raptor talos will make a lower priced mainboard, if you buy 10000
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
https://www.raptorcs.com/
They manufacture the libre software talos mainboard.
I emailed them, asking if they would make an about 500usd libre software mainboard. Raptor answered me. Rather good answers in terms of the possibility of getting a lower priced libre software mainboard.
Raptor wrote, they are technically able to sell a libre software powerpc mainboard having lower specification for a lower price if they got an order about more than 10000 mainboards. Raptor did not say, they could do it for 500usd per mainboard.
Raptor wrote, they have refrained from making such mainboard, because raptor assess they cannot sell the mainboard and making it would be a waste of money.
Raptor said, should they get an order about more than 10000 mainboards, payment to raptor secured, they would manufacture them.
I was not aware it is possible to make an all libre software mainboard, though it is not x86.
Raptor in plain language said, if you do a crowd funding, goal more than 10000 mainboards, raptor likely are going to manufacture the mainboards.
This tells me, it is up to the libre software people to get a libre software mainboard. We will probably not be able to organize a crowd funding, because who will head it? It has to be an entity, which is known to be reliable. The free software foundation should do it. Previously I have asked fsf to make a stand on organizing crowd fundings of hard ware. They did not.
Who can you write on fsf? Several times I have written name at domain about campaign suggestions. Not once have fsf been able to handle my emails effectively and clearly. Can you write Stallman?
This is very cool! I would think the free software movement should be able to get 10,000 signatures together to crowdfund this. We would need to get a firmer quote from them in terms of price/lead time; what the exact specs would be and verify that it would be 100% libre. If the price really is $500 for decent specs then I'd certainly put my name down (although I assume that wouldn't include a CPU?).
I agree that it would be great if the FSF could get involved with organizing the crowdfunding, as surely something like this would be a big deal for libre computing. Have you seen Stallman's personal site? There is an e-mail address there where you can reach him.
This would be awesome and I really hope it happens! :-)
Not just signatures but 10 000 orders with payments sent to Raptor to secure the orders
Just need to know how much for a mainboard
Is Raptor's product different (in terms of freedom) from Think Penguin's EOMA68?
I haven't gotten involved with any crowdfunding before, but I had a quick look at Kickstarter. It looks like the way it works is people pledge a certain amount of money, but no money is charged unless the target is met.
So, it seems like there isn't anything lose by giving it a try ..
I took a look at the EOMA68 (that looks very cool as well, btw). That's a laptop using a dual-core ARM Cortex 1.2 GHz chip, which looks similar to what is used in the Raspberry Pi. The Raptor board would be for a desktop and the Power9 CPU is more recent, so I expect it would be quite a bit more powerful (especially if it has more cores). In terms of freedom, it appears that both have applied to the FSF for RYF certification, but neither is listed on the FSF page yet, so I assume they are both pending.
Thanks Time4Tea -- that seems like a good way of coordinating this kind of effort
On 30/09/17 12:38, name at domain wrote:
> I took a look at the EOMA68 (that looks very cool as well, btw). That's
> a laptop using a dual-core ARM Cortex 1.2 GHz chip, which looks similar
> to what is used in the Raspberry Pi. The Raptor board would be for a
> desktop and the Power9 CPU is more recent, so I expect it would be quite
> a bit more powerful (especially if it has more cores). In terms of
> freedom, they both seem to be endorsed by the FSF, so they are both free.
EOMA68 has a power budget of 5W. its for low power efficient
SOCs/Processors. EOMA200 would have a high power budget for those little
power hungry monsters. Saying that, with good enough computing quad to
octa core 4gbram little beasts are available that fit the power budget
of eoma68 just about ;).
