Archive

Posts Tagged ‘solidrun’

First OpenCL Encounters on Cortex-A72: Some Benchmarking

November 14th, 2017 1 comment

This is a guest post by blu about his experience with OpenCL on MacchiatoBin board with a quad core Cortex A72 processor and an Intel based MacBook. He previously contributed several technical articles such as How ARM Nerfed NEON Permute Instructions in ARMv8 or OpenGL ES development on Ubuntu Touch.

Qualcomm launched their long-awaited server ARM chip the other day, and we started getting the first benchmarks. Incidentally, I too managed to get some OpenCL ray-tracing code running on an ARM Cortex-A72 machine that same day (thanks to pocl – an LLVM-based open-source OCL multi-platform implementation), so my benchmarking curiosity got me.

The code in question is an OCL (half-finished) port of a graphics demo from 2014. Some remarks of what it does:

For each frame: a single thread builds a sparse voxel octree from a dynamic voxel scene; the octree, along with current camera settings are passed to an OCL kernel via double buffering; kernel computes a screen-space map of object IDs from primary-ray-hit voxels (kernel utilizes all compute units of a user-specified device); then, in headless mode used in the test, the app discards the frame. Test continues for a user-specified number of frames, and reports the average frames per second (FPS) upon termination.

Now, one of the baselines I wanted to compare the ARM machine against was a MacBook with Penryn (Intel Core 2 Duo Processor P8600), as the latter had exhibited very similar IPC characteristics to the Cortex-A72 in previous (non-OCL) tests, and also both machines had very similar FLOPS paper specs (and our OCL test is particularly FP-heavy):

  • 2x Penryn @ 2400MHz: 4xfp32 mul + 4xfp32 add per clock = 38.4GFLOPS total
  • 4x Cortex-A72 @ 1300MHz: 4xfp32 mul-add per clock = 41.6GFLOPS total

Beyond paper specs, on a SGEMM test the two machines showed the following performance for cached data:

  • Penryn: 4.86 flop/clock/core, 23.33GFLOPS total
  • Cortex-A72: 6.52 flop/clock/core, 33.90GFLOPS total

And finally RAM bandwidth (again, paper specs):

  • Penryn: 8.53GB/s (DDR3 @ 1066MT/s)
  • Cortex-A72: 12.8GB/s (DDR4 @ 1600MT/s)

On the ray-tracing OCL test, though, things turned out interesting (MacBook running Apple’s own OCL stack, which, to the best of my knowledge, is also LLVM-based):

  • Penryn average FPS: 2.31
  • Cortex-A72 average FPS: 7.61

So while on the SGEMM test the ARM was ~1.5x faster than Penryn for cached data, on the ray-tracing test, which is a much more complex code than SGEMM, the ARM speedup turned out ~3x? Remember, we are talking of two μarchs that perform quite closely by general-purpose-code IPC. Could something be wrong with Apple’s OCL stack? Let’s try pocl (exact same version of pocl and LLVM as on ARM):

  • Penryn average FPS: 11.58

OK, that’s much more reasonable. This time Penryn holds a speed advantage of 1.5x. Now, while Penryn is a fairly mature μarch that has reached its toolchain-support peak long ago, could we expect improvements from LLVM’s (and pocl’s) support for the Cortex family? Perhaps. In the case of our little test I could even finish the Aarch64 port of the non-OCL version of this code (originally x86-64 with SSE/AVX), but hey, OCL saved me the initial effort for satisfying my curiosity!

What is more interesting, though, is that assuming a Qualcomm Falkor core is at least as performant as a Cortex-A72 core in both gen-purpose and NEON IPC (not a baseless supposition), and taking into account that the top specced Centriq 2400 has 12x the cores and 10x the RAM bandwidth of our ARM machine, we could speculate about Centriq 2400’s performance on this OCL test when using the same OCL stack.

