AMA Raspberry Pi AMA - Ask your questions now!

[ebenupton]

@ebenupton Thanks a lot for taking the time for this AMA! Overal I'm very impressed with how much improvement was crammed into a $35 price point, and happily surprised it launched this early!

Thank you - it's been a substantial effort for the team. We're very glad that it came ready earlier than we were expecting.

I would love to know some more behind the design and decision proces, and why which compromises where made.

1. Was a FinFET process considered and if so what killed it for this iteration (development cost, marginal costs, available IP, etc.)?
2. This is the first time difference memory tiers of the same Pi are offered, what made you guys decide to go for multiple market levels?
3. Double 4K monitor support is amazing. What made you decide on the two Micro HDMI connectors instead of one DisplayPort over USB Type-C for example? Could two regular HDMI sockets be stacked on top of each other?
4. How much is the Pi ecosystem locked in on Broadcom? Will a Amlogic or Realtek chip in the future be possible for example?
5. The SD-card speed increase is very welcome! Was an eMMC or M.2 slot considered?
6. I would be very interested in a 'server' variant without the whole display, audio and wireless parts, leaving only the Ethernet and USB. Do you think such an product will be viable in the future?
7. The current power usage is a lot higher then with previous generations, probably mainly because the way faster out of order CPU cores. Did you guys experiment with heat sinks and metal casings (used as heat sink)?

1. A little of all of those, but primarily projected unit costs.
2. Three things: we wanted to offer a $35 product; we could only afford 1GB of RAM at that price; and people had asked for (and our chip could address) up to 4GB.
3. USB-C is rather a poor connector for high speed signals (Google "short to VBUS" for details), and we would have either needed to include protection silicon or put up with a substantial in-field failure rate. We did look at stacked type-A, but it would have dominated the board z height and we were a bit squeamish about the torque on the board when inserting the upper connector, and the EMC implications of routing high-speed signals to it.
4. I think we're happy that the Broadcom silicon offers the right feature set, and of course total backward compatibility with earlier models.
5. We don't think there's a compelling advantage to socketed eMMC over SD. M.2 would have been fun, but we didn't like the form factor considerations, and had no spare PCIe lanes. I think USB 3.0 SSDs are the way to go for high-performance storage.
6. Possibly. The savings aren't particularly compelling (with the possible exception of wireless), so the rationale would have to be about form factor rather than cost.
7. The intention is that "regular" users will be okay without an additional thermal solution, but I think that FLIRC-type metal cases will be useful for power users.

 

[GetSmart]

How come you use ARM processors and not an x86 processor like Intel? Why not partner with AMD? Ryzen on Raspberry!

Actually already exists single board computers with x86 SoCs such as LattePanda, LattePanda Alpha, Up Squared, Udoo x86, Udoo Bolt, Odroid H2, Atomic Pi, etc. However the prices for these are all much higher also (except for that Atomic Pi).

 

[ebenupton]

Could you talk more on this? What is the perfomance increase? I thought x86 was industry standard because it was the most powerful. I am a gamer so Intel and AMD is all i know 😛

So, you can get higher total performance designs from Intel and AMD, but ARM designs win out if you normalise for performance per Watt (useful if you're a thermally constrained platform like us) or per square millimetre (and thus per dollar - useful if you're a cost constrained platform).

The argument is that there's less instruction set cruft to worry about in ARM land, so ARM cores can be leaner and meaner. Not sure I totally buy that (AArch32 is plenty crufty by now), but nobody's ever really fielded a decent low-power x86 implementation, so maybe there's something to it.

 

[bit_user]

Dr. Upton,

Speaking as a long-time & vocal Pi user, I would first like to thank you for pioneering and establishing this class of computing devices. I'd also like to thank you for all the wonderful improvements in the Pi 4. It's better than I dared hope, for the price. As I reflect on that achievement, I'm duly impressed.

As for my question, I wonder whether we can anticipate OpenCL support, for the VideoCore VI GPU.

I would like to close with a suggestion that, in future generations, the GPU hardware receive more focus, in order to balance out against the rapid series of improvements that have gone into the CPU cores. GPUs offer the most power-efficient form of programmable hardware for tackling a wide range of problems, not to mention playing an essential role in the user experience with 4k displays.

Also, I believe getting more capable GPUs into the hands of so many inspired and creative makers & developers would spur innovations in GPU programming tools, techniques, and uses.

 

[nmstoker]

Great work with the Pi 4!

With all the GPU power present on the Pi 4, do you expect to support ways to use it for non-graphical use (eg machine learning)?

