Facebook’s Big Sur



Big Sur is our newest Open Rack-compatible hardware designed for AI computing at a large scale. In collaboration with partners, we’ve built Big Sur to incorporate eight high-performance GPUs of up to 300 watts each, with the flexibility to configure between multiple PCI-e topologies. Leveraging NVIDIA’s Tesla Accelerated Computing Platform, Big Sur is twice as fast as our previous generation, which means we can train twice as fast and explore networks twice as large. And distributing training across eight GPUs allows us to scale the size and speed of our networks by another factor of two.

In addition to the improved performance, Big Sur is far more versatile and efficient than the off-the-shelf solutions in our previous generation. While many high-performance computing systems require special cooling and other unique infrastructure to operate, we have optimized these new servers for thermal and power efficiency, allowing us to operate them even in our own free-air cooled, Open Compute standard data centers. Big Sur was built with the NVIDIA Tesla M40 in mind but is qualified to support a wide range of PCI-e cards. We also anticipate this will achieve efficiencies in production and manufacturing, meaning we’ll get a lot more computational power per dollar we invest.

Servers can also require maintenance and hefty operational resources, so, like the other hardware in our data centers, Big Sur was designed around operational efficiency and serviceability. We’ve removed the components that don’t get used very much, and components that fail relatively frequently — such as hard drives and DIMMs — can now be removed and replaced in a few seconds. Touch points for technicians are all Pantone 375 C green, the same touch-point color as all of Facebook’s custom data center hardware, which allows technicians to intuitively identify, access and remove parts. No special training or service guide is really needed. Even the motherboard can be removed within a minute, whereas on the original AI hardware platform it would take over an hour. In fact, Big Sur is almost entirely toolless — the CPU heat sinks are the only things you need a screwdriver for.
Collaboration through open source

We plan to open-source Big Sur and will submit the design materials to the Open Compute Project (OCP). Facebook has a culture of support for open source software and hardware, and FAIR has continued that commitment by open-sourcing our code and publishing our discoveries as academic papers freely available from open-access sites. We’re very excited to add hardware designed for AI research and production to our list of contributions to the community.

We want to make it a lot easier for AI researchers to share techniques and technologies. As with all hardware systems that are released into the open, it’s our hope that others will be able to work with us to improve it. We believe that this open collaboration helps foster innovation for future designs, putting us all one step closer to building complex AI systems that bring this kind of innovation to our users and, ultimately, help us build a more open and connected world.








Open Source Hardware for Scientists

Joshua M. Pearce has an article on Scientific American’s blog, talking about making lab hardware using Open Source Hardware model.

Science for All: How to Make Free, Open Source Laboratory Hardware

[…] If you are a working scientist, it is in your best interest to share the designs for your hardware. Consider for example the humble lab jack, which is used to move heavy or bulky equipment small distances up or down (we use one in my lab to move solar photovoltaic materials into a light path, for example). I received a shocking $950 quote for one. Although much less expensive lab jacks are available, we wanted a customizable lab jack. We designed one for less than $5 and shared it on the web. Then the fun began. A Finnish maker recommended an improved assembly that made it better and since then over 3,700 others have downloaded the designs. Next, a French scientist shared a better design that reduces the number of non-3-D printed parts. Most recently a Seattle-based company posted a 100% 3-D printable version. It even prints assembled! Now anytime I need a lab jack I can print one out of recyclebot-made plastic filament for a few pennies. […]

Full article:

PS: Next year’s Open Hardware Summit will be in Portland Oregon in October 2016-timeframe:


Olimex ARM64 OSH laptop update

Olimex is working on an Open Source Hardware-based AArch64-based laptop, based on their Open Source Hardware-based AArch64 dev board. They have a update on the system. Including some prototype pictures:

“needless to mention this window button will become Tux. :-)”


I wonder about what firmware they’ll use, and if the use will be able to update it themselves, from source….


OSHWA announces Open Source Hardware Certification v1

Excerpting the initial text:
Open Source Hardware Certification Version 1

This is version 1 of an official certification for open source hardware housed in the Open Source Hardware Association.  It outlines the purpose and goals of such a certification, and establishes the mechanisms for the operation of the certification process itself.

Primary Goals
*  Make it easier for the public to identify open source hardware.
*  Expand the reach of open hardware by making it easier for newer members to join the open source hardware community.

Read the full document:




Andy Simpkins: Developing products in the open

Last month, at the DebConf, the annual Debian conference, there was a talk for OEMs on using Open Source Software and Free Software (FOSS), as well as using Open Source Hardware. Most have heard an Open Source advocate discuss the merits of open-source -vs- closed-source software. This talk is from the perspective of vendors who need to buy and build hardware.

Developing products in the open
Andy Simpkins
Over the last couple of decades the world of product development with embedded systems has changed considerably. Changing to Open Source (for hardware as well as software) is not easy. The world resists change, this is a brief history of where I have succeeded, where I have failed and the lessons learned. This is a not a technical talk, more a collection of observations.

The video can be downloaded in WBEM format from DebConf site, or watched online via Youtube:




Parallax announces JTAGulator:


Designed by Grand Idea Studio, JTAGulator is an open source hardware tool that assists in identifying OCD connections from test points, vias, or component pads on a target device.

The product is currently on sale for the next few days, 15% off until 2015-08-31 (US$136 instead of US$160).