13 ms·
Amid chip shortages, companies bet on RISC-V
- Reitet00 4y agoVery nice to see RISC-V growing like that and being in the area of interest by big names such as Google and Intel. Open solutions are critical for risk management but I wonder if desktop/server RISC-V processors are also planned.
- ghaff 4y agoThere are server specs but embedded was the initial priority. I don't think there's been a lot of attention paid to desktop which isn't a super interesting area in general although it could of course piggyback on server work from a spec perspective if anyone were actually interested in giving it a go.
- mistrial9 4y ago> super interesting area in general deterred by the size of the profits for the winners, and the losses for those who do not compete in global markets.. yet I suggest there are few things more interesting than a personal, general purpose computer
- packetlost 4y agoWithout a company like Apple to put in the resources and engineering effort for something like Rosetta, there's really going to be next to no demand for a RISC-V based desktop computer outside of specialized development and maybe competing with ARM-based Chromebook-likes. Desktop systems are absolutely beholden to their platform and architecture because that's what determines which apps you can run. Any serious use of a desktop (that is not programming) basically needs to use Windows or Mac OS (think CAD, professional video editing, etc.) so you're not just convincing hardware manufacturers: you're convincing thousands of app vendors, and that's just not going to happen.
- dontlaugh 4y agoTrue. But I think more importantly, there will be no demand until someone makes a RISC-V CPU that can actually compete with Intel, AMD and Apple on performance.
- packetlost 4y agoI don't think that's necessarily a prerequisite. At least you could conceivably see demand for low-power, moderate performance Chromebook(-like) devices. For true desktop computing, yeah, I agree.
- worldofmatthew 4y agoA fees dollars saved from not paying ARM will guarantee adoption in low end laptops, tablets and smartphones. $100 Laptops, $70 phones and $50 tablets will be a big target for RISC-V.
- justahuman74 4y agoARM could just start price-matching the various RISC-V vendors, making themselves the easier choice due to the software ecosystem
- worldofmatthew 4y agoThey would have to offer the core for free to the SOC designers.
- humanwhosits 4y agoThe (good) riscv cores designs that implement the spec aren't free, it's the ISA spec that is free. There are some open source riscv cores, but the paid ones make money for a reason
- giantrobot 4y agoAll of those things you mention take advantage of the economies of scale of ARM production. They exist because a few years ago the basebands they're using were top of the line and were used in upmarket devices. They can buy an old baseband/board design, attach a screen and battery, and have a cheapo device for essentially zero development cost. If RISC-V doesn't see the development for upmarket products it's not going to magically take over the downmarket segments. No one footing the development bill is going to selling $50 tablets.
- vbezhenar 4y agoWhat's the difference between desktop and embedded? I reverse-engineered some medical device. It houses tiny CPU and 3" touch screen. Inside it runs Linux with X Window, Chromium, Electron and software written with JS. It's definitely embedded device, it's tiny, works from accumulator, hand-held. But its software stack is not any different from desktop.
- rowanG077 4y agoFor desktop you'd definitely want things like GPU and at least video decode. So that instantly makes it much more involved.
- notsapiensatall 4y agoTraditionally, the main difference was a full-featured MMU which allows virtual address spaces. But these days, you have advanced 600MHz microcontrollers with simple GPUs, and full-featured CPUs which get used as an embedded platform. You can even build a Linux kernel for no-MMU platforms. It's a fuzzy line.
- Symmetry 4y agoThe flexibility that makes RISC-V so compelling in embedded roles (what if I want 64 bit address but don't need hardware floating point?) makes it a harder target for the sort of workloads that you'd usually run on a server or desktop. If I were going to create an open high performance core to challenge x86 and ARM's A series cores I'd probably use PowerPC as a base rather than RISC-V. But I do think that RISC-V has a bright future in other segments.
- mixmastamyk 4y agoWhat are the trade-offs of PowerPC vs RISC-V?
- Symmetry 4y agoIn addition to RISC-V being more flexible than ideal here, PowerPC is more complicated as an instruction set which is a drawback if you're a student learning to implement your first processor or in a deeply embedded role. But given the huge amount of effort that goes into a high performance core it gets lost in the noise and it lets you execute tasks in fewer instructions without crazy difficult to implement levels of instruction fusion.