This has been possible before but it feels like it didn't really take off. What barriers do you see with regard to this?

 

[GetSmart]

The argument is that there's less instruction set cruft to worry about in ARM land, so ARM cores can be leaner and meaner. Not sure I totally buy that (AArch32 is plenty crufty by now), but nobody's ever really fielded a decent low-power x86 implementation, so maybe there's something to it.

Perhaps you could look into Intel's upcoming LakeField, which will be the replacement for their low power Intel Atom series. Already some benchmark leaks have appeared. And perhaps Raspberry could consider creating a new future contender in the x86 single board computer space.

 

[Balance-]

@ebenupton Thanks for the extensive answers on my previous questions! I have one more:

Is an 64-bit version of Raspbian in development? Some applications benefit a lot from ARMv8 / 64-bit (NEON) optimizations.

 

[kasid]

Some people just want to use Raspberry Pis as just a lightweight PC for media consumption and web browsing. How important is making this experience go smoothly to the development team? And are they interested in enlarging the user base by attracting also the people who just want a simple, cheap PC and are not into using it for projects?

 

[flipontheradio]

I'm a huge fan of the official Raspberry Pi 7" touchscreen, are there any plans to release a larger (10"?) touchscreen down the road? Any possibility of an updated 7" touchscreen that will position the Pi in reverse (arranged to power does not plug in on top), and finally, will there be an updated case for the 7" touchscreen for the Pi 4?

 

[will9183]

@ebenupton Is there going to be support for HDMI input on the Raspberry Pi either through the second micro-HDMI connector or through a HAT in the future?

 

[ebenupton]

Dr. Upton,

Speaking as a long-time & vocal Pi user, I would first like to thank you for pioneering and establishing this class of computing devices. I'd also like to thank you for all the wonderful improvements in the Pi 4. It's better than I dared hope, for the price. As I reflect on that achievement, I'm duly impressed.

Thank you for the kind words. You can see from the blog post just how many people have contributed to the program: it's been the most remarkable thing to be involved with.

As for my question, I wonder whether we can anticipate OpenCL support, for the VideoCore VI GPU.

The hardware is certainly capable of it, and I am aware of at least one proprietary OpenCL implementation for VideoCore V. It may be that OpenGL ES Compute Shaders (in ES 3.1) are both a better abstraction and quicker to implement, so we might get those first.

I would like to close with a suggestion that, in future generations, the GPU hardware receive more focus, in order to balance out against the rapid series of improvements that have gone into the CPU cores. GPUs offer the most power-efficient form of programmable hardware for tackling a wide range of problems, not to mention playing an essential role in the user experience with 4k displays.

FWIW I think gross GFLOP numbers, and the OpenArena benchmarks that have appeared, understate the uplift we've just delivered (largely because VideoCore IV is extremely good at Quake III-type rendering for its age). On modern, compute-heavy benchmarks I think we'll see >4x increase.

But yes, more work to be done here.

Also, I believe getting more capable GPUs into the hands of so many inspired and creative makers & developers would spur innovations in GPU programming tools, techniques, and uses.

Absolutely agree. I remember chatting with Sophie Wilson (co-creator with Steve Furber of the ARM1) a few years ago. She was concerned that with Raspberry Pi 1 we were teaching children an obsolete view of how computers actually work (single core, homogenous compute). Bringing GPU compute into the mix doesn't just enable cool projects: it helps address that concern too.

 

[ebenupton]

Great work with the Pi 4!

Thank you!

With all the GPU power present on the Pi 4, do you expect to support ways to use it for non-graphical use (eg machine learning)?

This has been possible before but it feels like it didn't really take off. What barriers do you see with regard to this?

I'd certainly like to see it put to use for (e.g.) machine learning acceleration. Historical barriers:

• Lack of tooling (no OpenCL etc). This meant you had to hand-roll everything in QPU assembler. I did a couple of nice kernels, and Pete Warden (now at Google) did some ML, but it's tough going.
• Lack of good abstractions. The ratio of GPU FLOPS to CPU FLOPS on devices like 2711 is much lower than on a PC with a modern graphics card, so you need abstractions that don't come with a lot of overhead; otherwise you end up in a situation where you're better off just writing ARM code. OpenCL is, I think, quite a bad abstraction.

Hopefully once we get OpenGL ES Compute Shaders we might start to see some progress in this area.

 

[ebenupton]

@ebenupton Thanks for the extensive answers on my previous questions! I have one more:

Is an 64-bit version of Raspbian in development? Some applications benefit a lot from ARMv8 / 64-bit (NEON) optimizations.

