An other CPU discussion. Rockchip RK3399 sounds good


The RK3399 dev board I posted earlier looks like a perfect board to test the RK3399 with


zzuIH8F.png
 
How about just going for this one then? Seems to have almost the same

http://www.orangepi.org/Orange Pi RK3399/

RK3399_en.jpg

https://www.aliexpress.com/store/group/OPI-RK3399-2GB/1553371_512711117.html
Currently $109... which seems pretty cheap to me

Even better - and thank you for using spoiler tags!
[doublepost=1529421514,1529421138][/doublepost]If the RK3399 is really going to be considered - board availability seems high.

I think this is the same one @vcoleiro1 posted above.
@WizardStan - is this what you were looking for?

Four different versions with 2GB/4GB RAM and camera/nocamera, so they have those bits of contention covered.

Nice little case & setup. Looks power-on ready.

I still like the additional connectors on the board that @pmprog found better, but have no basis for preference beyond that.
[doublepost=1529422733][/doublepost]Is there a good way to compare the graphics performance of the RK3399 with the OMAP5432 and others?

The Pyra is multi-use, yes, but it has an obvious focus on gaming. Having accessible 3D drivers and high graphics performance should probably be a requirement - and may be more important than having "the fastest" CPU.
 
Yeah, that orangepi version seems to lack to gain an MiniPICe socket but loses a 3G/LTE socket (which looks remarkably like a MiniPCIe socket to me), lacks a dedicated buzzer and has the radiator mounting holes in a different orientation (although the chip still looks to be in the middle and they may be the same distance apart so it may not matter).

Edit: That first sentence fragment should read:
Yeah, that orangepi version seems to gain a MiniPCIe socket but loses a 3G/LTE socket (which looks remarkably like a MiniPCIe socket to me).

That's what happens sometimes when you decide to reword a sentence while still writing it but don't actually check what you've written.
 
Last edited:
that orangepi version seems to lack to gain an MiniPICe socket but loses a 3G/LTE socket (which looks remarkably like a MiniPCIe socket to me)
I struggled to parse that statement, but yes, it has a miniPCIe slot which can be used as with a modem (and there's a sim card slot on the board)
 
Which board? The original linked board has a PCIe socket labelled 3G/LTE modem, while the OrangePI version has a socket just labelled as a MiniPCIe socket. The form factor looks to me like it's designed to take pre-M2 memory cards, but I guess other peripherals are available in that connector, so perhaps the hardware is the same even though the labelled uses vary.
 
The form factor looks to me like it's designed to take pre-M2 memory cards, but I guess other peripherals are available in that connector, so perhaps the hardware is the same even though the labelled uses vary.
On the one hand mini PCIe actually has a different physical layout than M.2 B and M key (which are also incompatible to each other unless the card supports both!), on the other hand the original mini PCIe standard already defined multiple interfaces (PCIe and USB 2.0) of which none is mandatory. For example, the 2nd mini PCIe slot available in T60 Thinkpads that is supposed to be used for WWAN cards is USB only.

The newer M.2 standard is even more versatile, it may not only include an optional SATA interface but USB 3.0 as well. There are even more interfaces available, the selection is being limited by the M.2 key type.
 
Even better - and thank you for using spoiler tags!
[doublepost=1529421514,1529421138][/doublepost]If the RK3399 is really going to be considered - board availability seems high.

I think this is the same one @vcoleiro1 posted above.
@WizardStan - is this what you were looking for?

Four different versions with 2GB/4GB RAM and camera/nocamera, so they have those bits of contention covered.

Nice little case & setup. Looks power-on ready.

I still like the additional connectors on the board that @pmprog found better, but have no basis for preference beyond that.
[doublepost=1529422733][/doublepost]Is there a good way to compare the graphics performance of the RK3399 with the OMAP5432 and others?

The Pyra is multi-use, yes, but it has an obvious focus on gaming. Having accessible 3D drivers and high graphics performance should probably be a requirement - and may be more important than having "the fastest" CPU.

Totally disagree with this. We are not going to run modern games because there aren't such in linux arm. On the other hand emulation and productivity will make use of CPU over GPU all the time.

I would love to hear other's opinion. I really think that CPU and single core performance should be the priority.
 
We are not going to run modern games
So what is your definition of modern? I mean my definition may be a little skewed as I've been using an FPGA accelerated Amiga 2000 as my primary home workstation for the last couple months.
 
There are a few open source games that can tax a more advanced GPU than the Pandora has at least; games like OOlite and various 3D engines that run commerical game wads. But yeah, other than those exceptions, most things we'll be running on our Pyras are going to be CPU bound, not GPU.
 
There are a few open source games that can tax a more advanced GPU than the Pandora has at least; games like OOlite and various 3D engines that run commerical game wads. But yeah, other than those exceptions, most things we'll be running on our Pyras are going to be CPU bound, not GPU.

Things are CPU bound until they aren't. At some point the CPU ceases to be the bottleneck.

Example: If your use case is surfing the internet, any CPU of Core 2 Duo or later vintage will do just fine. Processors prior to that cross an invisible boundary below which things start getting bad in a big hurry.