- brucehoult 4y agoRISC-V does not in any way depend on instruction fusion. I don't why this meme persists, especially as no RISC-V cores in the market do instruction fusion. High end OoO cores in other ISAs are BREAKING DOWN complex instructions into µops. RISC-V is pre-broken down. The one exception to that is current x86 and ARM cores DO do instruction fusion, to combine a compare with a following conditional branch. Which is a singe instruction in RISC-V in the first place. SO it is actually the other way around. Low end cores shouldn't be doing either fusion or breaking down into µops. They are supposed to be as simple as possible, and doing either of those is a complication using significant silicon area and energy. There might be a place for instruction fusion in mid-range cores. Things in the ARM Cortex A53 / A55 / A510 range. The most popular RISC-V core in that segment, the SiFive U74 (used in the HiFive Unmatched, BeagleV Starlight, VisionFive v1, VisionFive 2, Pine64 Star64 .. and probably more yet to be announced) doesn't fuse multiple instructions into one instruction, but it does pair a forward conditional branch past a single instruction with that following instruction. Both instructions still exist, they each go down one of the two execution pipelines side by side, the same as if the branch was predicted not-taken. At the final pipe stage if the branch turns out to be taken, instead of taking a mis-predicted branch penalty and flushing the pipeline, the U74 simply does not write the result of that following instruction back to the destination register. That's the closest any currently shipping RISC-V core I know of comes to instruction fusion. "crazy difficult to implement levels of instruction fusion" does not exist anywhere, and is not needed. It's just an idea from an academic which doesn't actually exist in the real world, at least at present.
- worldofmatthew 4y agoI suspect most of RISC-V development is more towards moving away from western controlled ISA's that can be sanctioned for any reason. Points about avoid shortages from any fab being allowed to fabricate RISC-V CPUs is just the "cherry on top".
- packetlost 4y agoMaybe. It's not like sanctioned countries tend give a shit about copyright law, so how would sanctions prevent the use of something they already know how to make and use? I'm just not convinced this idea holds any water.
- worldofmatthew 4y agoNot breaking copyright makes it easier for a eastern block to form around RISC-V fabrication.
- pdimitar 4y agoCan you explain how? I'm not seeing it.
- fakeslimshady 4y ago
- conradev 4y agoI don’t know if sanctions are the reason, but moving away from foreign-controlled IP is certainly a thing. Chinese companies make STM32 clones (even fixing the errata in the original): https://hackaday.com/2020/10/22/stm32-clones-the-good-the-bad-and-the-ugly/ https://hackaday.com/2020/10/22/stm32-clones-the-good-the-ba...
- justahuman74 4y agoI think it's more about a potential future usage of ARM in a sanctions package. Somewhat similar to how Hauwai was was hit with a list of narrow prohibitions like access to the android store. Having an openly implementable ISA is just one less thing to worry about
- baybal2 4y agoI found it to be completely orthogonal how switching to a different ISA will solve the issue of lack of underlying physical semiconductor manufacturing capacity.
- mlindner 4y agoUsing RISC-V doesn't let you bypass chip shortages... Do they actually understand what they're talking about?
- martin1975 4y agoSometimes it's not about an actual shortage, but more of an imposed (political/business motivated) shortage. If RISC-V can alleviate these concerns, the alleged 'shortage' may disappear.
- mlindner 4y agoThat’s a shortage on the reseller side not the production side.
- worldofmatthew 4y agoAny fab is allowed to produce a RISC-V CPU. You will also have far more core designs that can work across different fabrication possess.
- SV_BubbleTime 4y agoHow many fabs are sitting around doing nothing saying "Oh gosh, I wish we were allowed to make something!". Less than zero, because new fabs being built already have orders and plans for what they'll be making Day1.
- cptskippy 4y agoThe limiting factor of a Fab is the process node. RISC-V doesn't magically boost the capability of large process nodes.
- yywwbbn 4y agoIt’s not obvious that any RISC-V CPUs that might be able to compete with ARM/x86 will be free in the future. It would take a lot of money to design them, why would the company that does that just give it away for free?
- sremani 4y agoAccess to high purity quartz is the real bottle neck. You can have your own fabs and designs.. but if America and Russia refuse to sell you some Quartz.. you are out of luck! https://www.persistencemarketresearch.com/market-research/high-purity-quartz-market.asp#:~:text=One%20of%20the%20most%20significant%20factors%20impeding%20industry,for%20more%20than%2095%25%20of%20global%20HPQ%20production https://www.persistencemarketresearch.com/market-research/hi....