Hypothetical Qualcomm Centriq 2400 server: Centriq 2400 48x Falkor @ 2200MHz-2600MHz, 6x DDR4 @ 2667MT/s (128GB/s)

Assumed linearly scaling from the ARMADA 8040 measured performance; in practice the single-thread part of the test will impede the linear scaling, and so could the slightly-lower per-core RAM BW paper specs.

Of course, CPU-based solutions are not the best candidate for this OCL test — a decent GPU would obliterate even a 2S Xeon server here. But the goal of this entire test was to get a first-encounter estimate of the Cortex-A72 for FP-heavy non-matrix-multiplication-trivial scenarios, and things can go only up from here. Raw data for POCL tests on MacchiatoBin and MacBook is available here.

Industrial Shields Industrial Panel PCs are Based on Raspberry Pi, Banana Pi, or HummingBoard

October 10th, 2017 4 comments

Boot&Work Corp., S.L. is a company based in Catalonia that sells industrial automation electronic devices under “Industrial Shields” brand. What makes their product noticeable is that they all appear to be based on maker boards such as Arduino or Raspberry Pi.

The company offers various Arduino based PLC modules with or without Ethernet that can be controlled with 10.1″ industrial grade panel PCs based on ARM Linux development boards.

Click to Enlarge

Currently three sub-families are available:

  • HummTOUCH powered by Solidrun HummingBoard-i2 NXP i.MX 6Dual Lite board
  • BANANATOUCH with either Banana Pi M64 (Allwinner A64 quad core Cortex A53) or Banana Pi M3 (Allwinner A83T octa core Cortex A7)
  • TOUCHBERRY with Raspberry Pi model B or Raspberry Pi 3 model B

Beside the different processors, the 10.1″ Panel PCs share some of the same specifications:

Industrial Shields Arduino PLC – Click to Enlarge

  • Display – 10.1″ resistive multitouch LVDS, 315 nits, 170° viewing angle, 1280×720 resolution
  • Video Input – MIPI CSI connector (HummTouch only)
  • System Memory – 512MB to
    • HummTOUCH – 1 GB RAM
    • BANANATOUCH – 2GB RAM
    • BERRYTOUCH – 512MB RAM or 1GB LPDDR2
  • Storage
    • All – micro SD slot
    • BANANATOUCH – 8GB eMMC flash (16, 32, 64 GB optional)
  • Connectivity
    • Fast or Gigabit Ethernet depending on model
    • BANANATOUCH and BERRYTOUCH 3 – 802.11 b/g/n WiFi, Bluetooth 4.0
  • USB – 2x to 3x USB ports
  • I/O Expansion – 8x GPIO, SPI, I2C, UART
  • Power Supply – 12V DC; supports 7 – 18V DC input up to 1.5A
  • Dimensions – 325.5 x 195.6 x 95 mm
  • Compliance – CE

The user manual lists further details about environmental conditions, for example for HummTOUCH models:

  • Temperature Range – Operating: 0 to 45°C; storage: -20 to 60 C
  • Humidity – 10% to 90% (no condensation)
  • Ambient Environment – With no corrosive gas
  • Shock resistance – 80m/s2 in the X, Y and Z direction 2 times each.

There’s no information about Ingress Protection (IP) ratings, so it’s safe to assume those have not been tested for dust- and waterproofness.

Back of BANANATOUCH M3 Panel PC

The company also have smaller 3.5″ and 3.7″ model based on Raspberry Pi 3 board only. HummTOUCH models are available with either Linux or Android, BANANATOUCH and BERRYTOUCH models are only sold with Linux (Raspbian),  but Ubuntu, Android and Windows 10 IoT are options if they are supported by the respective board.

The 10.1″ panel PCs are sold for 375 to 460 Euros, and the Arduino based PLCs start at 135 Euros. Documentation and purchase links can all be found on Industrial Shields website.

