Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Android Ported to RISC-V

theregister.com

Post reply on HN
Print view View original

61–70 of 98 posts

Re: Android Ported to RISC-V

#61

Post by thereddaikon »

AlexAltea wrote: post #8

Related: https://www.cnet.com/news/chip-designer-arm-ditches-huawei-b... Unsurprisingly, this has been done by Alibaba-owned T-Head Semiconductor. The obvious result if Chinese vendors start facing obstacles to obtain licenses to manufacture ARM-based chips (or any other non-Chinese architecture). How can anyone expect that software/hardware-bans to global superpowers can result in anything but their software/hardwar…

ISA's are to hardware what API's are to software. They simply provide a spec to conform to. ISA's say little to nothing about the hardware implementation itself. That's the real value and something that every firm that has CPU IP guards.

And sadly China isn't lacking in stolen valuable CPU IP. The Chinese branch of ARM went rogue last year. I think its a safe assumption that any sensitive information they may have had is no in the hands of the CCP and will be used to its full potential.

Re: Android Ported to RISC-V

#62

Post by analognoise »

sudosysgen wrote: post #60
simonh wrote: post #59

Earlier quoted context omitted.

Ok, I suppose on those terms we'd need to know the real cost to SpaceX after they have recovered the first stages and used them for a dozen satellite launches each. The fact is on commercial terms SpaceX has been undercutting the Russians for years now, and were doing so even before first stage and fairing recovery. Edit: It looks like NASA has already agreed to allow SpaceX to re-use first stages and Crew Dragon cap…

We really don't. Roscomos doesn't make the rockets, it's the Russian NASA. The cost here is because of the middle man and the US having been dependent. Be that as it may, today in 2021, the cheapest and most reliable way we have to send someone to LEO is to make a Soyuz and use that to send them. Maybe eventually SpaceX will catch up in cost. But in 2021 it hasn't yet.

So you're arguing it's cheaper in the case where nobody openly knows Roscosmo's numbers, nobody can obtain the services at those numbers, then comparing the cost at the FACTORY for Soyuz to the price of LAUNCH for SpaceX - before accounting for using the same rocket TEN TIMES - and still claiming Soyuz is cheaper?

Edit: Parent filled in data below, it looks like Soyuz in cheaper. :)

Re: Android Ported to RISC-V

#63

Post by alexhutcheson »

bhouston wrote: post #53

I expect RISC-V to match Apple's ARM implementations efficiency/speed within the decade. This is a long time but it is a very lofty goal as Apple's ARM implementation is a moving target to RISC-V has to move faster. But it likely will come close to converging around 5 years from now -- at least somewhat competitive in core count and single core performance. Interesting times. I look forward to the first AMD Ryzen-lik…

That might be true, but I wouldn't expect the microarchitectures that achieve this performance to be free (in either the "gratis" or "libre" senses of that word). Designing a superscalar application processor requires an enormous amount of work by engineers who are experts in their field, and that work needs to be compensated somehow.

The ISA may be open, but there is no requirement for the microarchitecture implementing the ISA to be, and it's hard to imagine anyone doing the work to produce a performance-competitive chip and then giving it away. Even if there is a "consortium" model where multiple companies benefit from a single design, they would likely set up some sort of licensing arrangement to prevent free-riding by companies that did not contribute to the design.

Re: Android Ported to RISC-V

#64

Post by AtlasBarfed »

AlexAltea wrote: post #8

Related: https://www.cnet.com/news/chip-designer-arm-ditches-huawei-b... Unsurprisingly, this has been done by Alibaba-owned T-Head Semiconductor. The obvious result if Chinese vendors start facing obstacles to obtain licenses to manufacture ARM-based chips (or any other non-Chinese architecture). How can anyone expect that software/hardware-bans to global superpowers can result in anything but their software/hardwar…

Yeah, will NVidia's acquisition result in a real push to RISC-V?

Re: Android Ported to RISC-V

#65

Post by ddingus »

varispeed wrote: post #13
d3ckard wrote: post #11

Earlier quoted context omitted.

