Raspberry Pi 3 B+ vs. Asus Tinkerboard

[Beorn_Bear]

Has anyone done a performance comparison with the new Raspberry Pi 3 B+ vs. Asus Tinkerboard?!

[Imperf3kt]

I expect this to be locked/removed pretty quickly - RPF dislikes anyone mentioning to ATB.

But anyway, if you want a comparison between the two, you should consider that the TB has a newer model that should be used instead.

[fruitoftheloom]

I expect this to be locked/removed pretty quickly - RPF dislikes anyone mentioning to ATB.

But anyway, if you want a comparison between the two, you should consider that the TB has a newer model that should be used instead.

The Tinkerboard and Tinkerboard S both use the same ARMv7 Rockchip RK3288 SoC.

[Imperf3kt]

I expect this to be locked/removed pretty quickly - RPF dislikes anyone mentioning to ATB.

But anyway, if you want a comparison between the two, you should consider that the TB has a newer model that should be used instead.

The Tinkerboard and Tinkerboard S both use the same ARMv7 Rockchip RK3288 SoC.

Yes, but as I understand it, there were numerous other improvements similar to what happened between a Pi3B and a Pi3b+ that also affect how it performs.

If anyone wants, I can provide the tinkerboard side of the comparison if you provide the benchmarks you wish to test and will compare against your own PI3B+, or wait until I buy a Pi3B+, which may take a while as I cannot afford one just yet.

Better act fast because within a week or so I plan to attempt something that may kill the Tinkerboard and I don't plan on buying another.

[fruitoftheloom]

I expect this to be locked/removed pretty quickly - RPF dislikes anyone mentioning to ATB.

But anyway, if you want a comparison between the two, you should consider that the TB has a newer model that should be used instead.

The Tinkerboard and Tinkerboard S both use the same ARMv7 Rockchip RK3288 SoC.

Yes, but as I understand it, there were numerous other improvements similar to what happened between a Pi3B and a Pi3b+ that also affect how it performs.

If anyone wants, I can provide the tinkerboard side of the comparison if you provide the benchmarks you wish to test and will compare against your own PI3B+, or wait until I buy a Pi3B+, which may take a while as I cannot afford one just yet.

Better act fast because within a week or so I plan to attempt something that may kill the Tinkerboard and I don't plan on buying another.

The basics are the same, anyway have we not given Asus enough free advertising ?

viewtopic.php?t=193038

https://liliputing.com/2018/01/80-asus- ... puter.html

Basically they have taken a SoC used in their Chrome OS products and offered it as a Pi Clone !

Comparison, the only similarity is it runs a fork of Debian ARMHF..........

[ejolson]

I expect this to be locked/removed pretty quickly - RPF dislikes anyone mentioning to ATB.

This is not a thread asking for help with a competing product, but a question about how Raspberry Pi compares. Since many people on this forum are Raspberry Pi experts, comparisons made here should be quite accurate. Since Raspberry Pi is orders of magnitude more popular for a reason, accurate comparisons should be beneficial.

From my point of view, it would be interesting to test both machines using the HPL high-performance Linpack benchmark. More details how to install this program on Raspbian is available here.

[jahboater]

The Tinkerboard and Tinkerboard S both use the same ARMv7 Rockchip RK3288 SoC.

The Pi3+ uses the ARMv8 Cortex-A53 processor.

[fruitoftheloom]

The Tinkerboard and Tinkerboard S both use the same ARMv7 Rockchip RK3288 SoC.

The Pi3+ uses the ARMv8 Cortex-A53 processor.

The RPi SoC is a VPU with a ARMv8 CPU

The TK is an ARMv7 CPU with a GPU..


FIP my ChromeBit with the RK3288 is speedier than any RPi I have owned, but then I have never run FlintOS on my RPis which is the nearest OS to ChromeOS..

[bensimmo]

Full specs here tinkerboard RaspberryPi
and a reference to the original being twice as fast as the Pi3 on the TinkerBoard
3+ is a 15% notch up on the 3 iirc.
So still slower I would asume.

of course you programmes and usage may vary.
The questions are, does it do KODI (265 for our german members) and RetroPi better.
Can I use Android, Android Thing, Windows 10 IOT and it the general OS faster.
Does it have an easily accessible, to new users, standard desktop with most things they need.
Can they claim Hawkings used it for his voice.

[jamesh]

Happy to leave this thread up and running, but please keep to the title - performance comparisons.

I expect the TB to be faster in some areas, not so good in others. We never claim the Pi is the fastest SBC, but we do like to think it is the best designed, and the best supported. It's certainly the best seller.

