HPC & Quantum
27 subscribers
11.4K photos
668 videos
3 files
30.6K links
Download Telegram
HPC Guru (Twitter)

The new era needs a new #network: Industry Heavyweights Form Ultra #Ethernet Consortium for #HPC and #AI

Founding members include AMD, Arista, Broadcom, Cisco, Eviden (Atos), HPE, Intel, Meta and Microsoft

https://t.co/BPGZWNeQkW

https://t.co/V4cwgGngDo

via @insideHPC https://t.co/rKVPGrb9fc
HPC Guru (Twitter)

.@Azure #Boost offloads virtualization processes traditionally performed by the hypervisor and host OS, such as networking, storage, and host management, onto purpose-built hardware and software

MANA smartNICs to counter AWS Nitro?

https://t.co/FeKBWzqZup

#Cloud #Network https://t.co/uAlSWmla56
HPC Guru (Twitter)

More on #MN5 @BSC_CNS:

o General Purpose: ~35 PF (HPL) 1120 nodes with 2x 40-core @intel Sapphire Rapids

o Accelerated: ~163 PF (HPL) based on @nvidia Hopper and Grace CPUs

o #Network: 100 Gbps / 800 Gbps

o #Storage: 248PB HDD, 2.8PB flash, 400 PB tape

#HPC #AI
HPC Guru (Twitter)

Why @Meta is determined to make #Ethernet work for #AI

@Meta is one of the founding companies behind the @ultraethernet Consortium, a group of companies -according to @TDaytonPM - united against a common enemy - #Infiniband

nextplatform.com/2023/09/26/…

#HPC #Network
HPC Guru (Twitter)

Cornelis Networks announces CN5000 #OmniPath High-Performance #interconnect solution for #AI and #HPC

hpcwire.com/off-the-wire/cor…

#SC23 #Network @CornelisHPC via @HPCwire
HPC Guru (Twitter)

Cornelis Networks announces CN5000 #OmniPath High-Performance #interconnect solution for #AI and #HPC

hpcwire.com/off-the-wire/cor…

#SC23 #Network @CornelisHPC via @HPCwire
HPC Guru (Twitter)

3/ Nvidia roadmap at 1-year rhythm

Rubin platform will succeed Blackwell platform

o Featuring new GPUs

o New Arm-based CPU - Vera

o Advanced networking with NVLink 6, CX9 SuperNIC and the X1600 converged InfiniBand/Ethernet switch

#AI #HPC #GPU #ARM #Network
HPC Guru (Twitter)

Tencent upgraded its #HPC #network to access idling #GPU capacity offering a 20% improvement in #LLM training

Xingmai network can support a single computing cluster with more than 100,000 GPUs

https://www.scmp.com/tech/big-tech/article/3268901/tencent-boosts-ai-training-efficiency-without-nvidias-most-advanced-chips

#AI #GenAI
HPC Guru (Twitter)

.@EuroHPC_JU is launching a call to build a high-speed, secure #network connecting #European #supercomputers and data centers, creating a hyperconnected and federated #HPC and #QuantumComputing ecosystem

https://eurohpc-ju.europa.eu/eurohpc-ju-launches-call-tender-hyperconnect-european-supercomputers-2024-12-16_en