Works on Arm News for September 28, 2018.
In the news
- Ampere Computing announces 64-bit Arm system, Lenovo as OEM partner
- Interview by Charbax at Linaro Connect
- EBBR demo at Linaro Connect
- NixOS building for 32-bit, 64-bit Arm systems
- Kubernetes 1.12
- OpenJDK 11
- Orthanc, embedded AI for medical diagnostic imaging
Ampere Computing announces 64-bit Arm system, Lenovo as OEM partner
Ampere Computing has announced a new 32-core, 3.3Ghz design for a 64-bit Arm server, to be known as eMag. The Register has the speeds and feeds:
In summary: it’s a 32-core 64-bit Armv8 CPU clocked up to 3.3GHz in turbo mode, with a shared 32MB L3 cache. It supports up to 1TB of DRAM from 16 DIMMS plugged into eight DDR4-2667 memory controllers, has 42 lanes of PCIe 3.0, draws up to 125W, and is a single-socket chip. It’ll be made using 16nm TSMC FinFETs. The cores were designed by Silicon Valley-based Ampere, and the whole package is branded the eMAG.
HPCWire has this quote from Lenovo on the partnership:
“We have been working closely with the Ampere team over the past several months and have been impressed with the team and technology,” said Paul Ju, VP/GM Hyperscale, Lenovo Data Center Group. “As the fastest growing server vendor in the world, Lenovo is rapidly growing as the new clear choice for hyperscale customers by designing customized boards and systems to each hyperscaler’s unique requests, and building through our own extensive global factory and supply chain network. Together with Ampere, we will deliver a new generation of servers designed specifically with these customers in mind, providing the leadership quality, consistency and value they’ve come to expect from Lenovo.”
Interview by Charbax at Linaro Connect
In this video shot at Linaro Connect, Ed Vielmetti, director of the Works on Arm cluster, discusses technical issues regarding integration, testing, Cloud Native and network workloads, and directions for the project for the coming year.
The Works on Arm cluster is run by Packet for Arm to provide test, development, and data center CI/CD resources for community projects to build on arm64. The project also includes a weekly video office hours, a weekly newsletter, and a channel on the Packet Community Slack and Freenode IRC (#worksonarm) for community discussion.
EBBR demo at Linaro Connect
EBBR is a developing standard for booting Arm-based embedded systems. The problem to date is that each single-board computer has had a slightly different boot and operating system load process, necessitating a lot of work by operating system and distribution developers to incorporate the quirks of each into their development.
A demo at Linaro Connect showed Debian, Fedora, OpenSUSE, and FreeBSD, each booting (with various measures of success) from the same boot media on a variety of single-board systems.
NixOS building for 32-bit, 64-bit Arm systems
The NixOS operating system has a new build server that will support both 32-bit and 64-bit Arm builds. It’s hosted on a 2 x 32-core HiSilicon Hi1616 @ 2.4Ghz system at Packet, part of the Works on Arm infrastructure.
Kubernetes 1.12 has been released. In addition the main new development that will lead to improved scaling and stability, this release also introduces “fat manifests” for all core Kubernetes components. This makes it even easier to support a multiarchitecture cluster in Kubernetes.
OpenJDK 11 has been released. This newest iteration of Java promises performance improvements, and you should see releases of it coming soon for arm64 from the usual sources including Adopt OpenJDK.
Orthanc, embedded AI for medical diagnostic imaging
Orthanc aims at providing a simple, yet powerful standalone DICOM server. It is designed to improve the DICOM flows in hospitals and to support research about the automated analysis of medical images. Orthanc lets its users focus on the content of the DICOM files, hiding the complexity of the DICOM format and of the DICOM protocol.
With the help of the Works on Arm project, Derek Merck from Brown University has ported Orthanc to arm64 and to Docker. The goal of this exercise is to build a suitable system for embedding medical image recognition algorithms and machine learning directly into diagnostic imaging devices.