Mastodawn

@nh @pervognsen iirc char can alias with anything and depending on the platform uint8_t may be char, so this potentially still aliases on some platforms?

Bas Nieuwenhuizen Feb 20

Kenneth Hoste Feb 19

If you're interested in RISC-V, this talk by @fclc from FOSDEM'26 is a must see...

RISC-V had 40 years of history to learn from: What it gets right, and what it gets hilariously wrong

https://fosdem.org/2026/schedule/event/X8TLY9-mistakes-in-riscv/

#riscv

FOSDEM 2026 - RISC-V had 40 years of history to learn from: What it gets right, and what it gets hilariously wrong

Show thread

Bas Nieuwenhuizen Feb 18

@karolherbst Strongly disagree. My problem with an AI is:

1. it says "You're right" and then probably is going to make the same/worse mistake.
2. They're unlikely to be more correct than an experienced contributor.
3. It doesn't make sense to take a chance that they'll be a great contributor later, they don't improve.

The reaction on "you're right" is just Pavlovian conditioning between the sentence and the odds of it being an AI.

Bas Nieuwenhuizen Feb 6

Carl Schwan

Feb 5

RE: https://meow.social/@chirpbirb/116014711339837040

We might not have frozen kizza in the store, but we have a lot of fancy accessories and clocks 😍

https://www.cafepress.com/shop/KDE

Show thread

Bas Nieuwenhuizen Feb 2

@GDCPresoReviews @dotstdy @TomF I don't think claims like that have been made for Mali, and I think that unless they advertise it widely I'd assume they don't

Show thread

Bas Nieuwenhuizen Feb 2

@GDCPresoReviews @dotstdy @TomF For CUDA, AFAIU yes: https://nvidia.github.io/cccl/libcudacxx/extended_api/execution_model.html

Execution model — CUDA Core Compute Libraries

Show thread

Bas Nieuwenhuizen Feb 2

@dotstdy @GDCPresoReviews @TomF I dunno if they promise forward progress in graphics, vs only in cuda where nobody promises maximal reconvergence. And for gfx Nvidia has instructions to force reconvergence. Could still be a nice latency hiding tool between two sides of an if, maybe?

Show thread

Bas Nieuwenhuizen Feb 2

@TomF @GDCPresoReviews I believe in NVidia this is sometimes used to not merge execution as soon as possible, to avoid e.g. deadlocks between lanes in the same warp with interleaved execution of both sides. That gets a lot more complicated in explicit predicated SIMD machines.

Bas Nieuwenhuizen Feb 2

QR code scanners at the station in Brussels Airport have been the worst ones I've seen in a while. Everyone taking 20+ seconds to placate the machine on distance/placement (or is it just slow?)

Bas Nieuwenhuizen Jan 27

Khronos Group Jan 27

The 2026 LunarG Vulkan Ecosystem Survey is LIVE! Your feedback shapes Vulkan's future: SDK tools, validation layers, extensions, platform support & more. Quick, anonymous, and it helps every dev in the ecosystem. Take it now! https://surveymonkey.com/r/LRFD7V6