NFV PicoPod is a Cluster of Six MACCHIATOBin Networking Boards for OPNFV, ODP, DPDK and OPF

June 29th, 2017 9 comments

If you are interested in networking applications, you may have already heard about Marvell ARMADA 8040 based SolidRun MACCHIATOBin board with multiple 10Gbps and Gbps network interfaces, three SATA ports, and more. PicoCluster has decided to make a cluster of 6 MACHIATOBin boards coupled with a Marvell Prestera DX 14 port, 10GbE switch for OpenDataPlane (ODP), Data Plane Development Kit (DPDK), OpenFastPath(OPF) as well as OPNFV (Open Platform for Network Functions Virtualization) .

Click to Enlarge

The main features of the full assembled kit –  named Cube – include:

  • 6x MACCHIATOBin quad core Cortex A72 boards (24 cores in total)
  • 1x Prestera DX 10GbE 14-port switch board
  • 6x fans
  • Power Supply
  • Acrylic case

The cluster kit comes with 64GB micro SD cards pre-loaded with the latest OPNFV Danube software release for ARM integrated by ENEA Software AB. The cluster is said to be compliant with the OPNFV Pharos specification.

Click to Enlarge

Three pre-order options are available with delivery scheduled for September 2017:

  • $1,699.00 – Starter kit with all required items  minus the MACCHIATOBin boards
  • $4,699.00 – Advanced kit with all required items including the Marvell boards, but it still needs to be assembled
  • $4,999.00 – Assembled Cube with everything as shown in the picture above

There’s also an option for 6 SSD mounts that add $100 to the kits. You may find more info on the product page, as well as on ARM Community’s blog.

Linux 4.11 Release – Main Changes, ARM & MIPS Architecture

May 1st, 2017 9 comments

Linus Torvalds has just released Linux 4.11:

So after that extra week with an rc8, things were pretty calm, and I’m much happier releasing a final 4.11 now.

We still had various smaller fixes the last week, but nothing that made me go “hmm..”. Shortlog appended for people who want to peruse the details, but it’s a mix all over, with about half being drivers (networking dominates, but some sound fixlets too), with the rest being some arch updates, generic networking, and filesystem (nfs[d]) fixes. But it’s all really small, which is what I like to see the last week of the release cycle.

And with this, the merge window is obviously open. I already have two pull request for 4.12 in my inbox, I expect that overnight I’ll get a lot more.

Linux 4.10 added Virtual GPU support, perf c2c’ tool, improved writeback management, a faster initial WiFi connection (802.11ai), and more.

Some notable changes for Linux 4.11 include:

  • Pluggable IO schedulers framework in the multiqueue block layer – The Linux block layer is know to have different IO schedulers (deadline, cfq, noop, etc). In Linux 3.13, the block layer added a new multiqueue design that performs better with modern hardware (eg. SSD, NVM). However, this new multiqueue design didn’t include support for pluggable IO schedulers. This release solves that problem with the merge of a multiqueue-ready IO scheduling framework. A port of the deadline scheduler has also been added (more IO schedulers will be added in the future)
  • Support for OPAL drives – The Opal Storage Specification is a set of specifications for features of data storage devices that enhance their security. For example, it defines a way of encrypting the stored data so that an unauthorized person who gains possession of the device cannot see the data. This release adds Linux support for Opal nvme enabled controllers. It enables users to setup/unlock/lock locking ranges for SED devices using the Opal protocol.
  • Support for the SMC-R protocol (RFC7609) – This release includes the initial part of the implementation of the “Shared Memory Communications-RDMA” (SMC-R) protocol as defined in RFC7609. SMC-R is an IBM protocol that provides RDMA capabilities over RoCE transparently for applications exploiting TCP sockets. While SMC-R does not aim to replace TCP, it taps a wealth of existing data center TCP socket applications to become more efficient without the need for rewriting them. A new socket protocol family PF_SMC is introduced. There are no changes required to applications using the sockets API for TCP stream sockets other than the specification of the new socket family AF_SMC. Unmodified applications can be used by means of a dynamic preload shared library.
  • Intel Bay Trail (and Cherry Trail) improvements – Intel HDMI audio support, patchsets for AXP288 PMIC, I2C driver, and C-state support to avoid freezes.