Heres what I posted in another thread:
Latest computer card luke has been working on:
RK3288 Quad-core, 4GB Ram, 32GB Flash:
*Base Info: http://rhombus-tech.net/rock_chips/rk3288/
*Progress: http://rhombus-tech.net/rock_chips/rk3288/news/
*Campaign
Update:https://www.crowdsupply.com/eoma68/micro-desktop/updates/prototype-progress?utm_source=EOMA68+Supporters&utm_campaign=7bde4d48f5-UPDATE_EOMA68_2016_12_27&utm_medium=email&utm_term=0_5e30e1da06-7bde4d48f5-112334465
Nexell EOMA68-S5P6818 Octa-Core 2GB Ram (I think):
http://rhombus-tech.net/nexell/s5p6818/
http://rhombus-tech.net/nexell/s5p6818/news/
a mailing List post:
http://lists.phcomp.co.uk/pipermail/arm-netbook/2016-August/011771.html
I'm a bit surprised there isn't more interest in this. Would anybody else be willing to pay $500-600 upfront for a fully open, modern mainboard alternative to Intel/AMD? Do people not think that target of 10,000 units could be achievable by crowdsourcing? I admit that I don't have a very solid handle on how big the free-software community is, but that price range seems much more accessible than the $2,300+ that Raptor are charging currently for their Talos II Workstation boards, which were mentioned in another thread. 10,000 doesn't seem like it would be impossible.
Another question though: what is the status of this board/chipset regarding firmware? Has Coreboot/Libreboot been ported to it? If a Talos board landed on my doorstep tomorrow morning, could I actually run it? I assume yes, if they are currently selling Talos II ;-)
>get a firmer quote from them in terms of price/lead time; what the exact specs would
My email to raptor was introductory. I did not list specifications and raptor did
not mention specifications either in their answer. I did not expect raptor to answer or
answer rejecting.
I think the price is the main issue. It has to be low enough, that most libre software
people can buy if they want to.
Specifications about a duo core x86 cpu, I should say suffice.
>verify that it would be 100% libre.
My impression is, all the source software regarding a ppc is available. The license
terms I do not know. Maybe the software license cannot get fsf approval. In that
case for certain fsf is not going to organize a crowd funding.
I have written think penguin, informing them about raptor. Think penguin supports
the pc card. Maybe that can stop think penguin from considering a raptor
mainboard.
I am going to write Stallman an email about raptor's lower priced mainboard.
> Specifications about a duo core x86 cpu, I should say suffice.
I'd be surprised if they would use a different CPU chip to the Talos II. It seems doubtful that IBM would make a custom Power9 CPU/socket just for a one-off run of 10,000 boards. I don't know, perhaps they would? Again, we would need to get more specific details of what Raptor have in mind.
I have to say, I'm a bit confused about the CPU that is being used with the Talos II. Raptor's website seems a bit scanty on details, except that each CPU package has '4 cores and is SMT4 capable'. However, according to Wikipedia's Power9 page and this PDF from IBM, the Power9 is only being made in 2 versions: 12x SMT8 cores or 24x SMT4. A 24-core CPU sounds like it would be massive overkill for a home desktop PC. I don't know, maybe that info is out-of-date, or Raptor have some agreement with IBM for a custom chip? Would be interesting if we can find out more ..
Anyway PowerPC is more powerful than a x86 CPU, and it doesn't have AMD PSP or Intel's ME shits. But PowerPC always was expensive.
Hi Tonlee,
> I am going to write Stallman an email about raptor's lower priced mainboard.
I applaud your efforts and encourage you to write others as well in hopes of building interest for a advanced Libre platform.
I often think about the momentum that comes from the gaming communities (developers/distributors/gamers/etc.) and wonder if Steam would have a comment or interest
http://store.steampowered.com/curator/6981492-Best-Linux-Games/?appid=324260
I do not know if any gaming community has ever expressed specific support to FSF or the Libre concept, but I hope it happens and a powerful Libre computer available from major distributors could help.
This FSF post: http://www.fsf.org/bulletin/2010/fall/the-free-game-lag
is a bit dated and a recent state of the Free gaming report would be welcomed.
I dream of freed game developers making amazing game changes and influencing how future generations imagine the freed world.
Thanks Heather.
What is #lgn ?
My quick web-search revealed:
https://libregamenight.xyz/
https://libregamewiki.org/Main_Page
I wasn't aware of these sites, perhaps you have more urls you can share?
And - oh for sure do not under-estimate gamers - because I believe gamers can Save the WORLD:
Ask Jane McGonigal What Is “Gameful”?
https://ronmamita.wordpress.com/2013/08/22/what-is-gameful/
Imagine playing the game of life like a awesome super-gamer; defeating the challenges to worldwide freedom FTW!
That is the message I share with the youth and game developers.
