SC23 2023
who it's for
ai infrastructure engineers, high performance computing engineers, cloud infrastructure engineers, data center architects, systems architects, storage engineers, research computing directors, supercomputing center operators
topics
ethernet high-scale interconnects; PCIe with CXL disaggregation; multi-GPU servers (Dell XE9680/XE9640); liquid immersion cooling racks; GPU cloud and niche GPU providers (CoreWeave, etc.); on-prem exaFLOP benchmarking; memory disaggregation top-of-rack trays; NVMe per-GPU local storage; streaming inference at scale; Open OnDemand remote access; quantum-assisted simulation; modular compact data centers
| title | author | length |
|---|---|---|
| AnalystANGLE | Savannah Peterson, Lisa Martin | 18m |
SC23 coverage argues AI reshapes HPC with concrete shifts: Dell runs a two-week exaFLOP stand-up program to baseline AI workloads on-prem, Cambridge cites a… | ||
| Kickoff & Analysis | John Furrier, Savannah Peterson, Lisa Martin, Dave Nicholson | 15m |
Live coverage claims AI is reshaping HPC and industry by forcing a collision of chips and cloud providers, with cloud-native GPU players like CoreWeave GPU… | ||
| Hasan Siraj, Broadcom | SC23 | Hasan Siraj | 18m |
Broadcom argues that modern AI training makes the network the “computer,” backing that with products and numbers: Jericho3 and Tomahawk 5 target congestion and… | ||
| Jas Tremblay, Broadcom, and Ihab Tarazi, Dell Infrastructure Group | SC23 | Jas Tremblay, Ihab Tarazi | 21m |
Dell and Broadcom describe concrete outcomes: Dell’s XE9680 multi‑GPU server (eight GPUs + memory fabric) delivers about 10× single‑GPU performance and ships… | ||
| theCUBE Day 1 Show Wrap | John Furrier, Savannah Petwerson | 20m |
Coverage argues AI in HPC shifts from hype to measurable business impact with vendors showing customer stories and platform workarounds, citing Dell retooling… | ||
| theCUBE Day 2 Kickoff | Savannah Peterson, Lisa Martin, John Furrier | 15m |
Conference coverage reports Microsoft unveiling custom chips like Azure Cobalt CPU and Maia 100 while cloud and on‑prem stacks converge; Aurora claims it ran a… | ||
| theCUBE Day 2 Wrap | Savannah Peterson, Lisa Martin, David Nicholson, John Furrier | 21m |
Coverage summarizes that SC23 highlights AI reshaping HPC through vendor-academic-government collaboration ecosystems, claims that inference as killer app will… | ||
| Paul Calleja, University of Cambridge, and Andrea Booker, Dell Technologies | SC23 | Paul Calleja, Andrea Booker | 20m |
A converged AI+simulation PowerEdge XE9640 delivers 20 petaflops and roughly a thousand accelerator cards using Intel Ponte Vecchio and oneAPI for… | ||
| Johnny Dallas, Zeet | SC23 | Johnny Dallas | 19m |
Argument: niche clouds plus a metacloud control plane let teams pick best-in-class services — example: CoreWeave focuses on GPUs and is Kubernetes-first rather… | ||
| Sumit Puri, Liqid | SC23 | Sumit Puri | 18m |
Announces a concrete product: a 16-way L40S Dell server built with Nvidia and Dell using Liqid’s fabric and software to attach 16 GPUs to one bare-metal… | ||
| Shreya Shah, Dell Technologies, and Dave King, Denvr Dataworks | SC23 | Shreya Shah, Dave King | 23m |
Dell Technologies and Denvr Dataworks describe concrete solutions: modular data centers that compress a football-field cluster into under 900 sq ft, immersion… | ||
| Brian Beeler, StorageReview.com | SC23 | Brian Beeler | 22m |
Conference reporting argues that hardware access and cost—not hype—are the real bottlenecks, citing 40 weeks average wait for H100-class GPUs, workstation… | ||
| Jonathan Ross, Groq | SC23 | Jonathan Ross | 24m |
Groq describes an LPU built as a compiler‑first system rather than a single chip, running inference across a 576‑chip factory that pipelines work instead of… | ||
| Eric Baissus, KALRAY | SC23 | Eric Baissus | 14m |
Kalray describes a 15‑year journey from semiconductor maker to data management platform with a custom DPU plus a parallel file system and a data orchestrator… | ||
| Armando Acosta, Dell Technologies, and Alan Chalker, Ohio Supercomputer Center | SC23 | Armando Acosta, Alan Chalker | 21m |
