cpgxiii

26 ms·

271.
by cpgxiii 6y ago
SpaceX's innovation was looking at the market and deciding that Iridium was right but about two decades too early. A lot of people in the industry (and a number of governments as well) lost big when Iridium went bankrupt, and quite und
272.
by cpgxiii 6y ago
Actually, through their Altera acquisition, they must have design licenses for a number of recent Cortex-A variants as well, since a number of their FPGAs have ARM cores.
273.
by cpgxiii 6y ago
I don't think it was a lack of experience with what ARM designs could do. If I were to guess, their hope was that a nominally-familiar architecture would make porting easier for customers interested in building low-end embedded + serve
274.
by cpgxiii 6y ago
In principle, graphics is different - they have to do graphics development given their need for integrated graphics, and they have an established (if quite rocky) history of highly parallel HPC accelerators/processors. Cell modem dev
275.
by cpgxiii 6y ago
You can already use AMD GPUs on some aarch64 machines, and anything with upstream Linux driver support is likely to work fine.
276.
by cpgxiii 6y ago
It's different. Xe LP, which exists right now in integrated and discrete forms, is comparable to low-end discrete GPUs from Nvidia and AMD and AMD's APUs. The most distinctive feature of the Xe LP variant is the emphasis on video
277.
by cpgxiii 6y ago
FPGA vendor support for higher-level synthesis tools is still pretty poor. Not surprised on that front. Offload-from-GC runtime tools do exist, e.x. Cudafy would translate .Net code into Cuda kernels and handle kernel dispatch. Of course, y
278.
by cpgxiii 6y ago
There are different kinds of "relaxed stability", largely depending on which axes of the aircraft are affected, and the magnitude of the instability. Longitudinal stability is something of a special case, in that essentially all s
279.
by cpgxiii 6y ago
it strikes me that the serious bug here is the design of GDM to check accounts-daemon for existing accounts, and if that check times out, start account creation. Regardless of the nominal protections afforded by the system, if you rely on a
280.
by cpgxiii 6y ago
That article, and Gregory Travis' assessment, offer zero actual evidence that the design is unstable. Lot of hand waving and "the engines are different so it must be dynamically unstable" but no actual evidence, which is obvi
281.
by cpgxiii 6y ago
No it is not. No regulators has yet been prepared to certify a fixed-wing aircraft that is intrinsically unstable, even after decades of proven fly-by-wire development. The 737 Max has different yoke force during pitch-up than predecessor 7
282.
by cpgxiii 6y ago
It comes down to gradients. When you have useful gradients to work with backprop on a GPU is vastly faster and more directed than GAs. One could uncharitably say that all of the impressive results from NNs in the last decade are the result
283.
by cpgxiii 6y ago
> As the article explains, HLK testing isn't really designed for non-hardware drivers and the solutions are strange at best. You're not actually improving quality in this case, you're just making us do convoluted busywork.
284.
by cpgxiii 6y ago
I will absolutely denigrate vendors who lag behind on driver quality and support, particularly vendors who sell expensive equipment tied to ancient and shitty drivers who can't be bothered to update them support new Windows versions. O
285.
by cpgxiii 6y ago
The title is a gross exaggeration of the policy change MS is making. Come 2021, what will go away is the ability to sign drivers for Windows 7 and 8 without passing WHQL/HLK compliance tests. That's it. You can still update compli
286.
by cpgxiii 6y ago
Considering the number of stories (or at least legends that people were afraid of) of people assumed dead and buried while still alive, no, before modern medicine people were very unsure when dead was dead in marginal cases.
287.
by cpgxiii 6y ago
Part of it is certainly an issue of volume, but in the consumer space low-cost (or "free" subsidized-by-plan) 3G/LTE mobile hotspots have long been available. Those models are basically the modems and radios of a mid- to high
288.
by cpgxiii 6y ago
There are a number of quite good modern primes that are tiny with big apertures, especially for smaller sensor sizes. Even then, though, lens plus camera is almost always much bigger than the current crop of depth cameras (K4A, Realsense D4
289.
by cpgxiii 6y ago
As far as large lenses and sensors not working well on sensors, I think you'd be surprised just how well large lenses designed for the sensors they belong on can work. Indeed, by far the reason why photography lenses are huge is becau
290.
by cpgxiii 6y ago
The Xeon Phis really shit the bed on developer experience, so hopefully that experience has been taken to heart. The Xeon Phi marketing promise was basically "take your existing parallel x86 code, run it in on more cores", with so
291.
by cpgxiii 6y ago
As difficult as it is, control for launching and landing a rocket is vastly easier computationally than even apparently simple robotics problems. Robotics is simply hard, from algorithmic, computational, and engineering perspectives. Only s
292.
by cpgxiii 6y ago
I would say some of the most promising work right now is in quasi-direct-drive actuators, basically low-gear-ratio brushless motors. Standalone electric motors are basically linear in terms of electrical power->torque, and even with a ge
293.
by cpgxiii 6y ago
Structure from focus is an established technique that can produce quite good results, but it's not quite as easy as the existence of good autofocus in cameras would suggest. SfF techniques often operate the the inverse of an autofocus
294.
by cpgxiii 6y ago
In an ML context, the global optima may often correspond to better performance and (hopefully) a more general solution. In a motion planning or controls context, a local minima can often mean a configuration or path that is infeasible due t
295.
by cpgxiii 6y ago
Oh sure, it's fixable and something you can work around, but having your build system responsible for "machine is unresponsive and acts like it's out of disk but you've got >100GB space remaining" is a real pain
296.
by cpgxiii 6y ago
Yes, in general if you can capture multiple views you can make up for many dropouts and artifacts. The more you can move, the more likely that you'll get a complete reconstruction. Single-view artifacts are more of a problem when reach
297.
by cpgxiii 6y ago
Near-IR, which many of the available ToF (Kinect2, K4A, Basler/Lucid/Sony), active stereo (Realsense D4xx, Mynt, plenty more), and many some structured light (all Primesense derivatives, some older Realsense) cameras use, as well
298.
by cpgxiii 6y ago
The project is our internal monorepo, with probably order ~1000 targets, mostly C++, with reasonably heavy use of templates. GCC needs somewhere on the order of 2-4GB to compile each C++ target.
299.
by cpgxiii 6y ago
The problem is that "distance" here is not necessarily in any intuitive or useful space, and knowing if you are close to the goal may be as hard as finding the full solution. You can be quite "close" to a solution in, sa
300.
by cpgxiii 6y ago
Generally speaking, no. But there are classes of problems (ex surgical robotics) where solution optimality (or at least near-optimality) is essential to having a usable and safe system.

More ›