I have written think penguin asking if they would be interested in endorsing a campaign and
crowd funding about a lower priced ppc libre software mainboard. Think penguin
answered yes, if we could come up with a midway acceptable plan to get 10000
mainboards sold. We both agree, we may not achieve the goal. Then we can say, libre
software people had an option and did not jump on it.
In another email to think penguin I sketched my approach, which I refer here.
We will not be able to sell a mainboard to people, who do not know about libre software.
Only those who already want a libre software computer will buy a mainboard. The task is to
inform them.
A campaign has to be scheduled to start on a set day. That day every one involved in
the campaign will know who to inform and how.
The campaign will have two groups of actors. Mass and front runners. Mass would be people
on trisquel's forum. We would coordinate the task to email various websites in different
languages. Front runners would be fsfe, fsf, think penguin, Snowden, Thomas Drake,
Rutkowska, Assange endorsing the mainboard. We will have to seek their support. Fsf could
play a role in this task. Snowden has 3000000 followers on twitter. He may response to
an email from fsf.
The mainboard to be manufactured will aspire to 2 requirements. A performance comparable
to an intel duo core x86 cpu. And the lowest achievable price. In result the mainboard will
come with no extras included in the price.
Raptor told me, they assess 10000 buyers are not present. Therefore raptor will not be part
of a campaign or crowd funding. They say, show us the money and raptor will make the
mainboard.
In raptor's email, they indicated, that they can make both atx and notebook mainboards. I
suggest a notebook mainboard should be made. A notebook mainboard would allow
people to explore making their own notebook. Instructions telling how to turn the
mainboard into a desktop should be provided.
I am in favor of getting the crowd funding run by think penguin. I do not know whether
think penguin will do it.
I have asked raptor to list an overview of specifcations.
Remarks or suggestions?
Hi tonlee, seems like you're doing some great work on this and I think your plan for campaigning sounds good. I have a couple of comments though:
Firstly, I strongly recommend going for a desktop ATX board over notebook. An ATX board is very standard, so most people who already own a desktop PC should be able to switch it in fairly easily, without having to buy any new hardware (other than the CPU). If you go for a notebook, that will either restrict it to those people who already have that specific hardware, or mean that people would have to go and buy a new laptop as well, in order to use it.
I'm surprised they even mentioned a notebook board. I would think designing an ATX board would be much easier, because it's nowhere near as constrained in terms of space and thermal requirements as a notebook (not that I know very much about PC board design ..).
Secondly, about the performance part - I've been thinking about this some more and I really think it would need to be better than an Intel Core 2 Duo. I have a laptop already that has a Core 2 Duo and is 100% libre, so if your board has a similar spec chip, then what is that giving me that I don't already have? A big part of the excitement over the Talos II is that it is using a libre chip that looks like it will give comparable performance to more recent Intel/AMD chips. I think the project would struggle to gain support if it isn't raising the bar of what free computing can do.
+1
"what is that giving me that I don't already have?"
I'd love me some >=Haswell-levels of battery life
Ok, fair enough. But, would you be willing to pay $500 for it, if it's not giving you any increased computing power over what is already available? I wouldn't, personally.
I may not commit to another motherboard (not sure) but I am willing to promote the campaign via blogging and dropping comments in forums.
Please keep us informed with price, urls and dates.
I wrote think penguin for remarks on my plan how to find 10000 buyers. He has not answered. I have resend the email. I wrote raptor for preliminary specification options. No answer. I have resend the email. If communication stops, I think we will get no further.
Maybe people should write think penguin and raptor to show insterest.
Has anyone contacted Richard Stallman yet? he's E-mail is: name at domain and he's active! he replied to my message today.
I am a translator!
>Has anyone contacted Richard Stallman yet?
Several times. Each year I send him a nice brief text with a gnulzy wallpaper in the attachments :)
I got an email from raptor.
Raptor wrote, they will only produce hardware which is approvable by the free software foundation. They will not state specifications of a mainboard. If gathering the pieces of information means costs to raptor, it is understandable. If a crowd funding is initiated, raptor will have to display specifications because people will want to know what they buy. Raptor is not in favor of a notebook mainboard. Raptor does not elaborate. It seems making a notebook ppc mainboard is more difficult. Requiring bigger sales numbers in order to keep a lower price. Instead raptor suggest a micro atx mainboard. I would prefer a notebook mainboard enabling people to make their own portable or notebook solutions.