New features and bug fixes specific to ARM architecture:

  • Allwinner:
    • Allwinner A23 –  Audio codec device tree changes
    • Allwinner A31 – SPDIF output support
    • Allwinner A33 – cpufreq support, Audio codec support
    • Allwinner A64 – MMC Support, USB support
    • Allwinner A80 – sunxi-ng style clock support
    • Allwinner H2+ – New SoC variant, similar to H3 (mostly with a different, lower end VPU)
    • Allwinner H3 – Audio codec device tree changes, SPDIF output support
    • Allwinner V3s – New SoC support, USB PHY driver, pinctrl driver, CCU driver
    • New boards & devices – LicheePi One, Orange Pi Zero, LicheePi Zero, Banana Pi M64, Beelink X2
  • Rockchip:
    • Renamed RK1108 to RV1108
    • Clock drivers – New driver for RK3328, and non-critical fixes and clk id additions
    • Tweaks for Rockchip GRF (General Register File) usage (kitchensink misc register range on the SoCs)
    • thermal, eDP, pinctrl enhancements
    • PCI – add Rockchip system power management support
    • Add machine driver for RK3288 boards that use analog/HDMI audio
  • Amlogic
    • Add support for Amlogic Meson I2C controller
    • Add SAR ADC driver
    • Add ADC laddered keys to meson-gxbb-p200 board
    • Add configurable RGMII TX delay to fix/improve Gigabit Ethernet performance on some boards
    • Add pinctrl nodes for HDMI HPD and DDC pins modes for Amlogic Meson GXL and GXBB SoCs
    • New hardware: WeTek TV boxes
  • Samsung
    • Add USB 3.0 support in Exynos 5433
    • Removed clock driver for Samsung Exynos4415 SoCs
    • TM2 touchkey, Exynos5433 HDMI and power management improvements
    • Added Samsung Exynos4412 Prime SoC
    • Removed Samsung Exynos 4412 SoC
    • Added audio on Odroid-X board
    • Samsung Device Tree updates:
      • Add necessary initial configuration for clocks of display subsystem. Till now it worked mostly thanks to bootloader.
      • Use macro definitions instead of hard-coded values for pinctrl on Exynos7.
      • Enable USB 3.0 (DWC3) on Exynos7.
      • Add descriptive user-friendly label names for power domains. This  makes debugging easier
      • Use proper drive strengths on Exynos7.
      • Use bigger reserved memory region for Multi Format Codec on all Exynos chipsets so it could decode FullHD easily
      • Cleanup from old MACHs in s5pv210.
      • Enable IP_MULTICAST for libnss-mdns
      • Add bus frequency and voltage scalling on Exynos5433 TM2 device (along with  necessary bus nodes and Platform Performance Monitoring Unit on Exynos5433).
      • Use macros for pinctrl settings on Exynos5433.
      • Create common DTSI between Exynos5433 TM2E and TM2E.
  • Qualcomm
    • Added coresight, gyro/accelerometer, hdmi to Qualcomm MSM8916 SoC
    • Clock drivers – Updates to Qualcomm IPQ4019 CPU clks and general PLL support, Qualcomm MSM8974 RPM
    • Errata workarounds for Qualcomm’s Falkor CPU
    • Qualcomm L2 Cache PMU driver
    • Qualcomm SMCCC firmware quirk
    • Qualcomm PM8xxx ADC bindings
    • Add USB HSIC and HS phy driver for Qualcomm’s SoC
    • Device Tree Changes:
      • Add Coresight components for APQ8064
      • Fixup PM8058 nodes
      • Add APQ8060 gyro and accel support
      • Enable SD600 HDMI support
      • Add RIVA supprort for Sony Yuga and SD600
      • Add PM8821 support
      • Add MSM8974 ADSP, USB gadget, SMD, and SMP2P support
      • Fix IPQ8064 clock frequencies
      • Enable APQ8060 Dragonboard related devices
      • Add Vol+ support for DB820C and APQ8016
      • Add HDMI audio support for APQ8016
      • Fix DB820C GPIO pinctrl name
      • etc…
  • Mediatek
    • Mediatek MT2701 – Added clocks, iommu, spi, nand, adc, thermal
    • Added Mediatek MT8173 thermal
    • Added Mediatek IR remote receiver
  • GPU – Add Mali Utgard bindings;  the ARM Mali Utgard GPU family is embedded into a number of SoCs from Allwinner, Amlogic, Mediatek or Rockchip
  • Other new ARM hardware platforms and SoCs:
    • Marvell – SolidRun MACCHIATOBin board, Marvell Prestera DX packet processors
    • Broadcom – BCM958712DxXMC NorthStar2 reference board
    • HiSilicon – Kirin960/Hi3660 SoC, and HiKey960 development board
    • NXP – LS1012a SoC with three reference board; SoMs: Is.IoT MX6UL, SavageBoard, Engicam i.Core; Liebherr (LWN) monitor 6;
    • Microchip/Atmel – SAMA5d36ek Reference platform
    • Texas Instruments – Beaglebone Green Wireless and Black Wireless, phyCORE-AM335x System on Module
    • Lego Mindstorms EV3
    • “Romulus” baseboard management controller for OpenPower
    • Axentia TSE-850 Data Radio Channel (DARC) encoder
    • Luxul XAP-1410 and XWR-1200 wireless access points
    • New revision of “vf610-zii” Zodiac Inflight Innovations board

