Largely parallel processor specializing in usually 16-bit float or 8 bit int matrix multiplication and a few other common fixed functions.
With a lot of the programmability and complexity out of the way you can get a lot more band for the buck from the same silicon.
Branch prediction? not needed. Cache sizes ideally suited to the fixed functions and how they're expected to be used and half or quarter the data width per work item etc.
This lets people figure out how dependent they are on their smart phones and eftpos cards.
I always tell people to keep enough cash on them for at least a day or two's expected expenses, a bit more if you want to plan for the unexpected.
Usually the response is along the lines of "but eftpos will never go down".. then things like this happen and they can't buy lunch.
GPU accelerated encoders are almost never the same quality or better than cpu encoders set to their highest settings for a given bitrate.
But that isn't their point. The point of them is to be able to encode in real time with very little overhead.
AV1 encodes in realtime even when possible on cpu (and many situations this just isn't realistic/possible) chug a lot of resources. If these resources are shared with whatever is attempting to be drawn to be encoded (for example a game) this can cause problems.
GPU accelerated encoding usually has at least some dedicated silicon, so as to not really impact performance otherwise.
Here's one of the best arguments I've heard against the extension of copyright
From an era where the public interest and the good of society was argued for over corporate and private interests.
A bit of a long read, but a standout paragraph imho -
I believe, Sir, that I may with safety take it for granted that the effect of monopoly generally is to make articles scarce, to make them dear, and to make them bad. And I may with equal safety challenge my honourable friend to find out any distinction between copyright and other privileges of the same kind; any reason why a monopoly of books should produce an effect directly the reverse of that which was produced by the East India Company's monopoly of tea, or by Lord Essex's monopoly of sweet wines. Thus, then, stands the case. It is good that authors should be remunerated; and the least exceptionable way of remunerating them is by a monopoly. Yet monopoly is an evil. For the sake of the good we must submit to the evil; but the evil ought not to last a day longer than is necessary for the purpose of securing the good.
Realistically it's a sliding scale, and even the 3ds was a lot more commodity than some prior offerings.
To me switching to a drop in mobile SoC is a clear "you've gone to the other side of the line" moment. No longer unique enough to bother with.
Where is the line before that? depends on your criteria.
Silicon Graphics pretty much designed the N64 for Nintendo.
Yes they did. The design and those chips were however not a typical SGI system. One might even say "bespoke custom hardware and toolchains that were fit for purpose"
None of the Nintendo consoles use mobile phone technology.
The nvidia tegra line first appeared in mobile phones, and at least by their wiki page, are still used in mobile phones
The newer consoles are designed by Nintendo just as much as the old ones, with tweaked versions of off-the-shelf chips.
Nobody is faulting nintendo for modifying a 6502 or the use of picture processing units, or the sony spc700 6502 derivative sound chip with dsps. But to say the overall design itself was not custom is a bit odd. Even the GBA's arm cores had some weird and wonderful things added.
Nintendo aren't redesigning what a transistor is and shouldn't be expected to, but there is a world of difference between the kind of customization you see in even a wii vs what you see in a switch.
Is it a sliding scale? sure. Was even the 3ds a bit more commodity than prior offerings? Yep.
The switch hardware just isn't interesting. At least nintendo used to have unique hardware or hardware combinations.
The 'New' 3ds was the last real bit of hardware designed by Nintendo. The end of the era of bespoke custom hardware and toolchains that were fit for purpose.
Their future direction seems to be along the switch route, off the shelf system on chips that are just overpriced old mobile phone technology.
A true shame considering over the years many nice things came from Nintendo research and engineering and Intelligent systems (who developed the toolchains/devkits)
Why wouldn't it matter? Why would you think having a visually jerky experience with chapter motions not smooth, display scrolling not smooth, and animations of the environment not smooth suddenly "not matter" simply because your view is not first person with a gun visible in the frame?
Low framerates in turn based games do not affect the gameplay in the ways they do in first person shooters and other real-time games. I did not specifically say this because I thought it self evident.
One frame per second will not greatly affect your chances of winning or losing chess.