On 11/10/17 15:39, name at domain wrote:
>
> Instead raptor suggest a micro atx
> mainboard. I would prefer a notebook mainboard enabling people to make
> their own portable or notebook solutions.
>
I suspect/take a guess that part of the problem is power consumption.
Which means a lot of cooling and for laptops that means not a generic
off the shelf heatsinks/cooling systems but custom designed and made
cooling system, i guess. Along with all the addition bits and
complications in making a notebook.
After watching luke making a eoma68 laptop, it is a lot of work. When if
you want portable you probably want something lower power with longer
batt life. Note that eoma68 has a power budget of 5W for the computer
card wheres i guess (idk) the talos will be double digit power needs.
A ATX board pops into existing PCs. so they just have to do the basic
stuff (the big power supply brick, for example. is already done for
them, laptop would mean diy power supply.) and i guess less trouble
fitting it all in the space of the atx pcb.
It’s a struggle to do anything. The challenge is to make something at all.
Anyway heres an existing powerpc laptop project:
https://www.powerpc-notebook.org/en/
It was mentioned/talked about on the eoma68/rhombus-tech/arm-netbook
mailing list:
http://lists.phcomp.co.uk/pipermail/arm-netbook/2017-September/014695.html
Thanks for the update, tonlee, although it is discouraging that Raptor will not be able to provide specifications. In my opinion, for the crowdfunding to have any chance at all, they would have to provide a firm quote first, including specs, price, lead time, warranty details. As you say, people would need to know what they are paying for.
I am not surprised that they are not in favor of a notebook board, as that would involve additional challenges, as I mentioned before.
Did you have any luck in contacting RMS or the FSF? Given Raptor's position, I wonder if the best way forward might be to kick off some sort of online petition, to gather signatures of those who would be willing to buy a new Power9 board for $500, to try to gauge if the interest is really there. If we could get 20-30,000 signatures, that might help to convince Raptor that the project is feasible.
https://www.powerpc-notebook.org/en/
This might be cheaper to support ;)
Manufactoring a notebook is not less expensive, than a mainboard.
In its latest email, think penguin says, it will not be part of a
ppc raptor lower priced mainboard crowd funding. Think penguin
believes it is impossible to find 10000 buyers of the mainboard.
My point of view is, it does not matter, if a ppc libre software
mainboard crowd funding is unsuccessful. Rather it is about
giving libre software people the option to buy a new libre
software computer. If volunteers did the campaigning, it would
not be an expensive crowd funding.
Raptor would have expenses devising a technical description of
the mainboard. The free software foundation would have
expenses reviewing raptors technical papers on the mainboard,
in order to make a decision about endorsing a lower priced
mainboard.
Now the situation is, we have a company which says, they can
and will manufacture a libre software computer, if they get
an order of 10000 mainboards, about 500usd each. But will
not do a crowd funding.
Another company, which I assessed could be the frame of the
crowd funding, will not do it.
What do you suggest? I will write Stallman in order to get
his opinion.
I think this is a pipe dream. POWER seems to be a candidate for the future of high-end computing, the really expensive stuff that businesses depend on. If the cost goes down to a few thousand dollars for a whole computer, Raptor could even provide high-end libre-friendly PCs, and that would be fantastic to have as an option. But for low-cost computing, EOMA68 is our best bet right now. And it's the low-cost stuff that sells more, because almost anyone can buy it (whereas only well-off people can buy really high-end, expensive stuff).
>POWER seems to be a candidate for the future of high-end computing
My impression from raptor's emails is, the main parameter is number of mainboards you order or raptor
sells. I have not asked raptor but I do not think ppc is inherently more expensive than similar x86 hardware. If you buy 300000 mainboards, why should the price per mainboard be much higher?
>cost goes down to a few thousand dollars
The mainboard in question here is about 10000usd for a ppc mainboard, cpu included, if raptor gets an
order of 10000 mainboards. It is a micro atx computer. If you
already have an atx computer, there
are no more costs. Maybe apart from a ram module.
>pipe dream
That would be the buyers decision. What was new to me is, you cannot make a new libre software x86
computer. You can manufacture a new ppc libre software computer.
The eoma has several difficult obstacles. To me the main
problems are performance and planning.
The current eoma is about an arm a20 cpu. It is nothing more than
a raspberry pi. It has nothing to do with a x86 notebook in
terms of performance. Have you used a raspberry pi?