Finally here are some of the change made to MIPS architecture in Linux 4.11:

  • PCI: Register controllers in the right order to avoid a PCI error
  • KGDB: Use kernel context for sleeping threads
  • smp-cps: Fix potentially uninitialised value of core
  • KASLR: Fix build
  • ELF: Fix BUG() warning in arch_check_elf
  • Fix modversioning of _mcount symbol
  • fix out-of-tree defconfig target builds
  • cevt-r4k: Fix out-of-bounds array access
  • perf: fix deadlock
  • Malta: Fix i8259 irqchip setup
  • Lantiq – Fix adding xbar resoures causing a panic
  • Loongson3
    • Some Loongson 3A don’t identify themselves as having an FTLB so hardwire that knowledge into CPU probing.
    • Handle Loongson 3 TLB peculiarities in the fast path of the RDHWR  emulation.
    • Fix invalid FTLB entries with huge page on VTLB+FTLB platforms
    • Add missing calculation of S-cache and V-cache cache-way size
  • Ralink – Fix typos in rt3883 pinctrl data
  • Generic:
    • Force o32 fp64 support on 32bit MIPS64r6 kernels
    • Yet another build fix after the linux/sched.h changes
    • Wire up statx system call
    • Fix stack unwinding after introduction of IRQ stack
    • Fix spinlock code to build even for microMIPS with recent binutils
  • SMP-CPS: Fix retrieval of VPE mask on big endian CPUs”

Read Linux 4.11 changelog – with comments only – generated using git log v4.10..v4.11 --stat, to get the full list of changes. You may also want to checkout Linux 4.11 changelog on kernelnewbies.org.

SolidRun MACCHIATOBin Mini-ITX Networking Board is Now Available for $349 and Up

April 24th, 2017 31 comments

SolidRun MACCHIATOBin is a mini-ITX board powered by Marvell ARMADA 8040 quad core Cortex A72 processor @ up to 2.0 GHz and designed for networking and storage applications thanks to 10 Gbps, 2.5 Gbps, and 1 Gbps Ethernet interfaces, as well as three SATA port. The company is now taking order for the board (FCC waiver required) with price starting at $349 with 4GB RAM.