[Imperf3kt]

I forgot I actually ran a Linpack benchmark about 3 weeks ago, but I ran a new test tonight on the original Tinkerboard.
Note: The following test was ran on a stock Tinkerboard, running the outdated TinkerOS 1.8

Written by A. Petitet and R. Clint Whaley, Innovative Computing Laboratory, UTK
Modified by Piotr Luszczek, Innovative Computing Laboratory, UTK
Modified by Julien Langou, University of Colorado Denver
================================================================================

An explanation of the input/output parameters follows:
T/V : Wall time / encoded variant.
N : The order of the coefficient matrix A.
NB : The partitioning blocking factor.
P : The number of process rows.
Q : The number of process columns.
Time : Time in seconds to solve the linear system.
Gflops : Rate of execution for solving the linear system.

The following parameter values will be used:

N : 8000
NB : 256
PMAP : Row-major process mapping
P : 1
Q : 1
PFACT : Left
NBMIN : 2
NDIV : 2
RFACT : Right
BCAST : 2ring
DEPTH : 0
SWAP : Mix (threshold = 64)
L1 : transposed form
U : transposed form
EQUIL : yes
ALIGN : 8 double precision words

--------------------------------------------------------------------------------

- The matrix A is randomly generated for each test.
- The following scaled residual check will be computed:
||Ax-b||_oo / ( eps * ( || x ||_oo * || A ||_oo + || b ||_oo ) * N )
- The relative machine precision (eps) is taken to be 1.110223e-16
- Computational tests pass if scaled residuals are less than 16.0

================================================================================
T/V N NB P Q Time Gflops
--------------------------------------------------------------------------------
WR02R2L2 8000 256 1 1 39.86 8.565e+00
HPL_pdgesv() start time Wed Mar 21 21:17:44 2018

HPL_pdgesv() end time Wed Mar 21 21:18:24 2018

--------------------------------------------------------------------------------
||Ax-b||_oo/(eps*(||A||_oo*||x||_oo+||b||_oo)*N)= 0.0025941 ...... PASSED
================================================================================

Finished 1 tests with the following results:
1 tests completed and passed residual checks,
0 tests completed and failed residual checks,
0 tests skipped because of illegal input values.
--------------------------------------------------------------------------------

End of Tests.
================================================================================

Temperature reached approximately 53 degrees celcius during the test.

According to the results, the test took 39.86 seconds and achieved 8.565 GigaFlops


About what I was saying earlier about possibly destroying this Tinkerboard; I plan to liquid cool it and overclock the crap out of it. Though, if I'm being honest, it already performs several times better than my laptop. Maybe I should overclock my laptop instead.
As soon as I can mill all the parts I need, I'll be proceeding, so if anyone has any further testing they'd like to see I would advise being somewhat quick about it.

[W. H. Heydt]

Happy to leave this thread up and running, but please keep to the title - performance comparisons.

I expect the TB to be faster in some areas, not so good in others. We never claim the Pi is the fastest SBC, but we do like to think it is the best designed, and the best supported. It's certainly the best seller.

I don't think there is any question about "best supported". The last comparison data I saw was that the Tinkerboard had buggy software and little to no support. The fastest board in the world is useless without properly functioning software.

[fruitoftheloom]

Happy to leave this thread up and running, but please keep to the title - performance comparisons.

I expect the TB to be faster in some areas, not so good in others. We never claim the Pi is the fastest SBC, but we do like to think it is the best designed, and the best supported. It's certainly the best seller.

I don't think there is any question about "best supported". The last comparison data I saw was that the Tinkerboard had buggy software and little to no support. The fastest board in the world is useless without properly functioning software.

TinkerOS (Debian) v2.0.5 released 20/03/18:

https://www.asus.com/Single-Board-Compu ... _Download/

[ejolson]

According to the results, the test took 39.86 seconds and achieved 8.565 GigaFlops

Thanks for running the test. A speed of 8.565 Gflops seems reasonable. Is any over clocking involved? Since you mention over clocking later in your post, I thought I'd ask to make sure. As an aside, If I'd recommend overcooking your laptop instead. Laptop CPUs are often thermally constrained. Therefore, if you add water cooling, the speed of your laptop may improve considerably.

I'm currently traveling and will get a Raspberry Pi 3B+ when I return. I'm expecting about 7.6 Gflops in 32-bit compatibility mode. My conjecture is that the Linpack benchmark on the 3B+ will be slightly slower in 32-bit mode and slightly faster in full 64-bit mode. However, the only way to know for sure is to test the actual hardware when it arrives.

[Imperf3kt]

According to the results, the test took 39.86 seconds and achieved 8.565 GigaFlops

Thanks for running the test. A speed of 8.565 Gflops seems reasonable. Is any over clocking involved? Since you mention over clocking later in your post, I thought I'd ask to make sure.

There is currently no overclocking involved. Everything on the board is stock at the moment except for the larger heatsink I put on it before putting the poor thing in a tiny acrylic enclosure.
The OS is outdated however, so improvements (or the opposite) may be seen when running the latest release. I will test again soon, using the latest updates.