That's good though. That means they have to invest in R&D instead of depending on exploiting cheap labour to brute-force their way into the market. If they invest in R&D, they need creative people who are up to date. Such people tend to bring change to the society at large(I'd say democracy is really good for creativeness to flourish). Basically, the goal is not to expunge them from the market, the goal is to make th…

Do you know that nothing stops them from rounding up engineers are move them to a camp with an order to develop certain technology in exchange for bread and water and that their families will live? Soviet Union for example used to build entire cities where they crammed engineers from all over the Union, however rather than being exploited, the engineers had special privileges. But that easily could change with a stro…

They can also round them up and let them live lavishly.

There still are walls, limits, secrets of course.

Just saying there are any number of ways to advance and they are all available to a sovereign willing to get after advancement. Size matters too. China has size enough for these things to be true.

Re: Android Ported to RISC-V

#66

Post by sudosysgen »

analognoise wrote: post #62
sudosysgen wrote: post #60

Earlier quoted context omitted.

We really don't. Roscomos doesn't make the rockets, it's the Russian NASA. The cost here is because of the middle man and the US having been dependent. Be that as it may, today in 2021, the cheapest and most reliable way we have to send someone to LEO is to make a Soyuz and use that to send them. Maybe eventually SpaceX will catch up in cost. But in 2021 it hasn't yet.

So you're arguing it's cheaper in the case where nobody openly knows Roscosmo's numbers, nobody can obtain the services at those numbers, then comparing the cost at the FACTORY for Soyuz to the price of LAUNCH for SpaceX - before accounting for using the same rocket TEN TIMES - and still claiming Soyuz is cheaper? Edit: Parent filled in data below, it looks like Soyuz in cheaper. :)

We actually can optain the numbers. Roscomos is a public agency. They buy the rockets from the Progress join venture, and a public the exact cost of both the rocket and the launch. The cost of only the rocket is around 5 million dollars.

The total cost in 2018 for a Soyuz launch to Roscosmos is 3.5 to 4 billion 2018 rubles, which is around 20-23 million dollars per seat.

The cost is kind of at the factory because Roscomos partially builds some systems of Soyuz. But in the end it costs Roscosmos 20 million to put someone into space using a Soyuz, and it costs NASA 55 million to do it via SpaceX.

I'm comparing the cost of launching an astronaut to the ISS. For Soyuz, since the rocket is not reusable, this is the cost of the rocket plus capsule plus payload plus préparation, tracking, etc.. For SpaceX, it is just the cost of their launch. Soyuz is simply so well engineered that it still is cheaper, even though the SpaceX rocket is reusable.

Re: Android Ported to RISC-V

#67

Post by fuoqi »

varispeed wrote: post #12
AlexAltea wrote: post #8

Related: https://www.cnet.com/news/chip-designer-arm-ditches-huawei-b... Unsurprisingly, this has been done by Alibaba-owned T-Head Semiconductor. The obvious result if Chinese vendors start facing obstacles to obtain licenses to manufacture ARM-based chips (or any other non-Chinese architecture). How can anyone expect that software/hardware-bans to global superpowers can result in anything but their software/hardwar…

The problem is that restrictions are not thought through. If they were banned from using Android altogether, they wouldn't be able (legally) to port it on that architecture in the first place. The same goes with the RISC-V as well - the law could be written to allow only certain countries to use it freely. These obstacles are kind of similar to patents. When you discover something and then you learn someone was first…

>The same goes with the RISC-V as well - the law could be written to allow only certain countries to use it freely.

Luckily, the people behind RISC-V were smart enough to protect it as well as they could against politicians using this open technology as a political tool. Not only it uses an open license, but also IP is owned by entity registered in Switzerland, so people and organizations all around the globe can collaborate and use results without being afraid to fall under yet another sanctions, which more often than not are simply a salvo in an ongoing economic warfare.

Re: Android Ported to RISC-V

#68

Post by justincormack »

zozbot234 wrote: post #33
kls wrote: post #31

Earlier quoted context omitted.

was the number 1 reason why the South had a garbage economy I am not sure what the implication of garbage economy is, but the south was fairly wealthy antebellum. So much so that they supported a large tax base for the federal government which they resented, it lead to the nullification crisis in 1832 and started the long road to the civil war. Southern universities where also considered to be some of the best in the…

> the south was fairly wealthy antebellum. Not so much when you properly account for the enslaved population. It's just not sensible to only count free citizens, that's clearly a misleading impression.

Its worse than that, its not about not counting the slaves, the slaves literally are the wealth they were counting, and were the largest capital asset in the country.

