BearOso

27 ms·

361.
by BearOso 4y ago
Wow. That intro with the joke about Hale-Bopp comet fans certainly wouldn't age well. I was surprised he made such a seemingly tonedeaf statement, but the whole Heaven's Gate fiasco didn't happen until a month later.
362.
by BearOso 4y ago
Yep, I've had this problem bite me and I had to go with CMake instead. I like Meson's python-like language much better, though.
363.
by BearOso 4y ago
I assume you're referring to Google pushing the Carbon language? That uses LLVM as the backend, so they're basically adding more resources.
364.
by BearOso 4y ago
There's actually 3 drivers for vulkan, amdgpu-pro, amdvlk, and radv. In any case, ROCm works with the completely open source stack, and is open source itself. amdgpu, the kernel component, is the only thing needed for that, and modern
365.
by BearOso 4y ago
> There's none that I can notice. Mice cursor are rendered with hardware support when available. Not sure what you reference is here. Some compositors don't use the hw cursor plane. All others that do don't update the curs
366.
by BearOso 4y ago
It's not the feature that's interesting, it's the story from an outsider describing an experience of working on it and getting it added to KDE.
367.
by BearOso 4y ago
If you're using a compositor you're paying the cost. The rounded corners are antialiased, so blending happens regardless.
368.
by BearOso 4y ago
I think it's the same problem as 3D TV. There's a novelty, but actually using the tech is straining. With 3D TV your eyes hurt a lot. With VR... your eyes hurt a lot, and you have this cumbersome device on your head. A problem wit
369.
by BearOso 4y ago
From what you describe we actually strongly agree. This stuff should work and be invisible to the user. I just meant that the environment should expose the DPI details to the programmer and not pretend everything is low dpi like it's a
370.
by BearOso 4y ago
That's because what seemed like the best way to do hidpi when macOS added it (fixed scale factors and assets) ended up being a problem. It's better to not abstract away the DPI differences. It makes for worse compatibility in the
371.
by BearOso 4y ago
Dosbox has basically unlimited bus speed, too, so you're not getting the delay from writing to a real VGA.
372.
by BearOso 4y ago
Exactly. Mark Rejhon, the Blurbusters guy, wrote a lot about it and had a proof of concept. A couple emulators implemented it.
373.
by BearOso 4y ago
And its new official name shall be the Trypophobigon.
374.
by BearOso 4y ago
They can both be equally authentic, there's nothing preventing either from doing the exact same thing as the other. But the software based one will have a lot more overhead keeping track of cycles and making sure things interact at the
375.
by BearOso 4y ago
You can get the precise scanout positions easily. For direct scanout: Windows has fullscreen exclusive mode, X11 has unredirecting, and Wayland recently added direct scanout surfaces. There's major API extensions to query the vblank ti
376.
by BearOso 4y ago
Interestingly, most 60hz monitors I've tested can tolerate the SNES/NES weird refresh rate without problems. In Windows you can create such a video mode with ToastyX's CRU, and with custom modelines on Linux. All you have to
377.
by BearOso 4y ago
They're hardware, but not the same set of logic as the original systems. They still rely on reverse engineering efforts that occur when developing software-based emulators. They produce the end results that were measured from the origi
378.
by BearOso 4y ago
Remember last time they tried a universal media filesystem with UDF? It was implemented in the most incompatible ways as a token gesture by both Microsoft and Apple. These companies want their own, patented, proprietary fs so they can main
379.
by BearOso 4y ago
The original python version uses pytorch/CUDA and requires nvidia. The Vulkan version doesn't.
380.
by BearOso 4y ago
To further elaborate on your "probably never," many people like myself live quite a distance from the usual locations of both cloud servers and game servers. If you're not in the east or west coast of the US, it's probab
381.
by BearOso 4y ago
Worse, QBasic was limited to one segment of 64k, or what Borland called the "small memory model." If the code and inline data was larger it couldn't run. Many programs would inline machine code and jump to it to get around li
382.
by BearOso 4y ago
I realize SBCL predates this folly, but I always regard things hosted on sourceforge with hesitancy after that sketchy ownership period. I'm surprised so many projects stuck with it, especially since their VCS servers were so slow duri
383.
by BearOso 4y ago
Well, they just added small BAR support to the Linux driver, so Intel certainly agrees.
384.
by BearOso 4y ago
Tim Cook said under oath that they don't need to differentiate themselves because they don't have to compete.
385.
by BearOso 4y ago
Profit can't grow forever. At some point your company is so saturated you have to wring water from stone, literally nickel-and-diming people.
386.
by BearOso 4y ago
> remove the resizable bar requirement They'd have to add support for changing the base address register. The easy default is just mapping all the memory directly. It's a lot more programming work to transfer data when you on
387.
by BearOso 4y ago
I don't know about it becoming the standard LL language yet. There's a lot of memory management in low level programming, and peppering everything with unsafe seems like it'd be tedious. I'd rather just write in C to beg
388.
by BearOso 4y ago
It's not even the architecture. It's the fact that Intel was stuck on 14nm for a decade. Apple also paid for the pick of the litter from TSMC, so they're always a step ahead of everyone else. Current AMD mobile processors are
389.
by BearOso 4y ago
Should be roughly the same. Are you using fractional scaling? That would add an extra layer of supersampling to the screen.
390.
by BearOso 4y ago
I'm curious--What GPU/driver were you using? I suspect Nvidia from my experience.

More ›