Of course, numbers aren't everything. As a general use computer, I'd say the ATB beats any SBC out there, but most people here don't want a general use computer, at least... those of us who aren't using them solely to play Retropie or as a Kodi media streamer.

When it comes to using the GPIO, despite being a copy of the Raspberry Pi (right down to leaving copyright notices in 'their' documentation), the GPIO is incredibly difficult to use on a Tinkerboard. It simply doesn't work.

[ejolson]

When it comes to using the GPIO, despite being a copy of the Raspberry Pi (right down to leaving copyright notices in 'their' documentation), the GPIO is incredibly difficult to use on a Tinkerboard. It simply doesn't work.

Three advantages that differentiate the Raspberry Pi from buying an x86 PC for $35 at used surplus or the local thrift shop is small size, manufacturer support for Linux and working GPIO. If you lose two of those, then only small size is left. At the same time, small size can be pretty compelling.

[HawaiianPi]

Happy to leave this thread up and running, but please keep to the title - performance comparisons.

I expect the TB to be faster in some areas, not so good in others. We never claim the Pi is the fastest SBC, but we do like to think it is the best designed, and the best supported. It's certainly the best seller.

This is really the key. Lots of "Raspberry Pi killer" single board computers have come to market, and none have even managed to wound the feisty little Pi, based on the astronomical sales figures (which I'm sure eclipse all other brands combined).

The Tinker Board has nice hardware (in synthetic benchmarks the TB is much faster), but there is literally no support from ASUS, and the OS is a steaming pile of poo compared to Raspbian.

[Imperf3kt]

the OS is a steaming pile of poo compared to Raspbian.

Funny you should say that because it uses the same kernel version of Debian Stretch with LXDE as Raspbian does.

[fruitoftheloom]

the OS is a steaming pile of poo compared to Raspbian.

Funny you should say that because it uses the same kernel version of Debian Stretch with LXDE as Raspbian does.

As posted above they released a new TinkerOS revision this week....

[jahboater]

the OS is a steaming pile of poo compared to Raspbian.

Funny you should say that because it uses the same kernel version of Debian Stretch with LXDE as Raspbian does.

There is much more to a Linux distro than the kernel.

[jamesh]

the OS is a steaming pile of poo compared to Raspbian.

Funny you should say that because it uses the same kernel version of Debian Stretch with LXDE as Raspbian does.

Which is mostly irrelevant, as what's in the kernel depends entirely on the build settings which are I suspect entirely different for the TB since its is an entirely different collection of HW.

[Imperf3kt]

the OS is a steaming pile of poo compared to Raspbian.

Funny you should say that because it uses the same kernel version of Debian Stretch with LXDE as Raspbian does.

Which is mostly irrelevant, as what's in the kernel depends entirely on the build settings which are I suspect entirely different for the TB since its is an entirely different collection of HW.

Fair enough, thats something I wasn't really aware of.
For comparison's sake, my personal experience with the TinkerOS 1.8 is that it works somewhat better than Raspbian ever has for me. At least, it did until I turned off the beta VC4 driver in Raspbian.
My one major gripe with it was that selecting what device to output audio over, is horrible. You set it per boot, per program. Start Chrome? You need to tell Chrome to use your audio devie of choice. Start Kodi after? Now you need to tell kodi to use the same audio output.
Restart the TB? You need to do it all over again.
I'm hoping that is fixed in the later OS versions.

I'll be installing the latest OS release hopefully tomorrow, so I'll post some new Linpack benchmarks, though I don't expect them to be much different.
If anybody can share some proper Pi3B+ Linpack benchmarks, that'd be great.

While I'm at it though, are there any specific things you guys were really looking at, or is this just another of those "which is faster" threads that don't really have any meaning in the real world?

[MaxK1]

Over the years I have acquired over a dozen pi's and a single TB as well as a BBB (The BBB was also touted as a "Pi killer" - LMAO). While the TB might perform a bit better, the support just isn't there and believe me it needs it. They better hope it doesn't become a "Pi Killer" as they seem to depend on the pi community for a lot of the expertise and add on hardware. I imagine that is true for other "Pi Killer" boards out there. They totally underestimated/ignored the importance of community/support. The TB is certainly not a board I would recommend for a new user.

I expect to receive my Pi3 B+ before the week is out and if it overclocks well, it may just replace the TB. The TB will then join the BBB in the drawer that I use to collect "stuff" I don't have an immediate use for...

[bensimmo]

Some Chromium and Flash speed tests (for scratch and general usage)
I'm sure most people use the browser on them.
If it's a V's thing. Some 3D FPS ?
speed at which GPIO can be switched on off ?

[robbes]

Phoronix has a review today:
https://www.phoronix.com/scan.php?page= ... plus&num=1