GPU

Tachyum Submits Bid for 20-Exaflop Supercomputer to U.S. Department of Energy Advanced Computing Ecosystems

Retrieved on: 
Tuesday, August 16, 2022 - 4:27pm

Tachyum today announced that it has responded to a U.S. Department of Energy Request for Information soliciting Advanced Computing Ecosystems for DOE national laboratories engaged in scientific and national security research.

Key Points: 
  • Tachyum today announced that it has responded to a U.S. Department of Energy Request for Information soliciting Advanced Computing Ecosystems for DOE national laboratories engaged in scientific and national security research.
  • Tachyum has submitted a proposal to create a 20-exaflop supercomputer based on Tachyums Prodigy, the world's first universal processor.
  • In its bid, Tachyum introduced its advanced universal processor technology and long-term roadmap that includes Prodigy, and Prodigy 2.
  • In addition, it will provide high performance, low TCO, low energy consumption, ease of deployment and low maintenance, said Dr. Radoslav Danilak, founder and CEO of Tachyum.

New 6U VPX Carrier Cards Host Dual XMC Modules to Support a Wide Variety of Signal Processing and Computing Options

Retrieved on: 
Tuesday, August 16, 2022 - 4:14pm

WIXOM, Mich, Aug. 16, 2022 /PRNewswire/ -- Acromag expands their OpenVPX™ carrier card selection with the addition of two new models that provide a simple and cost-effective solution for interfacing XMC modules to a VPX computer system. The VPX4840 and VPX4850 feature two XMC slots with support for front or rear panel I/O. They are available with VITA 42, VITA 61, or VITA 88 connectors to route power and interface bus signals to the plug-in mezzanine modules. Both models support a choice of direct PCIe connection to the VPX backplane via the data or expansion plane. The XMC sites have a 16-lane PCIe bus Gen 3 interface enabling rapid data throughput. By inserting XMC mezzanine modules on the carrier, including XMC processor (prXMC) modules, developers can now leverage hundreds of available function modules currently unavailable in a VPX platform. Pricing starts at $3995 for an air-cooled version with a 0 to 55°C range. Models with extended temperature ranges or conduction cooling support are also available.

Key Points: 
  • The VPX4840 and VPX4850 feature two XMC slots with support for front or rear panel I/O.
  • They are available with VITA 42, VITA 61, or VITA 88 connectors to route power and interface bus signals to the plug-in mezzanine modules.
  • Both models support a choice of direct PCIe connection to the VPX backplane via the data or expansion plane.
  • By inserting XMC mezzanine modules on the carrier, including XMC processor (prXMC) modules, developers can now leverage hundreds of available function modules currently unavailable in a VPX platform.

GAIMIN's Early Access Event opens its platform and monetization app to gamers

Retrieved on: 
Tuesday, August 16, 2022 - 3:00pm

ZUG, Switzerland, Aug. 16, 2022 /PRNewswire/ - Following recently completed initial testing of its monetisation and gaming platform, GAIMIN.io ltd is opening up its gaming and monetization platform to 10,000 gamers during August and September.

Key Points: 
  • ZUG, Switzerland, Aug. 16, 2022 /PRNewswire/ - Following recently completed initial testing of its monetisation and gaming platform, GAIMIN.io ltd is opening up its gaming and monetization platform to 10,000 gamers during August and September.
  • Starting on Monday 15th August 2022, we will be opening the download of our app to 10,000 users who can immediately benefit from its monetisation abilities and also receive bonuses for participating in this Early Access event."
  • We now need 10,000 gamers to test the performance of the app through high synchronous utilization, in readiness for general release."
  • To participate in this event, please follow GAIMIN on our social media channels, where full instructions will be provided throughout the Early Access period.

Verge.io Unveils Shared, Virtualized GPU Computing to Cut Complexity and Cost

Retrieved on: 
Tuesday, August 16, 2022 - 2:25pm

Current methods for deploying GPUs systemwide are complex and expensive, especially for remote users.

Key Points: 
  • Current methods for deploying GPUs systemwide are complex and expensive, especially for remote users.
  • Rather than supplying GPUs throughout the organization, Verge.io allows users and applications with access to a virtual data center to share the computing resources of a single GPU-equipped server.
  • Users/administrators can pass through an installed GPU to a virtual data center by simply creating a virtual machine with access to that GPU and its resources.
  • Alternatively, Verge.io can manage the virtualization of the GPU and serve up vGPUs to virtual data centers.

NVIDIA GTC to Feature CEO Jensen Huang Keynote Announcing New AI and Metaverse Technologies, 200+ Sessions With Top Tech, Business Execs

Retrieved on: 
Monday, August 15, 2022 - 2:00pm

GTC talks will explore some of the key advances driving AI and the metaverse -- including large language models, natural language processing, digital twins, digital biology, robotics and climate science.

Key Points: 
  • Huangs keynote will be livestreamed on Tuesday, Sept. 20, at 8 a.m. PT and available on demand afterward.
  • GTC will also feature a fireside chat with Turing Award winners Yoshua Bengio, Geoff Hinton and Yann LeCun discussing how AI will evolve and help solve challenging problems.
  • The discussion will be moderated by Sanja Fidler, vice president of AI Research at NVIDIA.
  • NVIDIA management will hold a Q&A session with financial analysts following the keynote; the webcast will be available at investor.nvidia.com.