That would not matter if there was a set plan for releases of
better performing pc cards. Which there is not. There is
a rk3288 which is not a x86 regarding performance.
If hewlett packard decided to make the pc card, they would have
the technical and financial volume to say when what hardware
would be on the market. And it would likely hold.
So far lkcl has to wait and see what relevant hardware gets on
the market.
> I do not think ppc is inherently more expensive than similar x86 hardware.
I didn't say it is. x86 is what I was thinking of when I said "high-end". You know, like gaming computers. The really expensive, cutting-edge stuff that people like gamers buy.
For lower-end stuff, ARM is largely overtaking x86; it was only there to begin with because of how immensely powerful Intel is. I don't think it's realistic to expect a new architecture to cover both the cheap computing niche and the high-end computing niche. It could happen, but it's far more realistic to focus on one.
> That would be the buyers decision.
No, it would be their wallets' decision. Most people can't afford to just chuck thousands of dollars on luxury devices. But a hundred dollars or two? Sure.
Note, $500 is only for the mainboard. It does not include the cost of other components.
> The current eoma is about an arm a20 cpu. It is nothing more than
a raspberry pi. It has nothing to do with a x86 notebook in
terms of performance. Have you used a raspberry pi?
No, but I've used the original OpenPandora (still do, actually), and its CPU power is perfectly fine. The A20 doesn't seem to be any worse. My experience with the OpenPandora is that its CPU is more than capable of not only all basic computing tasks, but many games as well. The main limitations are actually the low resolution (800x480) and the very low amount of RAM it has (256 MB).
In fact, both systems have ARMv7-A cores. The A20 is a dual-core Cortex-A7, while the OpenPandora uses a single-core Cortex-A8 (coming from its OMAP3530 SoC). There's a nice comparison chart here:
https://en.wikipedia.org/wiki/Comparison_of_ARMv7-A_cores
Each has its own advantages, but they are basically incredibly similar. So simply the fact that the A20 is dual-core should automatically make it a better CPU than what the OpenPandora has, at least when it comes to general purpose computing. Having 8 times as much RAM as the OpenPandora is even more of an advantage; 2 GiB is hardly phenomenal, but it's plenty to handle basic multitasking. The laptop housing being released is also a much more comfortable 1366x768 resolution, more than twice as many pixels.
So, in conclusion: I currently use a much worse system than the A20 computer card will be, so I know what the system will be like, and I can tell you without a doubt that it is perfectly reasonable for everyday computing, and even capable of plenty of basic gaming. Also, don't forget that it's upgradeable. The A20 card is just the first one.
>10000usd for a ppc mainboard, cpu included
Erroneously I wrote 10000usd. It is 1000usd.
My main argument is, if it was on sale in shops for
1000usd and you asked all gnulinux aware people, if
they wanted to buy one, with some probability
10000 units could be sold.
Main obstacles are, people are reluctant to buy an
about 1000usd item which has to be made from a
company they probably do not know. People do not
know about this option to get a libre software
computer. People want portable devices.
I agree. The EOMA doesn't excite me the way the POWER9 does. Not only is it a new architecture, it is the first architecture for a long time that fully supports libre computing. Like all of this stuff, cost will come down if it takes off and higher quantities get made.
I don't see what the EOMA offers me that my librebooted Macbook 2,1 doesn't already do, tbh.
That's because you have a preference for power. I have a preference for low price and low energy usage. That's fine; we have different preferences.
The problem with comparing EOMA68-based computers to your hacked Macbook, however, is that your hacked Macbook isn't new. No more of those will ever be made. So for those who can't get their hands on one, that just isn't a solution. The Libre Tea computer card, however, is. Raptor's POWER-based systems are, too, but the problem is most people can't afford what it costs. They can, however, afford the much cheaper EOMA68-based computers. That's why I think Raptor needs to get most of their head start elsewhere, in some niche market of people that can afford the initial massive costs.
>you have a preference for power.
No, he has not. As I have stated before, the cpu on the raptor microatx mainboard will not be
a particular fast one. It will be the lowest priced cpu, which raptor can provide, which has a
performance comparable to a x86 duo core cpu.
In a previously post, you have said, how you consider the pc card's cpu fast enough
to function as a daily use notebook for you. It may be so.