MACCHIATOBin board specifications:

  • SoC – ARMADA 8040 (88F8040) quad core Cortex A72 processor @ up to 2.0 GHz with accelerators (packet processor, security engine, DMA engines, XOR engines for RAID 5/6)
  • System Memory – 1x DDR4 DIMM with optional ECC and single/dual chip select support; up to 16GB RAM
  • Storage – 3x SATA 3.0 port, micro SD slot, SPI flash, eMMC flash
  • Connectivity – 2x 10Gbps Ethernet via copper or SFP, 2.5Gbps via SFP,  1x Gigabit Ethernet via copper
  • Expansion – 1x PCIe-x4 3.0 slot, Marvell TDM module header
  • USB – 1x USB 3.0 port, 2x USB 2.0 headers (internal),  1x USB-C port for Marvell Modular Chip (MoChi) interfaces (MCI)
  • Debugging – 20-pin connector for CPU JTAG debugger, 1x micro USB port for serial console, 2x UART headers
  • Misc – Battery for RTC, reset header, reset button, boot and frequency selection, fan header
  • Power Supply – 12V DC via power jack or ATX power supply
  • Dimensions – Mini-ITX form factor (170 mm x 170 mm)

Click to Enlarge

The board ships with either 4GB or 16GB DDR4 memory, a micro USB cable for debugging, 3 heatsinks, an optional 12V DC/110 or 220V AC power adapter, and an optional 8GB micro SD card. The company also offers a standard mini-ITX case for the board. The board supports mainline Linux or Linux 4.4.x, mainline U-Boot or U-Boot 2015.11, UEFI (Linaro UEFI tree), Yocto 2.1, SUSE Linux, netmap, DPDK, OpenDataPlane (ODP) and OpenFastPath. You’ll find software and hardware documentation in the Wiki.

The Wiki actually shows the board for $299 without any memory, but if you go to the order page, you can only order a version with 4GB RAM for $349, or one with 16GB RAM for $498 with the optional micro SD card and power adapter bringing the price up to $518.

SolidRun MACCHIATOBin is Another Marvell ARMADA 8040 Networking Mini-ITX Board

October 11th, 2016 34 comments

We’ve already seen SolidRun is working on a Marvell ARMADA 8040 quad core Cortex A72 community board for networking and storage applications, but based on a picture taken at Linaro Connect, the company is also working on a similar board with extra connectivity options called MACCHIATOBin.

Click to Enlarge

Click to Enlarge

Apart from the picture, there’s no info on the web about this board, so we’ll have to derive specs from the photo, the community board features, and info provided by Marcin Juszkiewicz, so all details are preliminary and subject to change:

  • SoC – ARMADA 8040 (88F8040) quad core Cortex A72 processor @ up to 2.0 GHz
  • System Memory – 1x DDR4 DIMM up to 16GB RAM
  • Storage – 3x SATA 3.0 port + micro SD slot
  • Connectivity – 1x Gigabit RJ45 port, 1x SFP SGMII @ 2.5Gbps, 2x 10Gbps copper (RJ45) with auto switchover to dual SFP+
  • Expansion – 1x PCIe-x4 3.0 slot, Marvell TDM module header
  • USB – 1x USB 3.0 port, 1x USB-C port
  • Debugging – 20-pin Connector for CPU JTAG debugger
  • Power Supply – 12V DC via power jack or ATX power supply
  • Dimensions – Mini-ITX form factor (170 mm x 170 mm)

That board is said to be SBSA compliant, meaning any ARM SBSA server distributions (like Red Hat) should work with mainlined kernel and bootloaders (U-Boot and UEFI). The price is said to be $350 with 4GB RAM, exactly what the community board is supposed to sell for, so MACCHIATOBin could also be the latest revision of the community board with a layout change, and most of the same features.

SolidRun ClearFog Base is a $90 Router/Networking Board with USB 3.0, M.2, mSATA, and Gigabit Ethernet Support

August 3rd, 2016 11 comments

SolidRun introduced ClearFog Pro and Base board based on Marvell Armada 380/388 processor at the end of last year, but at the time, only the higher-end ClearFog Pro board was available for $170 and up. Now the company  has officially launched the cheaper ClearFog Base board based on the same processor, two Gigabit Ethernet RJ45 ports, one SFP cage, a USB 3.0 port, an M.2 slot, mPCIe expansion slot, and more.

