Arduino Uno Tournament

[sje]

Hardware is at least 50% of the improvement from 1990 to today.

I'd say that figure is closer to 80%. Maybe 90%.

Reasonably priced consumer machines with 64 bit CPUs didn't appear until ten years ago, so anything before that had to run on 32 bit processors with a big performance hit for bitboard programs.

[DavidEather]

BBB is NOT going to happen. Your own figures show it is more than twice as expensive and for some people they will only use it once.

Official Uno's comes with a USB cable and the Uno can be powered from USB without any problems so no costs there.

100% compatible Uno clones cost:
https://www.sparkfun.com/products/12757

$20 for the made in America one above, or even less for Chinese made (I have some $12 one).

If you want to, you can make your own "Uno" at home with minimal tools. The only "must buy" item is the chip itself available with the boot-loader for $4. No one can make a Beagle Bone at home.

Arduino documentation, support and forums are much more developed: 9 years instead of 3 years for beagle bone or 1+ year for beagle bone black which IIRC was introduced in part to correct "unusual" behavior with the Beagle Bone that was stifling it's uptake. The Uno learning curve is also considerably easier.

The difference is so much that the support available for MULTIPLE people with the Uno may be enough to get it done - for the BBB, it won't be.

As has been pointed out there is essentially no challenge in using a BBB. It can run a full operating system and offers essentially unlimited memory, so it does not help the hidden agenda of this tournament:

To give "makers" a project that challenges their creative programing abilities and to give programmers an introduction to what is possible in the constrained world of microcontrollers.

[bob]

I was intentionally low-balling the estimate. I agree with you that it is MUCH higher than 50%. Certainly not 1% as that is nonsense.

[sje]

I have a BB Black. I have a Raspberry PI model B, I have TWO Arduino boards, one of them a Mega.

I have written chess programs for each of them. I have hooked up external components for each of them. I have gotten stand-alone applications running on each of them.

Hands down, the BB Black is the winner in every category except cost, and there the difference is less than a hour's wages for most engineers. And you are likely to lose that hour working around the relative shortcomings of an Arduino.

If there are a lot of support fora for the Arduino, then perhaps that's because a typical Arduino user needs more support. All of us were in that same boat one day, but not today.

The big idea behind the Arduino is to allow hardware novices and programming novices the ability to cheaply build prototypes of microcontroller gadgetry which use connections to external components. That's why the Arduino has a bunch of headers on each board. But most of the same signals, the same digital I/O pins are also on a Pi or a BB Black. And the Pi and the Black actually have more and faster pins.

I think that the idea of a uniform platform tournament may at last see a decent implementation with BB Black participants. Anyone with a somewhat portable chess program that can run under Linux will also run on a BB Black. If there are any tricky assembly language code in a program, it may be that the same functionality is available with use of the GCC intrinsics routines. (This is what I use in Symbolic for various bitboard hackery.) In fact, there are no differences in Symbolic's source with respect to ANY Linux target. A BB Black event will draw many more participants than an Arduino event, for there will be little work needed to port an existing program.

[DavidEather]

I have a great idea!

You try to design a tournament and make it as inclusive as possible - not everyone is a BSc.

Then you can do the work of hunting down possible sponsors, getting them on board, specifying EVERYTHING about the hardware and particularly the interfacing requirements so things can run without a hitch (and anyone who wants to can test their set up for compliance).

And then YOU can watch as someone has a dummy-spit because their favorite-of-the-day, more expensive, less supported, , less established, steeper learning curve hardware was not selected and the hardware that was, does not allow then to use their programing ideas.

Hell, when you get that done I might suggest your idea is no good just because you don't use an PCduino 8 (not related to Arduino - just a bad name). After all PCDuino 8 has 8 processors, twice the memory, twice the storage and will be able to stretch multi-core engines for less than twice the price too!

So, what are you going to do next? It won't take much to dissuade any sponsors from continuing. I suggest you take that route or ...

[DavidEather]

Just in case your listening..

The end result of hardware oneupmanship is that every piddly, little, job has to be done on a 16+ core I7 system costing a million dollars and then no one is happy.

BBB is nice - a said that before, but I intended a low cost inclusive tournament to included inexperienced people who might be intimidated by either hardware or software.

Less time starting with BBB? I bet you did Uno and Mega first?

[sje]

Less time starting with BBB? I bet you did Uno and Mega first?

No, the first single board computer I used to write a chess program had a Zilog z80 CPU; this was back in the 1970s.

If the goal is a Uniform Platform event, then then the 1 GHz BB Black is the way to go with the 700 MHz Raspberry Pi model B/B+ the best second choice. If for some reason one wants to include a camera or a touch screen, then the Pi would probably be the first choice as it has headers for these as standard equipment.

Asking people to learn Atmel assembly language coding, something that's needed to get a respectable strength chess program running on a 16 MHz Arduino, is asking too much when compared to asking them to learn Pascal, C, or C++, all of which are available free for the BB Black and the Pi. And asking too much means less applicants.

If you want sponsors (are they really needed?), then you might consider getting Texas Instruments, the makers of the core chips of the BB Black, to pony up. But their assistance may be unlikely as the BB Black is already selling very well and that's why it is often out-of-stock with long wait times. The story with the Pi is much the same.

[DavidEather]

Absolutely. I do hope no one expects to sit down with a cluster of Uno's and put a world champion in his place (unless that place was on-top).

The idea is to introduce concepts from the "maker" group and the "serious programmer" group to each other - by indulging in a relatively fun project.

I've known serious developers who have almost fainted at the sight of a bare PC and experienced technicians baulk at programing anything. Both would be better off if they knew a little of each others worlds.

On the other hand most programmers should never be trusted with a screw drive (when a program doesn't behave as it expected some instantly blame the hardware, rip the back of and replug and reconfigure until the machine won't work at all


At the end of the project the firmware will still hang around and hopefully become useful to people who want to try their hand at making things. Interfacing to some actual hardware and hacking existing software is a bit more interesting than some Arduino projects.

Cheers

David Eather

[stegemma]

I think it would be very difficult programming a chess engine on a 8bit microcontroller.
evene 16 bit is done through software library

My first trying in chess software were on 8 bit CPU and i think that it was not so hard.

[ZirconiumX]

Steven, *please* do the math.

Two Atmel ATMega328 chips - $8.

Two BeagleBone Black boards - $110.

$110/$8 = 1,375% more expensive.

And for $110 you can get a Pentium 20th Anniversary Edition, and a motherboard that allows basic over clocking.

Put that into an existing computer, and you've got something way more powerful than a beaglebone black.

Matthew:out