So 64-bit Raspbian would actually just be AArch64 Debian. It is probably feasible to build a work-alike version of our standard release against the AArch64 repos (we already do this for PCs and Macs with the i386 repos), but we're currently unpersuaded that there's a benefit.

Much more likely is that we move in due course to a 64-bit kernel with the 32-bit Raspbian userland on top of it. There's work to be done, but this would bring some performance improvements in some areas by reducing page-table maintenance cost versus our current 32-bit LPAE kernel.

 

[g-unit1111]

Since the 3rd generation could only support a max of a 32GB micro SD card, will the 4th generation have any limitations on the maximum amount of SD card storage a RP can hold?

 

[ebenupton]

Some people just want to use Raspberry Pis as just a lightweight PC for media consumption and web browsing. How important is making this experience go smoothly to the development team? And are they interested in enlarging the user base by attracting also the people who just want a simple, cheap PC and are not into using it for projects?

This is incredibly important to us: in fact it's the primary goal with Raspberry Pi 4. Getting a cheap, usable PC, pre-loaded with programming tools into everyone's (and particularly children's) hands is one of the ways we want to accomplish our charitable mission.

 

[ebenupton]

I'm a huge fan of the official Raspberry Pi 7" touchscreen, are there any plans to release a larger (10"?) touchscreen down the road? Any possibility of an updated 7" touchscreen that will position the Pi in reverse (arranged to power does not plug in on top), and finally, will there be an updated case for the 7" touchscreen for the Pi 4?

An updated touchscreen is drifting towards the top of our todo list. Definitely hear you on plug positioning.

We don't do a touchscreen case at the moment, but I'm aware of at least one third party who is investigating a Pi 4 + touchscreen case.

 

[ebenupton]

@ebenupton Is there going to be support for HDMI input on the Raspberry Pi either through the second micro-HDMI connector or through a HAT in the future?

If we did this, it would be via a HAT. In fact (I think this is public knowledge) we did prototype one a long time ago. Also in fact (and this isn't public knowledge) we use them as part of the test jig for Raspberry Pi 4 production, to verify that HDMI output is happening correctly.

The big challenge is that you struggle to do more than 720p over the two-lane CSI connector. Maybe we should revisit this once we've determined if the 28nm MIPI IP can reliably support 1.5Gbps operation.

 

[ebenupton]

Since the 3rd generation could only support a max of a 32GB micro SD card, will the 4th generation have any limitations on the maximum amount of SD card storage a RP can hold?

I'm not sure I recognise that limit. IIRC this was related to cards >32GB being pre-formatted with EXFAT, which we can't read, so if you took a blank card and copied the NOOBS files on it would fail to boot. You could always overwrite the card with a Raspbian image (with its small FAT32 boot partition) and get it to boot.

In any case, there won't be a limit on Raspberry Pi 4. If we find incompatibilities in the field we'll ship an updated SPI Flash image to fix them.

 

[james.Harking]

Hello Eben,

I have been a huge fan of the Foundation and the Raspberry Pi for many years now, indeed I was involved with support and promotion of the first version of Raspbian with Mike Thompson when working on the initial hard float community release when you were still shipping Fedora.

Anyway I wanted to say that the Pi 4 looks like a very nice upgrade, however I wanted to ask about the teared pricing for the RAM upgrades. The standard SKU is $35, however the additional 3GB version costs $55, 57% extra over the base cost, that appears quite an expensive upgrade.

I understand that the foundation needs to make money but this is still quite a large % increase in price. If you decided to keep that $55 price could a small amount of EMMC storage be included in the price, perhaps 8 or 16GB? This would certainly help with performance. In my experience MicroSD cards become corrupt within months of day to day usage, particulary if you are using the Pi as a low end desktop.

Is there any plans to decrease these costs with economies of scale over the coming months or will this be fixed now until any future hardware announcement, e.g. 4a?

In addition could I ask if there are any future plans for different form factors? I understand you probably couldn't confirm here, but it is something to think about.

As a teacher I would really like to see the Raspberry Pi 4 (or a future model) integrated directly in to a monitor with the ability to access the GPIO pins, perhaps with a flap of some sort. As you are now targeting 'desktop computing' I think something like this would be very valuable.

You could target class rooms and computer labs with a standard supported unit. If you could hit the magic sum of $100 with keyboard and mouse then I think you would be on to a real winner. There are many poorer countries (even poorer school districts) that would jump at the chance to purchase an item like this and it would suit the educational goals of the foundation while promoting free software and open learning without restriction.

