www.fudzilla.com Open in urlscan Pro
194.36.44.31  Public Scan

Form analysis
1 forms found in the DOM

POST /news/ai

<form class="form-search" action="/news/ai" method="post">
  <div class="search">
    <label for="mod-search-searchword">Search ...</label><input name="searchword" id="mod-search-searchword" maxlength="200" class="input" type="text" size="20" placeholder="Search ...">
    <input type="hidden" name="task" value="search">
    <input type="hidden" name="option" value="com_search">
    <input type="hidden" name="Itemid" value="626">
  </div>
</form>

Text Content

Fudzilla It News and reviews
 * Home
 * News
   * Mobile
   * Graphics
   * AI & ML
   * PC Hardware
   * Transportation
   * Gaming
   * Cloud
   * IoT
   * Network
 * Reviews
 * About

 * Follow Us
   Facebook Twitter Print Email AddThis 0Share
   

Search ...

Published in AI


XILINX ANNOUNCES ALVEO U55C MOST POWERFUL ACCELERATOR CARD

by Fuad Abazovic on17 November 2021

   
 * font size decrease font size increase font size
   
 * Print
   
 * Email




HBM2 Purpose-built for HPC and big data

Xilinx has introduced the most powerful accelerator card ever purpose-built for
HPC and big data workloads. It will launch with the Alveo U55C to target even
bigger big data workloads from its predecessor. The new card uses new HBM2
memory, doubles the capacity, and uses a new form factor.

The new Alveo U55C card is the company’s most powerful Alveo accelerator card
ever, purpose-built for HPC and big data workloads. It offers the highest
compute density and HBM capacity in the Alveo accelerator portfolio. Together
with the new Xilinx RoCE v2-based clustering solution, a broad spectrum of
customers with large-scale compute workloads can now implement powerful
FPGA-based HPC clustering using their existing data center infrastructure and
network.

Alveo U55C is the single-slot full-height, half-length card. The hardware
programmable part of the card can deal with 1,304K Look-Up tables, 2,607K
registers, and 9,024 DSP slices. The most significant change is that the total
capacity of HBM 2 memory doubled from its predecessor from 8 to 16GB per car.
The new card supports Gen3 x16 or 2x Gen4 x8 interfaces, and it is a passive
card. The typical power is set to 115W, but the maximum power for the card has
been reduced from 225W to 150W.


The U55C provides more compute in a smaller form factor for creating dense Alveo
accelerator-based clusters. It’s built for high-density streaming data, high IO
math, and big compute problems that require scale-out, like big data analytics
and AI applications.

The reduction of total power enables a massive scale-out for the existing
customers targeting workloads that were not possible before. Customers can now
get double the HBM 2 memory at lower power to enable superior performance per
watt to HPC and database workload.

“Scaling out Alveo compute capabilities to target HPC workloads is now easier,
more efficient, and more powerful than ever,” said Salil Raje, executive vice
president, and general manager, Data Center Group at Xilinx. “Architecturally,
FPGA-based accelerators like Alveo cards provide the highest performance at the
lowest cost for many compute-intensive workloads. By introducing a
standards-based methodology that enables the creation of Alveo HPC clusters
using a customer’s existing infrastructure and network, we’re delivering those
key advantages at a massive scale to any data center. This is a major leap
forward for even broader adoption of Alveo and adaptive computing throughout the
data center.”

Built for HPC and big data applications

The Alveo U55C card combines many key features that today’s HPC workloads
require. It delivers more parallelism of data pipelines, superior memory
management, optimized data movement throughout the pipeline, and the highest
performance-per-watt in the Alveo portfolio.

Leveraging RoCE v2 and data center bridging, coupled with 200 Gbps bandwidth,
the API-driven clustering solution enables an Alveo network that competes with
InfiniBand networks in performance and latency, with no vendor lock-in. MPI
integration allows for HPC developers to scale out Alveo data pipelining from
the Xilinx Vitis™ unified software platform. Utilizing existing open standards
and frameworks, it’s now possible to scale out across hundreds of Alveo cards
regardless of the server platforms and network infrastructure and with shared
workloads and memory.

Software developers and data scientists can unlock the benefits of Alveo and
adaptive computing through high-level programmability of both the application
and cluster utilizing the Vitis platform . Xilinx has invested heavily in the
Vitis development platform and tools flow to make adaptive computing more
accessible to software developers and data scientists without hardware
expertise. The major AI frameworks like Pytorch and Tensorflow are supported, as
well as high-level programming languages like C, C++ and Python, allowing
developers to build domain solutions using specific APIs and libraries, or
utilize Xilinx software development kits, to easily accelerate key HPC workloads
within an existing data center.