The pc card's cpu is nothing more than a slow phone regarding performance. And it will
for almost no one be fast enough to perform as a daily use notebook. I had a raspberry pi.
It was that slow, it was impracticable to use. I know no reason, the current pc card should
be different. The market gives the same answer. If the cpu in the pc card would be regarded
to be fast enough by consumers, we would have the pc card's cpu in our notebooks, but we
do not. To me it is silly to claim the current pc card can act as a daily use notebook.
The problem about the PFY Laptop Housing Kit is, if you insert the current pc card, you
do not get performance equivalent to a current x86 notebook. When will arm cpus be as fast
as x86 notebooks and when will such arm cpus be libre software capable? Lkcl cannot tell us.
>can afford the initial massive costs.
Calling 1000usd for a mainboard and cpu for massive costs is tendentious. Many people
can pay such amount for a computer, should they want to get one. It is more a matter
if people, who has the money also have an interest in libre software computers.
That is why I believe it is important to advertise the raptor microatx mainboard to
computer security circles.
In the event the raptor microatx mainboard and cpu gets sold in a higher volume,
than 10000 units the price will go down.
The fact, that we cannot provide a libre software raptor mainboard everybody is able to buy,
should not stop us from getting it manufactured. If the price for the raptor atx mainboard
and cpu lands on about 1000usd, I am not sure I am getting one. But that does not stop me
from wanting the mainboard to be made because every libre software mainboard getting on the
market is an improvement. Getting the raptor microatx manufactured also ensures the
required knowledge stays present.
I am not in favor of the ppc architecture. Actually I do not know much about ppc. But it
is an option and therefore must be pursued.
I would want the riscv cpus to get on the market. But riscv has stalled before. We must
look into every available option.
> As I have stated before, the cpu on the raptor microatx mainboard will not be
a particular fast one. It will be the lowest priced cpu, which raptor can provide, which has a
performance comparable to a x86 duo core cpu.
I think it's important to clarify on this. Did Raptor tell you that the CPU is only going to be comparable only to an x86 core duo? I would find that strange, as I would highly expect them to use the same POWER9 CPU that the Talos II is using - those are 4-core Sforzas. It seems very unlikely to me that they would get IBM to develop a completely different CPU package and socket for this - I can't see it happening. If it is using the same socket as the Talos II then, according to the POWER9 Wikipedia page, it should be compatible with higher core-count POWER9s, when they become available.
For this reason, I think it would make more sense for them to just supply the board by itself here, not bundled with the CPU. Firstly, it would be a lower upfront cost, which would help with the crowdfunding, and secondly, some people might want to buy the board now and wait until the more powerful POWER9 chips become available later on.
>I would highly expect them to use the same POWER9 CPU that the Talos II is using
Yes. From what raptor wrote me, there is not much of a wide range of options in terms of hardware. It will be a talos 2 power9 sforza. It will be winded down such that it can work on a microatx mainboard which has lower priced electronic components on it.
Raptor did not state an intel cpu, which would have an comparable performance. Only that the power cpu would match a x86 duo core cpu on performance.
>wait until the more powerful POWER9 chips become available later on.
No, that is not doable. My understanding is, the power cpu on the raptor microatx mainboard has to be winded down about performance in order to work on a mainboard having lower priced components. Otherwise the mainboard would have to be more expensive.
> It will be winded down such that it can work on a microatx mainboard which has lower priced electronic components on it.
I would be very interested to hear more specific details about how it would be 'winded down' and what that means. I.e. is it being clocked lower, or only 2 cores are enabled?
> Raptor did not state an intel cpu, which would have an comparable performance. Only that the power cpu would match a x86 duo core cpu on performance.
As I said earlier in the thread, if this is truly the case then I think I'm going to have to decline to get involved any further myself. I don't really see the point of the project, as I believe there are already free-software-compatible mainboard options available that give that level of performance for much lower cost. I wish you the best of luck with the crowdfunding though, if you do pursue it.
> When will arm cpus be as fast as x86 notebooks and when will such arm cpus be libre software capable?
EOMA68 isn't supposed to use the same amount of power as x86; high-power applications are supposed to be dealt with by a separate EOMA standard which allows for more energy usage and heat, because these are the actual limiting factors, not what architecture is being used.
