kllrnohj

29 ms·

511.
by kllrnohj 2y ago
> They chose poorly Did they? Immutable system images are a pretty solid feature to have. It's a lot less "fall over broken as shit randomly" than desktop Linux is...
512.
by kllrnohj 2y ago
> Why can't Windows and Android provide such a feature? Windows does. Android "can't" because the OS is a partition image with libraries not intended to be updated piecemeal, not a collection of loosely related extern
513.
by kllrnohj 2y ago
What OS ships the mitigation of disabling SMT by default? Surely they just meant things like the retpoline mitigations in syscalls?
514.
by kllrnohj 2y ago
It's not going to hammer on syscalls, either, so it won't have any spectre-related regressions.
515.
by kllrnohj 2y ago
"trivial" how do you figure? Remember these exploits bypass your own code's conditionals over a shockingly far duration. Unless you just mean for incredibly restrictive usages such as eBPF? possible absent any performance
516.
by kllrnohj 2y ago
What are you doing where you see anything remotely close to double-digit-% gains from disabling spectre mitigations?
517.
by kllrnohj 2y ago
The writing was on the wall for in-process sandboxing with Spectre, but that seems to have faded a bit. This just re-enforces that. Things like "safe in-process sandboxing with WASM" are just a fantasy, it can't be implemente
518.
by kllrnohj 2y ago
Or it's a hardware issue and they don't have any way to do a microcode fix for this
519.
by kllrnohj 2y ago
Although desktop GPU plane counts are much more limited, it's not as restricted as you're portraying. Here's the AMD SoC in the Steam Deck, for example: https://github.com/ValveSoftware/gamescope/blo
520.
by kllrnohj 2y ago
There really isn't such a thing as "the actual framebuffer". Instead the display hardware can do composition during scanout from a set of buffers at a set of positions with varying capabilities. These buffers then just bein
521.
by kllrnohj 2y ago
OpenGL flipping is iirc the weird name nvidia gives for vsync. Disabling it means you're going to have tearing artifacts, hence why it's not the default.
522.
by kllrnohj 2y ago
you know there's a pretty massive gap between "double lifetime sentence + 40 years w/ no chance of parole is too harsh" and "10 years w/ full pardon expunging the record is bad". A really, really fucking m
523.
by kllrnohj 2y ago
It kinda seems like they do - 5 years would only include the RX 6xxx and 7xxx. 5 years is not very long tbh.
524.
by kllrnohj 2y ago
Which would be an EO counter the constitution and obviously not durable itself. In 4 years the next DOJ can just enforce the law on the books with 4 years of evidence of companies openly breaking it. It'd be a slam dunk case
525.
by kllrnohj 2y ago
Trump isn't president and the ban went into effect before he was. There's no legal extension possible anymore under this specific case.
526.
by kllrnohj 2y ago
I'm pretty certain an executive order cannot overrule a law. So they're just hoping to either get an actual reversal of the law while Trump is in term or just hoping nobody after him will care. It's like betting on jury nulli
527.
by kllrnohj 2y ago
> Keeping UI rendering under 16ms is the gold standard for native apps. No, it isn't. If you're not including the actual rendering, so the actual code of the app only or at least only the code on the UI thread, the targets are
528.
by kllrnohj 2y ago
> Anyone know what native screen loads typically are? 500ms sounds about right for a cold launch but otherwise is pretty poor. 50-100ms for "minor" screen changes and 100-200ms for "major" ones are otherwise reasona
529.
by kllrnohj 2y ago
> any reasonable CPU from the past 15 years can handle software decoding of high resolution video just fine. 4k VP9 from youtube takes my 5950x around 20-25% CPU usage to handle with hardware acceleration disabled. The fastest consumer
530.
by kllrnohj 2y ago
Cayennes aren't exactly cheap, either.
531.
by kllrnohj 2y ago
> Pony cars have been dying for decades only if your definition of "decades" is ~5 years? Otherwise the resurgence of the pony car was only like 20 years ago in the first place (5th gen mustang & charger 2005, Challenger 20
532.
by kllrnohj 2y ago
> What makes a Real Mustang? A RWD coupe that'll spin out and crash trying to show off leaving a cars & coffee.
533.
by kllrnohj 2y ago
A huge chunk of the value of an RPI is the ecosystem & support. The actual hardware in & of itself is consistently mediocre to poor for the price these days. I think RPI is starting to lose sight of that with the Pi 5 and especial
534.
by kllrnohj 2y ago
A fancy modern turbo'ing CPU which has power availability feedback loops might just slow down in this scenario, but I don't think anyone has put anything remotely that fancy into a low power SoC that these routers were using. So
535.
by kllrnohj 2y ago
yeah but it means MFG still has the same fundamental problem of FG that the latency hit is the largest in the only scenario where it's meaningfully useful. That is, at low 15-45fps native FPS, then the additional impact of an additiona
536.
by kllrnohj 2y ago
Nobody is talking about DLSS 2 here so I don't know where that came from. The 2x, 3x, and 4x in my post are the number of generated frames. So 2x == DLSS 3, and 3x and 4x are then part of the new MFG in DLSS 4. Digital Foundry has actu
537.
by kllrnohj 2y ago
The context was people that don't care about AI or Nvidia GPUs and just want an AArch64 system. So inference performance is irrelevant here > lots of people will buy these machines to get an AArch64 Linux workstation—even if they
538.
by kllrnohj 2y ago
Digital Foundry just covered this. 3x and 4x both add additional latency on top of 2x. https://youtu.be/xpzufsxtZpA?si=hZZlX-g_nueAd7-Q
539.
by kllrnohj 2y ago
This isn't relevant to what I said?
540.
by kllrnohj 2y ago
Games other than esports shooters and slow paced story games exist, you know. In fact, most games are in this category you completely ignored for some reason. Also nobody is buying a 4090/5090 for a "fine" experience. Yes 60f

More ›