Re: Android Ported to RISC-V

#69

Post by BobbyJo »

coldtea wrote: post #48
sharpneli wrote: post #37

Earlier quoted context omitted.

> Using cheap labor, copying foreign products, and bypassing copyright worked for the US, Japan, Korea, and Taiwan when they were in their own economic rise, and thus far has been working for China. That's what makes the arguments from USA so amusing. I mean they basically nicked, against British law, the designs how to mass manufacture textiles. Is US willing to repay, with interest, all the economic damage that did…

> I mean they basically nicked, against British law, the designs how to mass manufacture textiles. And used slaves to pick the cotton! Talk about "cheap labor". It might sound quaint now, but textiles then where one of the engines of industry growth. The world didn't have mass textiles/clothes production until the industrial revolution. It's like "Yeah, my ancestors bootstrapped this country doing that, and I benefit…

I mean, if the chasm China had to cross was 'not have mass manufactured textiles' -> 'having mass manufactured textiles' I don't think anyone would have an issue. And we didn't. Why do you think China rose out of poverty so quickly?

Now, China is not only technologically competitive, but an authoritarian regime. Theft of IP no longer represents trying to get to where the rest of the world is, but rather trying to dominate it. At the same time, every advantage the CCP gets is quickly utilized not just to intimidate neighbors, but to oppress the Chinese people.

Even if we go so far as to say the situation is exactly the same otherwise, (I might get hate on HN for this) 'why' you do something matters too. If you do it for freedom and democracy vs. authoritarianism, that matters.

Re: Android Ported to RISC-V

#70

Post by agurk »

kls wrote: post #29
numpad0 wrote: post #20

Earlier quoted context omitted.

There is sense of mystique associated with the Soviet Union but the best they managed to build were AKs, Soyuz, Chernobyl, all interesting and robust in its own ways but not all that competitive in terms of performance, so,

I understand the sentiment you are trying to express and to some degree I agree but I would argue that the AK is a really bad example that actually is probably the greatest example of the Soviet simple utilitarian technology. Out of all of the battle rifles designed (The big three being the AR, AK and FAL/SCAR) the AK has gone under the least amount of revisions, the biggest revision being the AK-74 where the Soviets…

> of all of the battle rifles designed (The big three being the AR, AK and FAL/SCAR)

To be pedantic about terminology the Kalashnikov family of rifles are generally not Battle Rifles. Of your listed examples only the FN FAL and some models of the FN SCAR are considered Battle Rifles.

Battle Rifles as a term came out in the 80s to distinguish between autoloading rifles soldiers had fielded since WW2 that weren't assault rifles. Assault rifles are defined as a selective fire rifle using an intermediate cartridge with a detachable magazine.

An intermediate cartridge being one that is more powerful than a pistol but less so than a traditional rifle cartridge. Germany pioneered, or at least mass-adopted, this concept with the 7.92 Kurz [short] round in WW2. The USSR ran with this concept - through seeing the STG-44 (German 7.92 kurz assault rifle), capturing some German weapon designers and following the train of thought of the period. This ended up with the AK-47, then later variants on this pattern.

NATO countries generally also wanted to do this, see the .280 British - the original cartridge of the FAL (and the EM-2). With one big exception - the US Army Ordnance Corps pushed for adoption of a traditional cartridge, and NATO bowed (in part as the US agreed to adopt the FAL) and everyone ended up with the full powered 7.62 NATO round.

Despite pushing for it, the US soon found out that having a high powered cartridge (and the sort of weapon that goes with it) wasn't what they actually needed, and they chose the 5.56 NATO round as used in the original M16 rifle (AR-15 platform). The rest of NATO caught up soon enough, often with platforms based of the AR-18/AR-180. This left them in the 80's with a new type of rifle that we would call an Assault Rifle but they also needed to refer to the (more) modern rifles they'd been fielding since the 2nd World War.

This resulted in the term Battle Rifle being created to refer to a self-loading rifle using a full power cartridge with a slew of (optional) features like full auto (the SLR [UK's inch pattern FAL] was semi-auto only), pistol grips (the M14 had a traditional stock) and the like. The 7.62 NATO calibre versions of the FN SCAR are generally considered Battle Rifles, whereas the 5.56 NATO versions aren't.

Post reply on HN