It changes it from game breaking and actual quality of gaming, vs non gameplay affecting.
The increased latency of all these features is virtually imperceptible in an FPS game.
Depends on what you're going to/from. The benefits of higher frame rates in specific competitive FPS games has been tested and shown. In real frames there's a strong benefit in going from 60hz to 120hz, and a drastically reduced one going from 120hz to 240. Returns diminish damn quickly.
If you're using inter-frame generation to go from 30hz to 60hz? yeah the latency is there and strong. 120hz to 240hz? Completely agree, not a real issue.
If so they can always buy AMD. They aren't doing that though, why do you think that is?
Well.. they do, what do you think ps5 and xbox series x run, nvidia? no. The PC world has a lot of non-graphic things (cough cuda) locking people into nvidia, the inertia is strong.
Also how much input lag is too much when playing Baldur's Gate 3?
The perfect style of game where 30fps doesn't even matter.. coincidence no?
Fixed function is great until you don't need that particular function.
The AMD 6800 XT's are/were pretty great. For gaming they just don't have the fixed function 'ai' hardware nvidia use for their dlss and other bits. Honestly I prefer it without. Real-time inter-frame generation is cool and all but it adds to input latency significantly, no thanks.
Don't get me wrong, I ran nvidia also until maybe 2014 or so. That's when the better drivers won me out. That some years later they are neck and neck performance wise (so long as you aren't faking frames etc) is just a bonus.
What's preventing someone else supporting CUDA, anyway?
Probably being hardware specific and constantly changing. Even if you have stuff written for cuda it won't necessarily work on all nvidia hardware, there's specific support levels etc. That said I think Rocm or such is out to 'recompile' cuda based things to something amd can handle.. I'd rely more on vulkan-compute for cross platform capability and longevity, that stuff can even run on phones.
OpenCL (and vulkan-compute) are for the programmable units on a gpu. This is possibly not what they're referring to when they talk about accelerators (especially if comparing to cuda)
Aside from typical arbitrary computation stream processors (still can't believe industry called them shaders, even in the graphical context fragment programs made more sense, I blame microsoft) nvidia have fixed function blocks. Being fixed function they are a lot more efficient per power and silicon area, so you might have a function that multiplies two 8x8 matrices for example.
The fixed functions change depending on the hardware, and there has to be a means of handling it, an abstraction layer. CUDA handles this.
Nvidia went all in with dedicated functions with the 3000 series cards, amd still had general computing with 'shaders' with the 6800 line. Some fixed function hardware was added in the 7000 line just to compete.
I'm not that much of a fan of fixed functions with deprecation rather than programmable units.. but you can see there's performance benefit to be had if you know those fixed functions will be used in specific ways.
Normally I call out the silly idea of using nuclear to solve the climate crisis due to times involved,
I'm sure people in the 2050's will be saying the same thing, when if we started now it's still better than starting in a few decades.
I remember people being shouted down about proposing nuclear as a green option in the 90's.. if only we'd started then.
Something for your air gapped linux machine cli
It's an open standard, (oath) and there are multiple indepdendent implementations. One of the better two factor methods.
People in the past understood it was to maximize the public good (a long read, but well worth it, when copyright length extension was discussed in the past.)
an excerpt from some ways down -
I believe, Sir, that I may with safety take it for granted that the effect of monopoly generally is to make articles scarce, to make them dear, and to make them bad. And I may with equal safety challenge my honourable friend to find out any distinction between copyright and other privileges of the same kind; any reason why a monopoly of books should produce an effect directly the reverse of that which was produced by the East India Company's monopoly of tea, or by Lord Essex's monopoly of sweet wines. Thus, then, stands the case. It is good that authors should be remunerated; and the least exceptionable way of remunerating them is by a monopoly. Yet monopoly is an evil. For the sake of the good we must submit to the evil; but the evil ought not to last a day longer than is necessary for the purpose of securing the good.
Kyocera are still good too.. but you certainly pay for it (up front at least, not in consumables)
Don't hit the keys so hard, it hurts.
