6 ms·
Wouldn’t the effect be Apple and others creating a new (and hopefully better) replacement for ARM?
by irrational 5y ago
Wouldn’t the effect be Apple and others creating a new (and hopefully better) replacement for ARM?
- ivirshup 5y ago"just create a new instruction set"? It's taken the better part of a decade for workable alternatives to CUDA to emerge.
- irrational 5y agoBy that logic there will never again be another workable alternative. Why not start now rather than wait a few decades?
- InvertedRhodium 5y agoThere's nothing that prevents you from doing that as it stands.
- nerdponx 5y agoInnovation is not an on-off switch. Market forces create an incentive to attempt to innovate, but they don't magically make innovation happen.
- vxNsr 5y agoIt's likely that there are many new designs being spun up everyday by various people, but that doesn't mean that we want to remove competition from the market.
- techdragon 5y agoI hope you’re not calling ROCm a workable alternative… it’s really not close. I love that they are trying, I just honestly feel they need to be trying with 10 times as much funding/resources behind them. Or do you mean Apple’s Metal Compute? useful only with inconvenient (at scale) Apple hardware. If workable CUDA alternatives were really here, I feel people would be talking about them more given how large the hate is for Nvidia on multi fronts from multiple different user groups.
- ivirshup 5y agoI was, but mostly through XLA or Vulkan layers. Also, I’d define workable as a pretty low bar.
- techdragon 5y agoThat’s fair, for me it never worked with any of the hardware or software I wanted it to and I’m not enough of a GPU expert to really fix that myself, so for me and many others I found with similar problems it definitely has yet to become a workable solution, but I was probably a bit harsh if I came across like I thought it didn’t work at all, I’ve seen it demoed and I know it works, but for such a restrictive set of hardware and software that Ive never seen it in use myself outside of demonstrations. As for Vulkan, I tend to think of CUDA at the driver level given how closely coupled it is to Nvidia’s hardware. So I wasn’t really thinking about broader cross compatible APIs like Vulkan which tend to get implemented on top of the drivers providing low level access like CUDA gives… but with the exception of Apple (because we know they will never change their mind) I’m hoping the broader industry gets behind Vulkan compute shaders enough we finally get something that delivers on the promises made when the OpenCL effort began. I want clean understandable abstractions over the top of all the different SIMD and MIMD capabilities we have these days it doesn’t have to magically compile my code down to FPGA hardware but how about actually being able to get the best matrix multiplication performance out of both my CPU or GPU or both if I have a CPU with the right kind of integrated GPU that they can efficiently enough share memory and both work together without slowing each other down when working on an embarrassingly parallel matrix multiplication task. Is this really too much to ask of software in an an era that gives us near magical JIT performance in multiple languages, and manages to build cross compile/transpilation tools that can convert entire assembly code bases into JavaScript and all the other nice things we have as software developers… it never felt like too much to me but the fate of OpenCL appears to disagree.
- torginus 5y agoI never understood the dominance of CUDA. I did GPGPU more than half a decade ago, and I feel like stuff has hardly changed. There are thread groups, group shared variables, global, and group shared atomics etc. Underneath, all GPUs run the same-ish wide SIMD architecture, so I'm pretty sure most stuff is semi-performance portable. DirectCompute, OpenCL and CUDA and probably all the others expose this same programming model. Why is CUDA so dominant?
- michaelt 5y ago> Why is CUDA so dominant? nvidia decided they wanted to be the market leader, and spent $$$$ developing tools, training materials, and libraries (like cuDNN) which they gave away for free. Other GPU vendors chose not to outspend them.
- ivirshup 5y agoThey also hire software engineers to integrate CUDA into open source projects. On the one hand, those projects can now take advantage of advances in hardware. On the other, users will have to buy NVIDIA to use it, so how open is it? Also NVIDIA will not provide any help with CI, so those costs are borne by the project – or the code isn't tested.
- ekianjo 5y agoWhich alternative? When it comes to ML there is no alternative to CUDA.
- ivirshup 5y agoJAX/ XLA seem quite popular, especially since you can use them in Colab on TPUs. IIRC, there are people who have managed get jax compiled with ROCm support. Tensorflow lets you deploy to quite a few backends. oneAPI support in GPUArrays.jl seems to be coming along.
- sircastor 5y agoApple has a very broad and safe license for their Designs. They would largely be unaffected by the merger and likely not change anything. That said, the replacement ISA already exists IMO: RISC-V
- scotty79 5y agoThat's the idea behind all the revolutions, that good things will spontaneously come after we destroy something we have. They rarely do. It's better to create a better thing first, even unfairly, and let it outcompete the incumbent.
- AshamedCaptain 5y agoARM is hardly a "good thing". These things rise and fall almost purely by chance since the differences between ISAs are menial at best. If the stars had aligned differently this could very well have been a SuperH world. Or a MIPS world.
- formerly_proven 5y agoWhat's that about sparc? Or a PA-RISC world, maybe even an AlphaWorld. (Granted, the last two were way less likely than the others)
- Fnoord 5y agoAlpha and PA-RISC died before the smartphone revolution (end 90s vs mid 00s). x86-32 killed them, before AMD64 (x86-64) took off. Laptops were all x86-32/64. There has been an UltraSPARC laptop (one AFAIK), and Apple switched away from POWER because of power inefficiency. Also before the smartphone revolution. Intel was busy with Moblin and PowerTOP back in the 00s. This before iPhone. MIPS and RISC-V also exist. Without ARM it would've happened but slower/later.
- adrian_b 5y agoNone of those 3 even tried to have a product for applications with low cost and low power consumption. The only competition was from Motorola/IBM POWER (and MIPS in some niche products), but neither Motorola nor IBM ever tried to offer low cost products. During 2000 to 2005, I have ported some products with embedded computers from POWER to ARM. The porting was not very easy, because the ARM CPUs had much lower performance, but they were much cheaper, so the conversion was done anyway. The ARM advantage has always been in its business model, which created a lot of competing CPU vendors, willing to give you the best price, not in any technical advantages of its CPU architecture. The NVIDIA acquisition certainly threatens this business model, despite all contrary claims of the NVIDIA management.
- Eelongate 5y agoApple supposedly has a indefinite 'architectural' license to design and manufacture ARM chips. Presumably that cannot be revoked.
- freemint 5y agoDoes that also include access to new versions of ARM not yet released?
- bni 5y agoWhy would Apple even care about those?
- marcan_42 5y agoConsidering they've even gotten away with breaking the rules (custom ISA extensions), I imagine nothing ARM does could realistically cause Apple to be unable to continue manufacturing ARM-compatible chips.
- garblegarble 5y agoYeah, turns out when you're one of the founders of ARM you get a pretty broad license!
- handelaar 5y agoI would imagine that Apple (which put money into Acorn and ARM JVs over thirty years ago and I think was the first third-party company to do so, by a very wide margin) has a much older license than anybody else.