HPC customer use cases

CSIRO, Australia’s national research organization along with the world’s largest
radio astronomy antenna array, is utilizing Alveo U55C cards for signal
processing in the Square Kilometer Array radio telescope. Deploying the Alveo
cards as network-attached accelerators with HBM allows for massive throughput at
scale across the HPC signal processing cluster. The Alveo accelerator-based
cluster allows CSIRO to tackle the massive compute task of aggregating,
filtering, preparing and processing data from 131,000 antennas in real time.
The 460Gbps of HBM2 bandwidth across the signal processing cluster is served by
420 Alveo U55C cards fully networked together across P4-enabled 100Gbps
switches. The Alveo U55C cluster delivers processing performance with overall
throughput at 15Tb/s in a compact power and cost efficient footprint.
CSIRO is now completing an example Alveo reference design in order to help other
radio astronomy or adjacent industries achieve the same success. 



Ansys LS-DYNA crash simulation software is used by nearly every automotive
company in the world. The design of safety and structural systems hinges on the
performance of models as they mitigate the costs of physical crash testing with
computer-aided design finite element method (FEM) simulations. FEM solvers are
the primary algorithms driving simulations with hundreds of millions of degrees
of freedom, these enormous algorithms can be broken out into more rudimentary
solvers like PCG, Sparse matrices, ICCG. By scaling out across many Alveo cards
with hyperparallel data pipelining, LS-DYNA can accelerate performance by more
than 5X in comparison to x86 CPUs. This results in more work per clock cycle in
an Alveo pipeline with LS-DYNA customers benefiting from game changing
simulation times.



“In the spirit of relentless innovation, we’re excited about collaborating with
Xilinx to significantly accelerate the finite-element solvers, which can
represent 90% of the compute workload for implicit mechanics, in our LS-DYNA
simulation application,” said Wim Slagter, strategic partnerships director at
Ansys. “We look forward to Xilinx acceleration helping us in our mission to
support innovators in engineering what’s ahead.”



TigerGraph, provider of a leading graph analytics platform, is using multiple
Alveo U55C cards to cluster and accelerate the two most prolific algorithms that
drive graph-based recommendation and clustering engines. Graph databases are a
disruptive platform for data scientists. Graphs take data from silos and bring
focus to the relationships between data. The next frontier for graph is finding
those answers in real time. Alveo U55C accelerates the query times and
predictions for recommendation engines from minutes down to milliseconds. By
utilizing multiple U55C cards to scale up analytics, the superior computational
power and memory bandwidth accelerates graph query speeds up to 45X faster
compared to CPU-based clusters. The quality of scores is also increases by up to
35 percent, resulting in greater confidence dramatically lowering false
positives to low single digits.

Product availability and easy evaluations

The Alveo U55C card is currently available on Xilinx.com and through Xilinx
authorized distributors. It’s also available for easy evaluation via public
cloud-based FPGA-as-a-Service providers, as well as select colocation data
centers for private previews. Clustering is available now for private previews,
with general availability expected in the second quarter of next year.

Xilinx is showcasing the Alveo U55C accelerator card, along with partner
solutions, at the SC21 conference taking place this week.

Last modified on 17 November 2021
Rate this item
 * 
 * 1
 * 2
 * 3
 * 4
 * 5

(0 votes)


Tagged under
 * xilinx
 * alveo
 * alveo u55c
 * hbm2
 * ai
 * HPC
 * big data




More in this category: « AI systems are really racist


back to top



MOST POPULAR - NOTEBOOKS

 * Xilinx announces Alveo U55C most powerful accelerator card
   
   
   
 * 


LATEST COMMENTS

 * bangskij
   
   what? when crossplatform steam works so well already /s
   
   Sweeny calls for universal app store · 5 hours ago

 * Dmitry
   
   Another sex pest (or at least the sex pesting facilitator) busted, but still
   denying everything...
   
   Activision Blizzard walk out over boss’s “cover-up” · 11 hours ago

 * Andrea Sibaldi
   
   "new chip in 2023".... gee... kind of hard to get excited about this. Intel
   and AMD will...
   
   Qualcomm prepares to give Apple’s chip a good kicking · 14 hours ago

 * Richard Vere Compton
   
   I own the ipr on a SPLAT engine concept. This engine mix 3 fuels together
   High pressure...
   
   The internal combustion engine is not dead yet · 16 hours ago

 * JohnW
   
   Hmm. A single point of failure and censorship. What could possibly go wrong?
   
   Sweeny calls for universal app store · 16 hours ago




 * Home
 * News
   * Mobile
   * Graphics
   * AI & ML
   * PC Hardware
   * Transportation
   * Gaming
   * Cloud
   * IoT
   * Network
 * Reviews
 * About