If we're going to seriously discuss a successor Pyra SoC, we're entering an age where the CPU side of the SoC dies are, if anything, way overpowered (mostly in core count). On the Pyra we will generally need 1 relatively fast core for the foreground application and maybe half of a second core to take care of background tasks. Modern ARM SoCs are tending towards 6-10 cores of varying speeds. Unless you can point to truly multi-threaded applications that will run...?

So, that one primary main CPU core will likely have some sort of cap on it. I realize that it isn't this simple, but people understand Ghz even if they don't. So, if two SoCs both have more or less equivalent magnitude primary CPU cores (1.7 to 2.5 Ghz) but one has a substantially "better" GPU (speed, driver accessibility, etc.), the exact speed of the CPU portion may then become be less relevant.

An industrial ARM SoC might have a pile of blazing fast cores (and the thermal footprint to match) but near enough to nothing in the GPU department.

A mobile ARM SoC will likely be more balanced. It gets specs credentials for the CPU core count and Ghz of it's fastest core, but needs the GPU to be good enough to feel speedy. 99% of the time the 3rd through Nth CPU core will be idle.

A modern automotive SoC will have several CPU cores, but much more emphasis on the GPU. It will be expected to drive fluid graphics on upwards 6+ displays simultaneously.

Example using R-car H3:
[doublepost=1529438772,1529437199][/doublepost]As noted above, processor efficiency (battery & thermal impacts) are generally better the smaller the lithographic process is.

The OMAP5432 is built on a 28nm process
The RK3399 is built on a 28nm process
The R-car H3 is built on a 16nm process

However, the A72 in the RK3399, if it were made on the same process, would be faster core to core at the same Ghz than the A53 that the R-car H3 uses as it's primary core. But, the above lithography difference in efficiency might well make up for that - I don't know.

Which brings me to the question of what is out there with a <= 16nm process and an A72 or better core?
Can we get a Helio X30?
10nm process...
mobile so balanced CPU/GPU
graphics?

Planet Computers was able to get the X27 for their Gemini - and it is running both Android and Debian Linux, so there has to be some taste for niche products and Linux.
 
A modern automotive SoC will have several CPU cores, but much more emphasis on the GPU. It will be expected to drive fluid graphics on upwards 6+ displays simultaneously.
There's a wide range of automotive SoCs for a lot of different purposes, including whole independent CPU architectures like Infineon's Tricore and Renesas' RH850 series - a lot of them are still single core. PowerPC Book E is also still alive and kicking in this area. Many don't have any sort of public information available, though - which often means that even the documentation that is not public is not necessarily complete or at least somewhat free from flaws... (this includes the R-Car chips)

Please spare me with the 3rd gen R-Car series, I already had the... "privilege" to work with them. Some parts of their basic CPU design are seriously flawed.
But by far the worst are the fans on the V3M Eagle devkits - never ever have I seen such a crappy fan ever in my life, these things will drive you insane. They literally sound like one of the bearing balls might break loose at any time, flying across the room just to penetrate your skull at full speed - and that chip only has 2 A53 + 2 R7 cores.
 
There's a wide range of automotive SoCs for a lot of different purposes, including whole independent CPU architectures like Infineon's Tricore and Renesas' RH850 series - a lot of them are still single core. PowerPC Book E is also still alive and kicking in this area. Many don't have any sort of public information available, though - which often means that even the documentation that is not public is not necessarily complete or at least somewhat free from flaws... (this includes the R-Car chips)

Please spare me with the 3rd gen R-Car series, I already had the... "privilege" to work with them. Some parts of their basic CPU design are seriously flawed.
But by far the worst are the fans on the V3M Eagle devkits - never ever have I seen such a crappy fan ever in my life, these things will drive you insane. They literally sound like one of the bearing balls might break loose at any time, flying across the room just to penetrate your skull at full speed - and that chip only has 2 A53 + 2 R7 cores.

Fair enough. I don't have any direct experience with them. From what I was able to tell, though, the R-car H3 'fit' the ports, XYZ space, has open Linux documentation, readily available dev boards, etc... But, if you're going to state that they're flawed to the point off non-consideration, OK. I deffer to your direct experience. What is better?

I still think we should be looking for a future SoC in systems on a <=16nm process. The efficiency gains over the OMAP5432 and the RK3399 (both 28nm) should be substantial enough to justify it.
 
What about the embedded version of the snapdragon 820 that was announced in February? Could it be obtained for the Pyra?
 
What about the embedded version of the snapdragon 820 that was announced in February? Could it be obtained for the Pyra?
If Qualcomm is as receptive as they were when EvilDragon was shopping around for SoCs before the OMAP5 was chosen, then likely no.
 
The board I posted earlier is $164. I emailed them for a price
It is 2 years old. We should be looking 1 to 2 years forward.
It is a 28nm chip as is the OMAP 5432. Gains would be slight when compared to 14nm or 12nm or 10nm SoCs.

I wonder if there are dev boards for a <16nm A73 SoC...?
 
Back
Top