But if rather than comparing to the most recent high-end x86 CPUs, you compare to older but still usable or lower-end x86 CPUs, then yeah, tons of ARM CPUs exist that are as fast or even faster. You do remember that many Chromebooks use ARM CPUs, right? At least one of them which is fairly recent can even run a 100% libre software stack, and incidentally, the SoC used in that Chromebook is the one lkcl is looking at for the next EOMA68 card.
>most recent high-end x86 CPUs
I do not. I have a lenovo t400. It is fast enough to have several programs running at the same time. Display youtube hd vidoes. Have a dozen tabs open in firefox. It happens that the computer stalls. But not often. That would be my benchmark.
It is unlikely the current pc card will be able to manage the above tasks.
>tons of ARM CPUs exist that are as fast or even faster
I know of no one, which is libre software, for sale and performing equal to a lenovo t400 cpu.
>is the one lkcl is looking at for the next EOMA68 card.
If it is the Rockchip RK3288, I will have to see it showing a performance like a x86 duo core cpu, before I believe it.
I do not doubt we will get arm cpus, which will be able to act like a x86 notebook. We do not know when? And if it will be small enough to fit in a pc card? Because arm software is not libre software, lkcl can only get it by software error, reverse engineering or leaks. That is an uncertain strategy.
>EOMA68 isn't supposed to use the same amount of power as x86
If you put a notebook cabinet on sale, people have a just expectation at some point getting a pc card, which will make the notebook run as fast as a x86 notebook, including common gnulinux software.
> I have a lenovo t400.
That's not a low-end computer. That's mid-range at least. Low-end, e.g. budget computers in 2008 would have used Celeron processors, which at that time likely would have been single-core. They would also only have 2-4 GiB of RAM, and around a 100 GiB hard drive.
The kinds of laptops you should be looking at are the $300-$400 budget laptops (sometimes as little as $200) found at Walmart and other similar stores. Those also include Chromebooks. I already told you that the SoC being used for the next computer card (yes, the Rockchip one) is actually used in a fairly recent Chromebook, the C201 (released sometime in 2015, afaict). In fact, this Chromebook is still easy to find new, so you could easily buy one, install Debian or Parabola on it, and see how well it works.
So, yes. EOMA68 will meet the speed needs of many people. That's what matters, not how they compare to the laptop you personally use. EOMA standards will cover this market. POWER will cover the higher-end markets you are clearly a fan of. They are separate and have almost no audience in common with each other, so they are not competing. This just doubles our chance of success at manufacturing libre software friendly computers.
> Because arm software is not libre software, lkcl can only get it by software error, reverse engineering or leaks.
No, this is not true. Please do not make claims about things you know nothing about.
I am a member!
> The kinds of laptops you should be looking at are the $300-$400 budget
> laptops (sometimes as little as $200) found at Walmart and other similar
> stores. Those also include Chromebooks. I already told you that the SoC being
> used for the next computer card (yes, the Rockchip one) is actually used in a
> fairly recent Chromebook, the C201 (released sometime in 2015, afaict). In
> fact, this Chromebook is still easy to find new, so you could easily buy one,
> install Debian or Parabola on it, and see how well it works.
Is it hard to build GuixSD for ARM? IIRC, the only thing I might have
to actually compile is Guile and Guix.
--
Caleb Herbert
OpenPGP public key: http://bluehome.net/csh/pubkey
Funny you should mention that. Apparently someone is working on a GuixSD port for the A20 computer card. :) I'm excited for that, GuixSD seems like a really cool system.
Sure, different people are going to want different things and one piece of hardware is never going to suit everybody. I agree, of course, that the Macbook is old hardware and, in 10 years' time, it isn't going to be as easy to find one on Ebay.
I think there are a lot of people that do want more computing power though, so if that is something that is currently putting them off of free computing, then the POWER9 could be huge. There will be a place for both EOMA and the POWER9.
> There will be a place for both EOMA and the POWER9.
I agree completely. I just don't think POWER9's success is going to be by a mass-production crowdfunding campaign aimed primarily at the libre software community. I think Raptor needs to also give a lot of attention to some larger niches that POWER9 is or can be strong in. I don't know whether or not they're doing this, but if they are, then they're going about this perfectly.