Thanks and kind regards,

Jim

 

[computerjoe314]

Why did you decide to add 2 display outputs?
Do the tarrifs on China effect raspberry pi pricing?
Why was it allowed for a more expensive Raspberry Pi this time, but not any other times?

 

[bit_user]

OpenCL is, I think, quite a bad abstraction.

For machine learning, or in general?

I see Broadcom is listed as a contributor member of Khronos, but I don't see the Raspberry Pi Foundation.

The Khronos Group

Deploying and developing royalty-free open standards for 3D graphics, Virtual and Augmented Reality, Parallel Computing, Neural Networks, and Vision Processing

www.khronos.org

As Khronos plots the the future evolution of OpenCL (sometimes referred to as "OpenCL-Next"), they seem fairly receptive to input. If you have any suggestions or ideas, perhaps it might be worth reaching out to them.

Hopefully once we get OpenGL ES Compute Shaders we might start to see some progress in this area.

I'm all for that, but what I've read about OpenGL Compute Shaders gave me the distinct impression they're essentially a subset of OpenCL.

I don't mean to seem argumentative - just trying to understand your concerns about OpenCL. Thanks.

 

[bliq]

Aside from the extra RAM, does the new Pi4 have better virtualization support? I tried using a Pi3 as an openstack compute node and it turns out that extensive modification would have been necessary to get it to work.

 

[shardros]

Hello Eben,

It looks like you and the team have done a great job with the Raspberry Pi 4. (Big fan of USB 3.0 and videocore VI).

I was just wondering what regrets or what you would do differently if you were where you were back in 2011?

Thanks again

shardros

 

[fdlouw]

Good day,

Thanks for bringing the world such a wonderful product. I wish you lots of success for the future.

Relating to the earlier question about tarrifs, what were some of the challenges you encountered in order to keep the price at $35? I recall somebody (might have been yourself) saying GbE and USB 3 being a difficult thing to implement if you wanted to keep the price down.

Do you foresee a situation were the industry's price/performance ratio stagnates somewhat and you're forced to increase the price in order to release better hardware?

Finally, do you have any thoughts on RISCV and if it can potentially be used on your products?

Thanks again.

 

[ebenupton]

I have been a huge fan of the Foundation and the Raspberry Pi for many years now, indeed I was involved with support and promotion of the first version of Raspbian with Mike Thompson when working on the initial hard float community release when you were still shipping Fedora.

Thank you for your efforts there. I think the move to Raspbian is one of the critical things that made Raspberry Pi a success early on.

Anyway I wanted to say that the Pi 4 looks like a very nice upgrade, however I wanted to ask about the teared pricing for the RAM upgrades. The standard SKU is $35, however the additional 3GB version costs $55, 57% extra over the base cost, that appears quite an expensive upgrade.

I understand that the foundation needs to make money but this is still quite a large % increase in price. If you decided to keep that $55 price could a small amount of EMMC storage be included in the price, perhaps 8 or 16GB? This would certainly help with performance. In my experience MicroSD cards become corrupt within months of day to day usage, particulary if you are using the Pi as a low end desktop.

Is there any plans to decrease these costs with economies of scale over the coming months or will this be fixed now until any future hardware announcement, e.g. 4a?

A couple of thoughts here:

• The tiered pricing structure largely reflects the underlying cost of RAM. If you look at a typical brand-name SODIMM you'll see an end-user price of roughly $5/GB, which is close to how we price the upgrade.
• We like the "stateless" nature of the device, which means it's very hard to brick and easy to update. I'm surprised you're seeing issues with SD card: it's been a long time since I've been able to corrupt a brand-name card (though I do tend to power down cleanly).
• We don't expect much further downward movement in RAM prices.

In addition could I ask if there are any future plans for different form factors? I understand you probably couldn't confirm here, but it is something to think about.

No plans at present, though of course we're interested in doing a CM4.

As a teacher I would really like to see the Raspberry Pi 4 (or a future model) integrated directly in to a monitor with the ability to access the GPIO pins, perhaps with a flap of some sort. As you are now targeting 'desktop computing' I think something like this would be very valuable.

You could target class rooms and computer labs with a standard supported unit. If you could hit the magic sum of $100 with keyboard and mouse then I think you would be on to a real winner. There are many poorer countries (even poorer school districts) that would jump at the chance to purchase an item like this and it would suit the educational goals of the foundation while promoting free software and open learning without restriction.

That's a very interesting thought: thanks for sharing.