pcwalton

63 ms·

811.
by pcwalton 5y ago
> The 3x to 4x decompression speedup here is nice, but I can't help but feel a little sad about how poorly this format is taking advantage of available compute resources. (The ultimate dream would be a format which is so parallel-fr
812.
by pcwalton 5y ago
There are many other popular vector graphics formats beyond those mentioned here. Lottie has already displaced SVG for animation/motion graphics, and for static content IconVG exists and is backed by Google. There's also PDF, Post
813.
by pcwalton 5y ago
FatELF exists, though not widely used, and would make more sense than adopting Mach-O, which has worse tooling than ELF. Because of tooling, it'd be preferable for Apple to switch from Mach-O to ELF in fact (though they won't do s
814.
by pcwalton 5y ago
It's helpful if you're, for example, running unit or integration tests and see stuck threads.
815.
by pcwalton 5y ago
In Linux there's work being done on a language-agnostic lightweight threading model, User-Mode Concurrency Groups or UMCG: https://lwn.net/Articles/863386/ One could imagine a similar, runtime-independent con
816.
by pcwalton 5y ago
A bunch of cryptocurrency startups. For a while now I've been wondering whether the cryptocurrency bust—which is inevitable in my view—will mean the end of these kinds of articles and the resurgence of the Bay Area. Many of the other u
817.
by pcwalton 5y ago
Tanenbaum was in fact right. Monolithic kernel design has caused no end of problems, which is why modern OS's (including Linux!) are moving a lot into userspace. See Windows moving font parsing out of the kernel, macOS DriverKit, Linux
818.
by pcwalton 5y ago
Well, one thing that does work for decoding regular old PNGs is separating the entropy decoding (in PNG, DEFLATE) from the prediction (what PNG calls filtering). You can create a pipeline where one thread is working on the entropy coding an
819.
by pcwalton 5y ago
This is an example of how certain optimizations (specifically, making vtables immutable) are hard in LLVM, because of how low-level it is. Language-specific high-level IRs, such as Swift SIL, can allow compilers to perform these kinds of op
820.
by pcwalton 5y ago
> It would hard for me to make a case that eventually cryptocurrency will fail. It's easy for me to make that case. Ransomware aside, there's no underlying value in cryptocurrency, unlike currencies which have the full faith an
821.
by pcwalton 5y ago
Having a Ph.D. and getting referred by insiders in no way reduces the amount of LeetCode you have to grind for Google interviews.
822.
by pcwalton 5y ago
ESR's proposal here is basically just another way of spelling Python list comprehensions or Perl 5 map (which is likewise a builtin, with special parsing rules). It's not "new PL territory" by any means.
823.
by pcwalton 5y ago
I think there are a lot of complete indie WebGL games, but they aren't nearly as well known as console titles from the late 90s. And that's because the market and technology are much more mature since then: demand for retro games
824.
by pcwalton 5y ago
Because it's fun.
825.
by pcwalton 5y ago
Is C++ adequate, when compile time and maintainability are taken into account?
826.
by pcwalton 5y ago
I'm not aware of any memory safety that works in Zig, other than maybe Boehm GC. The GeneralPurposeAllocator quarantines memory forever, which is too wasteful to work in practice as one allocation is sufficient to keep an entire 4kB pa
827.
by pcwalton 5y ago
We just had an article about this. I am loath to link to Scott Alexander, but it's the most likely thing to persuade HN users. https://astralcodexten.substack.com/p/ivermectin-much-more-t...
828.
by pcwalton 5y ago
Overflow bailout is 1 instruction in RISC-V (same as x86/ARM) if either the addition is unsigned or you know the sign of one of the operands, one of which is usually the case. (If neither is the case, it's 3 instructions.) And I b
829.
by pcwalton 5y ago
I'm a bit confused. First, JavaScript doesn't use bignum arithmetic much. JS Numbers are doubles. JavaScript JITs do speculatively use integers where possible, but this workload doesn't need add-with-carry, just efficient ove
830.
by pcwalton 5y ago
Sure, but this email is in the context of GMP, which should be using the vector extension, no?
831.
by pcwalton 5y ago
Doesn't RISC-V have an add-with-carry instruction as part of the vector extension? I see it listed here: https://github.com/riscv/riscv-v-spec/releases/tag/v1.0
832.
by pcwalton 5y ago
It's probably stdout locking. Call Stdout::lock.
833.
by pcwalton 5y ago
You can basically take any LLVM IR that Clang would produce and produce Rust that compiles down to it (modulo irreducible control flow, I think, but the C++ doesn't have that). Because of that, C++ vs. Rust language (as opposed to li
834.
by pcwalton 5y ago
Zig is, however, not memory safe, which outweighs all of those benefits in this context.
835.
by pcwalton 5y ago
I mean, if you don't care about efficiency, then you don't need any fancy mitigations: just use Boehm GC and call it a day. Performance is the reason why nobody does that.
836.
by pcwalton 5y ago
4chan's /pol is probably the largest neo-Nazi discussion forum on the Internet. I'm actually unsure as to what could even be more toxic than that.
837.
by pcwalton 5y ago
Metal is also a subset of Vulkan. As Raph Levien points out here [1], there are certain important algorithms (decoupled lookback) that can't be implemented at all on Metal. [1]: https://raphlinus.github.io/gpu/20
838.
by pcwalton 5y ago
Interesting. Rust does the same thing with Obj-C selectors, except each selector is looked up lazily at the first call. I wonder why they didn't do that for these bindings.
839.
by pcwalton 5y ago
Zig is not memory safe, and is not meaningfully different from C and C++ in this way.
840.
by pcwalton 5y ago
A quick Google search for [child killed because mistaken for thief] brings up this 2017 case in Alabama: https://www.wjhl.com/news/family-22-year-old-mistaken-as-thi... In this case the killer was indicted and pleaded

More ›