EDGE COMPUTING & AI
Ampere debuts 80-core ARM
server processor
Ampere Computing LLC has announced a 64-
bit ARM-based processor for servers comprising
80 cores and targeting 7nm manufacturing
process technology. Altra is a follow-up to the
32-core Skylark processor, otherwise known
as eMAG, which has been manufactured in
TSMC’s 16nm FinFET process. Even as that
product came to market in 2018 the company
was advising of a 7nm processor to follow (see
Startup Ampere prices ARM server chips ). Altra
is an 80-core processor based on Neoverse N1
license, designed for predictable high performance,
security isolation, scalability and leading
power efficiency. It comes in single socket
and dual socket versions, which are sampling
to customers. The device is scalable up to
80 cores per chip with a power consumption
footprint of up to 210W. The cores are specificed
up to 3.0GHz clock frequency. The core is
a four-instruction wide superscalar out-of-order
processor based on the ARMv8.2 instruction
set architecture. Each core is single-threaded
to provide reduced performance variability
and increased security against system-level
side-channel attacks. There are 64kbyte levelone
instruction and data caches per core and
1Mbyte level 2 data caches per core. There is
also a processor-wide 32Mbyte system-level
cache. There are 8 72bit-wide DDR4 interfaces
per chip and 128 lanes of PCIe Gen4 for highest
IO bandwidth. Altra is Ampere’s cloud-focused
product, and first in a new class of CPUs rolling
out on an annual basis from Ampere’s roadmap.
The increased performance and power efficiency
will make Altra suitable for many workloads
including data analytics, artificial intelligence,
database, storage, telco stacks, edge computing,
web hosting and cloud native applications.
Jeff Wittich, senior vice president of products
at Ampere, said that for some intensive training
loads arrays of GPUs or some more specialized
tensor processors or neural network accelerators
could possibly provide superior power
efficiency.
Ampere Computing
https://amperecomputing.com/
Smart voice control SoC targets home appliances
Xsail Technology, a subsidiary of set-top box chipset provider ALi Corporation,
has launched the G1, a low power and cost-effective AI chipset to be designed
into all kinds of voice-controlled home appliances. Xsail has developed its own
proprietary Edge Voice Process Unit (EVPU) for NN
(Neural Network) based voice recognition performed
by the G1 chipset, allowing various home appliances
to be controlled by online/offline voice and voice print.
The optimized voice pre-processing technology also
supports noise-reduction, echo cancellation and beam
forming in order to keep a good recognition rate under
different complex circumstances. The offline voice control module of G1 is to
be used in many smart products, supporting multiple languages in offline mode,
Chinese, English, German, French, Spanish, Japanese, Korean and Hindi etc.
G1 also includes a high performance dual core CPU that can support Linux and
RTOS. It also supports the playback of various audio formats, security encryption/
secure boot, LED display and all kinds of control interfaces.
Ali
www.alitech.com
Silicon photonic engine supports ML/AI workloads
Photonics company Ranovus has announced the Odin platform, scaling its
100Gbps per lambda silicon photonics engine from 800 Gbps to 3.2 Tbps in a
single chip. Delivered as a module or co-packaged with optics, the Odin platform
integrates the company’s 100Gbps silicon photonics
based on Micro Ring Resonator modulators
and photodetectors, 100Gbps driver, 100Gbps
TIA and control IC. The Odin platform delivers
50% reduction in power consumption/Gbps and
a 75% cost/Gbps reduction over today’s solutions,
claims the company’s CEO Hamid Arabzadeh.
The silicon photonics engine supports ML/AI applications with a 0.4ns
low latency and a transmission distance of 10m to 2km in CWDM and DWDM
applications. It offers 8 optical channels of 100Gbps/64Gbps/50Gbps PAM4 or
50Gbps/32Gbps/25Gbps NRZ. Suitable for integration in QSFP-DD and OSFP
module form factors, the new engine supports 25.6Tbps and 51.2Tbps Ethernet
Switch configurations.
Ranovus
www.ranovus.com
Fleet management integrated video solution
leverages AI to improve safety
Verizon Connect Integrated Video uses artificial intelligence (AI) to intelligently
capture and automatically classify video according to how severe an event is,
showing only what is relevant and important
to business owners and operations managers,
as well as machine learning to help businesses
improve driver behaviour and protect the bottom
line. Verizon Connect Integrated Video also
provides real data insights to help commercial
drivers stay safe on the road and protect them
against false claims. Additional features include speed overlay, which enables
operations managers to easily view the speed of the vehicle directly within the
video clip to determine if speed is the factor that caused the event, and video on
demand, which allows operations managers to request 40-second increments of
available footage.
Verizon Connect
www.verizonconnect.com
www.eenewseurope.com eeNews Europe April 2020 News 43
/
/
/www.alitech.com
/www.ranovus.com
/www.verizonconnect.com
/www.eenewseurope.com
