What is La Frite?
La Frite is a modern low power computer board with four 64-bit cores, up to a gigabyte of DDR4 RAM, ARM Mali GPU powering high definition HDMI, two USB ports, and ethernet. It is loosely based on the Raspberry Pi Model A+ and maintains similar GPIO header arrangement.
What is it suitable for?
La Frite is a smaller, less-featured, and more cost-optimized version of our first board Le Potato. Think of it as the Raspberry Pi 0 equivalent of the Raspberry Pi 3 if that is comparable to Le Potato. The new form factor enables smaller or more cost sensitive applications while offering nearly the same features. It utilizes the same software work that has already gone into Le Potato and will continue to be improved. The lower height design enables backwards-compatibility with HATs and more compact form factors.
It supports Ubuntu, Debian, LibreELEC, Lakka, RetroPie, Android Oreo, and many more Linux-based distributions. The low cost and power consumption enables more edgy use-cases like software DSP and IoT.
The possibilities are far reaching since the board comes with onboard SPI ROM for customized booting of upstream Linux from a variety of sources including USB flash drives, USB hard drives, and PXE. This along with the powerful Renegade Elite are part of our second generation platform initiative.
- Quad 64-bit ARM Cortex-A53 CPU Cores at 1.2GHz
- 2 Geometry + 3 Pixel ARM Mali-450 GPU Cores
- 512MB or 1GB DDR4 @ 2400MHz
- 128Mb SPI NOR
- HDMI 1.4 with 1080P Output
- 100Mb Fast Ethernet
- USB 2.0 Host
- USB 2.0 OTG
- IR Sensor
Some people have pointed out that we constantly come out with new boards instead of supporting old ones. This is a mis-conception of our open source software engineering initiative that we started last year with our first board Le Potato. Despite having different hardware, our platforms runs the same common software which is upstream Linux. All of the boards we have released gets all of the new feature enablement brought with every new version of Linux. Our goal is to have you pick the platform with the price and features you need for a specific task and run any upstream operating system image.
By following upstream Linux and u-boot, we can provide support for the latest software like Linux 4.19 instead of relying on obsolete vendor board-support-packages with outdated security-hole-ridden Linux kernels. This means that our platforms are up-to-date, secure, and will be supported for years or even decades to come. This kind of work is very expensive and not as fast since the quality standards are much higher. To date, none of the popular board makers have invested as much money as we have in this kind of open source software infrastructure. Their hardware becomes un-supported after a year or two when the silicon vendors deprecate support. By supporting us and using our hardware platforms, you will get more capabilities with time as Linux and other open source projects improve.
Most of the software engineering we sponsor for different platforms are not at the level that typical customers would understand. Rarely, it bubbles up to the surface into something easily consumed. As an example, Maxime Jourdan of BayLibre recently demoed hardware accelerated codec running on Chromium through the V4L2 Linux subsystem at Embedded Recipes Conference in Paris. Underneath that is layers of software engineering bits for various subsystems required for to get to that point. Soon, Le Potato and La Frite will be able to play H.265, H.264, and VP9 content on upstream Linux with Chromium, an industry first.
Unified Images with ARM64 Linux Kernel
64-bit ARM brings with it the possibility of an unified Linux kernel which means that software running on our boards with the exception of the bootloader (think of it like BIOS in a traditional computer) can be the same. What this means is that we can provide a single image that is bootable on any second generation platform. Software can then be packaged in the way that it is packaged for traditional AMD and Intel computers.
The SPI NOR built into La Frite will hold the boot firmware and provide a common interface for booting Linux. This will become useful once ARM finalizes the EBBR boot specifications. We aim to become one of the first providers to offer such a hardware/software integrated stack.
That is not to say this won't be supported on our first generation boards like Le Potato, Tritium, and Renegade*. They will simply have to write the bootloader into the empty space at the beginning of an image instead of loading it from the SPI NOR.
* Rockchip still needs to improve RK3328 upstream.
In our experience, the work that goes into changing engineering paradigm takes around two years for those with the grit and know-how. Since our project started almost one year ago, we have been making significant progress in all directions pertaining to user-friendliness of our boards infrastructure in terms of software, hardware, and web. By Q1 2019, we will have reached the completion of our initial vision. When you receive the boards, most of this will already be evident. We thoroughly value your feedback and have provided numerous avenues to report bugs and inquire about improvements.
People can build their ideas without worrying about the non-enterprise aspect of embedded systems and treat it like a traditional computer running Ubuntu, Debian, Arch, Fedora, or SuSE. By following standards, there is an eureka moment when you realize that you can focus on building your idea of a product or service instead of re-inventing the wheel through backporting kernel modules, security patches, stability fixes, and needless-proprietary re-inventions. We shouldn't all have to become embedded software engineers and learn a new platform every time we get started.
Porting new open source software to our platforms will not be an additional task that requires proprietary tweaking. The architecture agnostic code and stack you run on a desktop computer for Python, PHP, Java, Go, MySQL, Apache, or whatever else have you will run nearly identically on our platforms. That know-how follows established standards and are not proprietary to a vendor. It works the open source way and will finally make sense to developers.
We hope that you participate in our project and we hope these tidbits of information will provide you with a clearer understanding of what we are about and what you are supporting.