Enhanced Power Connector Series from Amphenol Features Low Profile, Capacity to Carry High Current

Retrieved on: 
Thursday, August 11, 2022 - 4:52pm

ENDICOTT, N.Y., Aug. 11, 2022 /PRNewswire/ -- Amphenol Industrial Operations, a global leader in interconnect systems, has enhanced its Amphe-PD series with two low profile, two-way power connectors that have the capacity to carry high current. The Amphe-PD Mini 5.7mm and Gen2 5.7mm Amphe-PD can be used to transfer power from a PSU/PDU to a CPU, GPU PCB board.

Key Points: 
  • The Amphe-PD series of connectors are ideal when space is limited, like in rack and mount applications
    ENDICOTT, N.Y., Aug. 11, 2022 /PRNewswire/ -- Amphenol Industrial Operations, a global leader in interconnect systems, has enhanced its Amphe-PD series with two low profile, two-way power connectors that have the capacity to carry high current.
  • The Amphe-PD Mini 5.7mm and Gen2 5.7mm Amphe-PD can be used to transfer power from a PSU/PDU to a CPU, GPU PCB board.
  • The Amphe-PD Mini 5.7mm is designed to carry up to 120A at 500V DC.
  • The Gen2 5.7mm Amphe-PD can carry a maximum current of 150A at 500V DC.

TurbineOne Announces Strategic Partnership with FEDDATA to Empower American Warfighters Serving at the Tactical Edge

Retrieved on: 
Thursday, August 11, 2022 - 2:00pm

TurbineOne , the frontline perception company dedicated to empowering first responders and warfighters with the worlds best technology, today announced a new product partnership with FEDDATA to deliver AI to the comms-contested tactical edge.

Key Points: 
  • TurbineOne , the frontline perception company dedicated to empowering first responders and warfighters with the worlds best technology, today announced a new product partnership with FEDDATA to deliver AI to the comms-contested tactical edge.
  • Thankfully, major advances in edge computing now make it possible to deliver the power of the cloud and Artificial Intelligence (AI) to the tactical edge.
  • TurbineOne is excited to partner with FEDDATA, a Dell Titanium Partner, to empower the American warfighter with greater situational awareness, said Ian Kalin, CEO of TurbineOne.
  • Founded by veterans from elite organizations within Silicon Valley and the Department of Defense, TurbineOne delivers stronger situational awareness at the tactical edge through machine learning.

Adept Builds a Powerful AI Teammate for Everyone with Oracle and NVIDIA

Retrieved on: 
Thursday, August 11, 2022 - 1:05pm

AUSTIN, Texas, Aug. 11, 2022 /PRNewswire/ -- Adept, a machine learning research and product lab, is using Oracle Cloud Infrastructure (OCI) and NVIDIA technology to develop a universal AI teammate capable of performing a range of tasks people execute on their computer or on the internet. Running thousands of NVIDIA GPUs on clusters of OCI bare metal compute instances and taking advantage of OCI's network bandwidth, Adept can train large-scale AI and ML models faster and more economically than before. As a result, Adept has been able to rapidly advance its general intelligence roadmap and develop its first product – a rich language interface for the tools knowledge workers use every day to be productive and creative.

Key Points: 
  • With OCI as its preferred cloud platform, Adept obtains the scale and high performance necessary to run massive AI models without excessive compute costs.
  • To fully support Adept with the compute capacity it required, Oracle and NVIDIA customized their offerings to ensure Adept had access to thousands of NVIDIA A100 Tensor Core GPUs needed to train its complex models.
  • Adept, which recently closed a $65 million funding round, is training a giant AI model on OCI using NVIDIA's most powerful A100 GPUs connected with best-of-breed RoCE network powered by NVIDIA (NICs).
  • "By workingwith Oracle to provide Adept with an industry-leading GPU engine and a wide range of AI and ML software tools, we're making innovative AI systems possible."

LattePanda Team and Global Partners Jointly Launch LattePanda 3 Delta - the Fast and Pocket-sized Single-board Computer

Retrieved on: 
Thursday, August 11, 2022 - 7:19am

SHANGHAI, Aug. 11, 2022 /PRNewswire/ -- The LattePanda Team launched the world's thinnest pocket-sized hackable computer - LattePanda 3 Delta with global electronic components distributors.

Key Points: 
  • SHANGHAI, Aug. 11, 2022 /PRNewswire/ -- The LattePanda Team launched the world's thinnest pocket-sized hackable computer - LattePanda 3 Delta with global electronic components distributors.
  • "LattePanda Team is so proud to cooperate with the global electronic components distributors for this joint launch.
  • It delivers an exciting message to our customers that they can gain fast, easy access to our high-performance and hackable LattePanda 3 Delta anywhere in the world.
  • LattePanda 3 Delta is poised to define a new era of computing and drive mega creativity for tech enthusiasts and industry innovations.

General Atomics Scientists Leverage DOE Supercomputers to Advance Fusion Energy

Retrieved on: 
Wednesday, August 10, 2022 - 9:21pm

A key element of the drive to practical fusion energy is accurately simulating the behavior of the fusion fuel during operation.

Key Points: 
  • A key element of the drive to practical fusion energy is accurately simulating the behavior of the fusion fuel during operation.
  • These simulations allow researchers to test theories and design more effective experiments on devices like the DIII-D National Fusion Facility.
  • Such simulations require enormous computing power, and many can only be run effectively on high-end supercomputers like those operated by the DOE.
  • https://www.ga.com/images/BlogFeaturedImages/ga-scientists-leverage-DOE-...
    About General Atomics: Since the dawn of the atomic age, General Atomics innovations have advanced the state of the art across the full spectrum of science and technology from nuclear energy and defense to medicine and high-performance computing.