ClearFog_BaseClearFog Base board specifications:

  • Processor – Marvell ARMADA 388 (88F6828) dual core ARMv7 processor (Cortex A9 class) @ up to 1.6 GHz with 1MB L2 cache, NEON and FPU
  • System Memory –  1GB RAM by default (2GB optional)
  • Storage – 1x micro SD slot, optional 4GB eMMC flash, 1x M.2 slot, 1x mSATA/mPCIE
  • Connectivity – 2x dedicated Gigabit Ethernet ports, 1x SFP cage
  • USB – 1x USB 3.0 port
  • Expansions
    • 1x mini PCI Express slots (shared with mSATA )
    • 1x M.2 slot with USB 3.0, SATA, GNSS, 3G modules
    • mikroBUS socket for GPIOs, MikroElektronika Click Boards
    • 2x SIM card sockets
  • Debugging – micro USB port for serial console
  • Misc – RTC battery header, LEDs, user push buttons
  • Power Supply – 9 to 32V DC input; PoE expansion header
  • Dimensions – 103 x 75 mm (optional metal enclosure)

The board is comprised of a baseboard and a microSoM (in green), and runs OpenWrt or a Yocto Project build based on Linux 3.10.x, and other operating systems such as Arch Linux ARM, and Debian also appear to be supported. Hardware and software documentation can be found in the Wiki.

ClearFog_Base_M2_mPCieTypically applications for such boards include home media clouds (NAS), IoT gateways, and secure routers.

The board sells for $90 without power supply, nor internal storage, but 110V or 220V power adapters, a blank 8GB SD card, and a 4GB eMMC flash are all available as options.

Via Liliputing

ARMADA 8040 Networking Community Board with a Quad Core ARM Cortex A72 SoC Coming Soon for $300 and Up

June 29th, 2016 21 comments

Developers interested in ARMv8 server or networking boards are starting to have more and more affordable choices. After AMD Opteron A1100 series based LeMaker Cello board, and Softiron Overdrive 1000 server, SolidRun is now working on ARMADA 8040 networking community board powered by Marvell ARMA8040 quad core Cortex A72 network processor.

Click to Enlarge

Click to Enlarge

ARMADA 8040 networking board (mrvl8040) preliminary specifications:

  • SoC – ARMADA 8040 (88F8040) quad core Cortex A72 processor @ up to 2.0 GHz with MoChi architecture
  • System Memory – 4GB DDR4 DIMM by default
  • Storage – 4x SATA 3.0 port + micro SD slot
  • Connectivity – 1x Gigabit RJ45 port, 1x SFP SGMII @ 2.5Gbps, dual 10Gbps copper with auto switchover to dual SFP+
  • Expansion – 1x PCIe-x4 3.0 slot, Linaro 96Boards expansion slot exposing GPIO, UART, I2C and SPI, Marvell TDM module header
  • USB – 1x USB 3.0 port, 2x micro USB ports
  • Debugging – Console port (UART) over microUSB connector; 20-pin Connector for CPU JTAG debugger; OpenOCD debugger support over FTDI device
  • Power Supply – 12V DC via power jack or ATX power supply
  • Dimensions – Mini-ITX form factor (170 mm x 170 mm)

The complete hardware specifications have not been released yet, so many of the features above are derived from the 3D renders of the board. The board targets OpenDataPlane (ODP), OpenFastPath (OFP) and ARM network functions virtualization (NFV) ecosystem communities. The software will include a fully open source ODP implementation with  U-Boot 2015.x, mainline U-Boot, UEFI EDK2, Linux LTS kernel 4.4.x, mainline Linux, Yocto 2.1 and netmap.

Marvell_ARMADA_8040_Block_Diagram

Marvell ARMADA 8040 Block Diagram

ARMADA 8040 community networking board is scheduled to ship early September (early access) or mid October, but SolidRun is already taking pre-orders with a $50 discount bringing the price down to $299 with 4GB RAM, but you may want to add the power supply for $10 more. Marvell also plans to launch 8-,16- and 32-core versions of ARMADA 8040 SoC in in Q1 2017, but it’s unknown whether they’ll make it in to community boards.

Via ElectronicsWeekly