State-funded Ohio Supercomputer Center aggregates four Dell clusters—about 60,000 cores and roughly $25,000,000 in hardware—and runs Open OnDemand which now… | ||
| David Flynn, Hammerspace | SC23 | David Flynn | 23m |
Hammerspace claims it installs installed in an afternoon onto existing NAS (Isilon) and immediately serves a unified metadata plane so data moves without… | ||
| Dr. Lisa Perez, Texas A&M High Performance Research Computing, and David Schmidt, Dell Technologies | Dr. Lisa Perez, David Schmidt | 18m |
ACES is a test‑bed for hardware composability that composes disaggregated resources (Dell R760s plus Nvidia and Intel GPUs) onto bare‑metal compute nodes so… | ||
| Vaughn Stewart, VAST Data | SC23 | Vaughn Stewart | 25m |
VAST Data argues the old tradeoff between high performance and low cost disappears by combining NVMe flash, disaggregated compute nodes and a software data… | ||
| Kimberly Leyenaar, Broadcom, and Steen Graham, Scalers AI | SC23 | Kimberly Leyenaar, Steen Graham | 20m |
Broadcom and Scalers AI describe running inference on 70B Llama models and fine-tuning Llama 2 on a PubMed corpus of tens of millions of records in a… | ||
| Renu Raman, Terizza | SC23 | Renu Raman | 19m |
Claims that large language models become model as unit of compute and may replace databases as systems of record, driving a shift to declarative… | ||
| Kevin Cochrane, Vultr | SC23 | Kevin Cochrane | 22m |
Vultr says it runs 32 global data centers and provisions Nvidia GPUs on Dell hardware to move customers from model training to regionalized inference clusters… | ||
| Bill Leslie, Dell Technologies | SC23 | Bill Leslie | 20m |
Claims that VxRail integrates GPUs into first-run and day-two lifecycle management and that VMware’s vSAN ESA single-tier design (removing cache) improves… | ||
| Joseph George and Andrew Wheeler, HPE | SC23 | Joseph George, Andrew Wheeler | 20m |
HPE says its systems occupy top five on Top500 placements (with partners AMD and Intel) and that four of the top five on GREEN500 are HPE machines, framing AI… | ||
| David Wohlford, IBM | SC23 | David Wohlford | 12m |
Claims a new industry-leading storage system that doubles prior throughput to 277 GB/s and pushes IOPS into the multi-millions, while adding computational… | ||
| Hemal Shah, Broadcom, and James Wynia, Dell Technologies | SC23 | Hemal Shah, James Wynia | 21m |
Broadcom shows production silicon for large AI fabrics: Jericho3-AI (a 100 billion+ transistor supercomputer switch chip) alongside Tomahawk 5, a top-of-line… | ||
| Dr. Happy Sithole, CSIR-NICIS | SC23 | Dr. Happy Sithole | 17m |
Government-funded National Integrated Cyber Infrastructure builds HPC, broadband and data management capacity across South Africa and parts of Africa, claiming… | ||
| Tommy Minyard, TACC | SC23 | Tommy Minyard | 21m |
TACC announces Stampede3 as an Intel Sapphire Rapids system with at least 560 liquid‑cooled Dell nodes (60 kW per rack), funded by a $10M NSF award and running… | ||
| John Zawistowski, Sycomp; Leon Platts, IBM; and Hugo Saleh, Google Cloud | SC23 | John Zawistowski, Leon Platts, Hugo Saleh | 21m |
Sycomp delivers a cloud-ready parallel file system called Sycomp Storage Fueled by IBM Storage Scale that hydrates data from Google Cloud Storage into a global… | ||
| Serban Zirnovan, Dell Technologies, and Guy D'Hauwers, atNorth | SC23 | Serban Zirnovan, Guy D'Hauwers | 22m |
atNorth builds HPC and AI only data centers across the Nordics that use direct liquid cooling to capture waste heat and sell it to district heat networks… | ||
| Scott Fadden and John (JZ) Zawistowski, Sycomp | SC23 | Scott Fadden, John (JZ) Zawistowski | 16m |
Sycomp describes a managed storage service that ties disparate silos across 46 countries into one workflow using IBM storage technology, offering… | ||
| theCUBE LIVE at SuperCompute 23 | Official Trailer | John Furrier, Lisa Martin, Dave Nicholson, Savannah Peterson | 01m |
A short trailer promises live reporting from SuperCompute 23 with a roving news desk for SiliconANGLE delivering on-site interviews and demos of the latest… | ||
| nothing matches | ||