We have some breaking news to cover today as ML Commons releases the results of their training, HPC, and tiny benchmark suites. Every company in the machine learning space found something to talk about here, with Intel showing Sapphire Rapids and Habana Gaudi 2, Nvidia showing H100, and ARM’s tiny inferencing results really shining. But not everyone was happy, with some players complaining that the tests are being gamed by the big players.
0:46 | HPE Introduces Next-Gen Servers
The next generation of HPE servers have arrived. The eleventh generation of hardware from the computing giant is focused on a whole host of hot workloads like AI, Cloud-native Apps, and virtualization. One of the biggest advancements in the platform is the use of HPE’s iLO remote management system to verify and authenticate all the server hardware through a secure root of trust. Also of note is the lineup of processors available for Gen11. You can choose from AMY EPYC as well as Ampere. However, there is no Intel option listed on the site as of the time of this recording.
Read More: Hewlett Packard Enterprise introduces next-generation compute engineered for a hybrid world
6:26 | SolarWinds Settles with Investors
The SolarWinds supply chain story is still going strong. News broke this week that the monitoring giant has finally settled a shareholder lawsuit. The suit, which ended up costing $26 million, alleged that SolarWinds misled them about their security posture in advance of the cyberattack that saw their software compromised. The original intrusion happened back in 2019 and the shareholder suit was filed back in January 2021, shortly after the disclosure of the breach. Solarwinds also said that they’ve received notice from the SEC that they face future enforcement actions, most likely fines.
Read More: SolarWinds reaches $26m settlement with shareholders, expects SEC action
10:12 | Microsoft and Github Sued over Co-Pilot AI
Microsoft has found itself in the crosshairs of a proposed class-action lawsuit over AI. Specifically, the lawsuit is looking at Github’s use of the maligned Copilot. The software, which uses AI to help programmers write better code, has been accused of wholesale stealing of copyrighted code throughout the entire site. Researchers have found that the program has replicated large blocks of code without proper attribution to the original authors or owners of the programs. The lawsuit could change the face of how AI can be used to aid users or how the inputs of publicly available sources need to be attributed.
Read More: The lawsuit that could rewrite the rules of AI copyright
16:08 | Patch Tuesday Tries to Prevent Six Exploits
The November Patch Tuesday event from Microsoft was filled with some important fixes. Six zero-day exploits were listed in the notes, two of which were high priority Exchange patches that were actively being exploited. Microsoft was previously stated that the biggest exploit in Exchange has only been the work of a single threat actor but they made sure to include patches to prevent that from spreading. Another patch came from a notification of a Google group that tracks nation state actors using exploits against their enemies.
Read More: Patches for 6 0-days under active exploit are now available from Microsoft
19:24 | Intel Admits Sapphire Rapids HPC Is Late
Intel is releasing the fourth-generation Xeon CPUs, code name Sapphire Rapids with High-Bandwidth Memory for HPC! But it’s not quite ready yet… The company admitted at SC22 that, although Sapphire Rapids CPUs and Ponte Vecchio GPUs are already appearing, the HBM-equipped Max platform won’t be widely available for another quarter. Intel had planned to use the Atom-based Xeon Phi for Argonne’s Aurora but the new configuration uses the so-called Max Series CPUs instead. And it’s still on the way.
Read More: One New Feature for Intel’s HPC Compute Engines: Contrition
25:19 | AI Performance is On Display with MLPerf
We have some breaking news to cover today as ML Commons releases the results of their training, HPC, and tiny benchmark suites. Every company in the machine learning space found something to talk about here, with Intel showing Sapphire Rapids and Habana Gaudi 2, Nvidia showing H100, and ARM’s tiny inferencing results really shining. But not everyone was happy, with some players complaining that the tests are being gamed by the big players.
Read More: Latest MLPerf Results Display Gains for All
Intel Delivers Leading AI Performance Results on MLPerf v2.1 Industry Benchmark for DL Training
NVIDIA Hopper, Ampere GPUs Sweep Benchmarks in AI Training
36:50 | The Weeks Ahead
Supercomputing, November 13-18
Security Field Day 8, November 16-18, 2022
The Gestalt IT Rundown is a live weekly look at the IT news of the week. It broadcasts live on Facebook every Wednesday at 12:30pm ET. To watch along, “Like” our Facebook page. Be sure to subscribe to Gestalt IT on YouTube for even more weekly video content.
