptive. For example C is listed as faster than C++ fannkuch-redux benchmark: benchmarksgame-team.pages.de… But the C code is also valid C++ code, so they're lying: godbolt.org/z/63M83sWha
1
1
with few exceptions C code is valid C++ code, so if their result is that C is faster that means their benchmarks are wrong.