- xadhominemx 4y agoHigh purity quartz is definitely not a bottleneck in the electronics industry.
- sremani 4y agoMay be you are right, but can you add some color and context to your commentary. Like a good article link etc. I am more than happy to chance my mind, if provided evidence.
- SV_BubbleTime 4y agoConsidering that quartz has a ton of limitations, MEMS and other digital oscillators are ready to step in the moment that quartz is cost prohibitive. We're already at a breaking point for low frequency quartz. You literally can't get it small enough for modern packages until you up the frequency. Find a 2x2mm 8Mhz part, I'll wait while you fail. We're JUST NOW starting to get standardization on SMD oscillators in common packages, and these packages have typically pin incompatible (an enable pin instead of a two pin pierce oscillator setup) but same footprint digital alternatives about. Maybe you read an article, but is incorrect to say we're being held up by quartz availability.
- crote 4y agoBest I can do right now is TXC AV08000301, which is 3.2x2.5mm. I'm surprised you even need one like that. Most modern electronics to use clock dividers or multipliers anyways, especially once your chip is small enough that you need a 2x2mm xtal. Or they just have an internal low-accuracy oscillator. In my experience one of the plentiful 4-pin 2520 or 3225 ones usually does the job, and the market has done a decent job standardizing their footprint.
- mysterydip 4y agoWhat I don't get is there's millions of "good enough" e-waste motherboards, RAM, CPUs etc. out there. Wouldn't recycling be better than having nothing or waiting indefinitely? Are there that many workloads that can only run on brand new hardware?
- fipar 4y agoI think that's a good idea but then companies doing that wouldn't be able to sell their products as new, and we probably need a cultural shift before widespread adoption of refurbished and used products is doable. Additionally, and not knowing much about the hardware side of things, if I put myself in the shoes of a manufacturer, it seems challenging to ship a product where the expected lifetime of some of its components is unknown. Support and warranties would be affected too.
- mysterydip 4y agoThose are fair points. With the amount of e-waste currently in existence I'd think we would want to address those sooner rather than later, but it's easy for everyone to just kick the can down the road.
- galleywest200 4y agoIt may be more prudent for us to perfect our material extraction process to reuse these metals and plastics, as opposed to reusing the same piece of hardware. Once that is done, we could probably have e-cycling be picked up by trash collectors just like regular recycling.
- cptskippy 4y agoYou can't really slap an AT motherboard in an LED light bulb and call it a day. Fabs produce all manner of chips, not just GPUs and CPUs. Over the last 40 years all manner of circuits composed of discrete components have been replaced with chips. Voltage regulation is a chip, battery protection is a chip, rectification is a chip.
- tregoning 4y ago"Yeah. RISC is good" .... https://twitter.com/Hackers_bot/status/1547829853231075331 https://twitter.com/Hackers_bot/status/1547829853231075331
- kyaghmour 4y agoThere's so much more to a chip than its ISA. I'm not sure how the ISA changes anything. If I design a board around an STM32, for instance, even within that family I'd have to find a chip that has the same exact footprint to be able to replace one that I couldn't find. And even then, I'd like have to reconfigure the software ... So unless there was a pin-for-pin/electrical standard for chips, the ISA is of no effect.
- Lind5 4y agonot just the chip shortage. The rise of RISC-V coincides with a couple of other events in the industry. The first is the slowing of Moore’s Law, meaning that increases in total processing power no longer comes along with each new fabrication node. The second is the meteoric rise in machine learning, demanding massive increases in processing power. https://semiengineering.com/why-risc-v-is-succeeding/ https://semiengineering.com/why-risc-v-is-succeeding/
- joe_the_user 4y agoHmm, Is there any discussion of how RISC-V designs could be incorporated into a GPU or TPU that could train deep learning systems? Your link doesn't say anything about that but it's an interesting question.
- arnaudsm 4y agoLanguage models and image generation make fun demos, but do we have transformative use cases that'll actually require large ML compute in the future ? Voice recognition and translation are the only ones that comes to mind, yet don't require that much power.
- mise_en_place 4y agoI am personally very excited for RISC-V. I like the boot process of OpenSBI/U-Boot, can also directly boot a Linux kernel. So far I have only used the qemu virt machine w/ riscv64 cpu. OpenBSD and Debian/Ubuntu have great support, I have ordered a VisionFive board, curious to see how well it runs on physical hardware
- nikolay 4y agoRussia's new main arch, I guess.