If I may summarize, I think EOMA68's success has to be from the bottom up, whereas Raptor's success has to be from the top down, just because of their price points. I think this is great, though, because that's twice the chance of succeeding (since due to their massive differences in price point and power they aren't really competing with each other).
>POWER9's success is going to be by a mass-production
I do not call 10000 mainboards a mass production. The goal of a crowd funding is, if there is 10000 persons, who want to spend 1000usd on a libre software mainboard and cpu, we will find them and give them the option to buy the computer. If more want it, fine.
You can have the opinion, there are no 10000 persons. You are probably right. I think there is a small chance. And if it can be tested at small costs for raptor, it should be done.
A market survey could give us the answer telling us if a crowd funding will be fruitful. But we do not have one and are not going to get one.
Raptor has emailed a list of specifications should a ppc libre software micro atx mainboard, cpu included, get crowd funded.
1 POWER9 "Sforza" CPU
4 DDR4 DIMM slots
2 PCIe slots (1 x16, 1 x8)
On board audio
On board HDMI video (2D)
2 GbE NICs
On board USB 3.0
On board SATA (AHCI)
Optional remote control via BMC (can be fully disabled via hardware /
firmware)
The list is non obligor.
If appears the hardware options are limited. The only option
getting the price down is ordering big numbers of
mainboards.
I agree, selling 10000 mainboards,cpus included, for about
1000usd each is difficult. If the computer can work for a
decade, it is an acceptable price.
I asked raptor about the software system situation. Raptor
answered, several gnulinux distributions are available.
I asked raptor about qubes.
Qubes intends to modify x86 mainboards, such that they will
not be able to compromise qubes. I consider that to be a
questionable matter.
Raptor said, if kvm is provided for ppc, a ppc computer
will be able to run qubes. A libre version
of qubes on a ppc computer could be of interest for
computer security circles.
That sounds pretty good! I would definitely take one for $1000 (including CPU). 2 PCIs seems a little bit low, but I could work with it.
I wrote Stallman an email, explaining about how fsf could endorse given crowd fundings of pieces of libre software or hardware which runs on libre software. I suggested fsf should look into endorsing the crowd funding of a lower priced raptor ppc atx mainboard.
Stallman answered, he was unable to read my english. He asked me to get another person to write the email for me. I replied no. I rewrote my email to him and emailed it. Stallman has not responded on the rewritten email I sent.
If one of you has understood, what I think should be fsf's part in a crowd funding of a raptor ppc atx mainboard and you think, you are able to explain it to Stallman in an english written email, you may send Stallman an email.
I have resend the email, Stallman did not respond on.
Ok, I'll send him an e-mail a bit later on and see if I can explain. Seems like he is pretty good at getting back to e-mails.
Good effort, tonlee! :-)
Seriously, we need to make this happen for the future of free computing. I'll be willing to volunteer some time as well for the crowdfunding effort. I don't really see where the risk is, tbh - if we don't meet the target, no money gets charged, no-ones really lost anything.
I had a reply from rms - he sent me a copy of what he said to tonlee the day before.
In summary: he said he's going to ask the FSF people to take a look and let him know if they think the project would be a benefit to free computing. He also said that the FSF would only endorse a crowdfunding campaign if it is run on a site that uses free Javascript (which is understandable).
He said there are two crowdfunding sites that are known to use free Javascript: Crowd Supply and Goteo.
I had a look at Crowd Supply and they seem pretty good. It looks like they are quite strongly geared towards free software-compatible hardware projects. According to their Wikipedia page, they have a high rate of success, partly because of their close management of projects that they run. It seems that they also have the ability to help with fulfillment (i.e. shipping and handling), which could also be a plus for this.
Perhaps I should get in touch with them and see what they think about these Raptor boards?
Stallman answered, he is not knowledgeable on the matter. He will ask people who are familiar with fsf's hardware certifying system to decide if fsf should engage in endorsing crowd fundings, which are about libre software. I assume fsf will return, when they have made a decision.
Not sure a raptor mainboard crowd funding should be in the domain of a well known crowd funding website. What do they cost? Costs if the crowd funding target fails? I would rather set up a crowd funding website on for instance raptor's website, if they agree.
Raptor would have expenses for technical and price calculations, which raptor would have to bear. In case the funding target is not achieved, is there a fee for returning already received payments? If so who pays?
- Vous devez vous identifier ou créer un compte pour écrire des commentaires