Just curious! Thank you! 
Iām pretty sure I heard the guys mentioning this as a possibility on BeagleCast if enough people
voice their interest; be aware that the cost increases exponentially with size though,
if the prices on Mouser is to be believed.
Itās kind of a bummer that the number of hard CPUs are still at 4 and still running at 667MHz,
but size wise the MPFS250 is a real beast.
Probably a tad too high of a price point in general, but still quite fair compared to a
similar sized Xilinx Zynq for example.
That price is so far out its not even worth thinking about. Unless you have a need for a very small board other solutions are much less.
They must be targeting the defense market with that one, seems out of range for a broad market base.
Also, the larger part would not really be a good match for the small beaglebone form factor. As it is, I wish the Fire had twice as many IO connectors. Most of the pins on the current part go nowhere. If you add an HDMI framer to the Fire, you really donāt have that many IOs left.
True, there does seem to be a lot of āwastedā pins on the package, but to be fair, I am still blown away by the fact that virtually all of the P8/P9 header pins are available for general IO; something that has always irked me with the am335x based boards. On top, we even got a MIPI and SYZYGY port to play with, so I for one, I am happy as a clam.
Would I shell out for a MPFS250 board? ā¦ Probably not, but for the right person, Iām still sure itās a good value proposition.
Not a Beagle platform, but hereās a decent example:
https://shop.trenz-electronic.de/en/TEM0007-01-CHE11-A-Microchip-PolarFire-SoC-FPGA-250T-FE-1-GByte-LPDDR4-SDRAM-4-x-5-cm?c=718
It looks like a product that is made for companies with deep R&D pockets. I did not find much substance on their website regarding actually getting that board to work. Looks more like an info-commercial to me.
Here is the āhidden agendaā
That verbiage is a red flag, it says, you will have enough information to get āblinkyā running. But the serious programming you will have to pay them to do. It is a very common ploy, and very expensive if you donāt notice how they manipulate you by tossing out only trivial information.
I canāt say that I completely disagree with the notion that there may be a hidden agenda, but I will say that I was merely attempting to address the OP. The Beagle user community is very strong, but there may not be an impetus to develop an MPFS250T based product. Trenz boards are very nice industrial components. Iāve used them in the past, and they are more esoteric in nature than Beagle. However, if someone has successfully gotten the Beagle product to work for them, and they need more fabric, the Trenz board is unlikely to be as much of a challenge after getting the Beagle to work, than going straight to the Trenz product. In fact, have a look at this:
https://wiki.trenz-electronic.de/display/PD/Modified+TE0703+for+Microchip+Getting+Started
The future TEB2000 may end up looking a lot like a Beagle board. Maybe someone should start a hardware project to make the Trenz boards drop into a Beagle outline? Up-gradable modules. Sounds like a winning concept.
The Trenz boards Iām familiar with were not locked to any given tool vendor. You can use Libero and gcc just like with the Beagle. There would be a task in port mapping to different boards, but thereās a good bit of that going on anyway.
That is interesting, I am going to look at there product again.
At the present we need to run DSP and the M7 is an option not really my first choice, would much rather have dedicated horse power.
Not being locked to a tool vendor is a major issue and that is nice. How long did it take you to get productive on that product?
I canāt say it was immediate and painless, but at the time I had a lot of help from other folks working on the same project. We used Libero and Softconsole along with gcc in both Windows and Linux to run FreeRTOS on the board. It came together pretty quickly due primarily to the expertise of the other folks on the team. We were using FreeRTOS because the chip was not a PolarFire. Presumably one would use Linux on a PolarFire based board.
1 Like