Category Archives: Stiri IT Externe

The Spectre/Meltdown Performance Impact On Linux 4.20, Decimating Benchmarks With New STIBP Overhead

As outlined yesterday, significant slowdowns with the Linux 4.20 kernel turned out to be due to the addition of the kernel-side bits for STIBP (Single Thread Indirect Branch Predictors) for cross-HyperThread Spectre Variant Two mitigation. This has incurred significant performance penalties with the STIBP support in its current state with Linux 4.20 Git and is enabled by default at least for Intel systems with up-to-date microcode. Here are some follow-up benchmarks looking at the performance hit with the Linux 4.20 development kernel as well as the overall Spectre and Meltdown mitigation impact on this latest version of the Linux kernel.

Some users have said AMD also needs STIBP, but at least with Linux 4.20 Git and the AMD systems I have tested with their up-to-date BIOS/microcode, that hasn’t appeared to be the case. Most of the AMD STIBP references date back to January when Spectre/Meltdown first came to light. We’ll see in the week ahead if there is any comment from AMD but at this time seems to be affecting up-to-date Intel systems with the Linux 4.20 kernel.

One of the most common request since yesterday’s article when bisecting it down to STIBP as the cause for the Linux 4.20 performance drop, many Phoronix readers were curious to know the overall performance cost of all the Spectre / Meltdown mitigations that have come about so far this year. I happened to do some tests on the latest Linux 4.20 Git both in its default mitigated state “KPTI + __user pointer sanitization + Full generic retpoline IBPB IBRS_FW STIBP RSB filling + SSB disabled via prctl and seccomp + PTE Inversion; VMX: conditional cache flushes SMT vulnerable” and then again when disabling the mitigations permitted at run-time.

The comparison of the overall Spectre/Meltdown cost on Linux 4.20 was done with the Intel Core i9 7980XE.

Following that comparison are some tests of the dual Intel Xeon Gold server with Linux 4.19, Linux 4.20, and then Linux 4.20 with no mitigations enabled. Those results are compared to the current AMD EPYC performance for seeing how the introduction of STIBP affects that positioning.

All of these benchmarks were facilitated in a fully-automated and reproducible manner using the open-source Phoronix Test Suite benchmarking software.

Acumos Project’s 1st Software, Athena, Helps Ease AI Deployment | Software

By Jack M. Germain

Nov 16, 2018 5:00 AM PT

LF Deep Learning Foundation on Wednesday announced the availability of the first software from the
Acumos AI Project. Dubbed “Athena,” it supports open source innovation in artificial intelligence, machine learning and deep learning.

This is the first software release from the Acumos AI Project since its launch earlier this year. The goal is to make critical new technologies available to developers and data scientists everywhere.

Acumos is part of a Linux Foundation umbrella organization, the LF Deep Learning Foundation, that supports and sustains open source innovation in artificial intelligence, machine learning and deep learning. Acumos is based in Shanghai.

Acumos AI is a platform and open source framework that makes it easy to build, share and deploy AI apps. Acumos standardizes the infrastructure stack and components required to run an out-of-the-box general AI environment, freeing data scientists and model trainers to focus on their core competencies, and accelerating innovation.

“The Acumos Athena release represents a significant step forward in making AI models more accessible for builders of AI applications and models, along with users and trainers of those models and applications,” said Scott Nicholas, senior director of strategic planning at The Linux Foundation. “This furthers the goal of LF Deep Learning and the Acumos project of accelerating overall AI innovation.”

The challenge with AI is that there are very few apps to use it, noted Jay Srivatsa, CEO of
Future Wealth.

“Acumos was launched to create an AI marketplace, and the release of Athena is a first step in that direction,” he told LinuxInsider.

The Acumos AI Platform

Acumos packages toolkits such as TensorFlow and SciKit Learn, along with models that have a common API that allows them to connect seamlessly. The AI platform allows for easy onboarding and training of models and tools.

The platform supports a variety of popular software languages, including Java, Python, and R. The R language is a free software environment for statistical computing and graphics.

The Acumos AI Platform leverages modern microservices and containers to package and export production-ready AI applications as Docker files. It includes a federated AI Model Marketplace, which is a catalog of community-distributed AI models that can be shared securely.

LF Deep Learning members contribute to the evolution of the platform to ease the onboarding and the deployment of AI models, according to LF Deep Learning Outreach Committee Chair Jamil Chawki. The Acumos AI Marketplace is open and accessible to anyone who wants to download or contribute models and applications.

“Acumos Athena is a significant release because it enables the interoperability of AI, DL and ML models and prevents the lock-in that usually occurs whenever projects are built using disparate configurations, systems and deployment techniques,” explained Rishi Bhargava, cofounder of

It will ease restrictions on AI, DL and ML developers by removing silos and allowing them to build standardized models, chain each other’s models together, and refine them through an out-of-the-box general AI environment, he told LinuxInsider.

“The efficiency of learning models is hugely contingent on the quality and uniqueness of data, the depth and repeatability of feature engineering, and selecting the best model for the task at hand,” Bhargava said. “Athena will free developers of extraneous burdens so they can focus on these core tasks, learn from each other, and eventually deliver better models to businesses and customers.”

Athena Release Highlights

Athena’s design is packed with features to make the software quick and easy to deploy, and to make it easy to share Acumos AI applications.

Athena can be deployed with one-click using Docker or Kubernetes. The software also can deploy models into a public or private cloud infrastructure, or into a Kubernetes environment on users’ own hardware, including servers and virtual machines.

It utilizes a design studio graphical interface that enables chaining together multiple models, data translation tools, filters and output adapters into a full end-to-end solution. Also at play is a security token to allow simple onboarding of models from an external toolkit directly to an Acumos AI repository.

Models easily can be repurposed for different environments and hardware. This is done by decoupling microservices generation from the model onboarding process.

An advanced user portal allows personalization of marketplace view by theme and data on model authorship. This portal also allows users to share models privately or publicly.

“The LF Deep Learning Foundation is focused on building an ecosystem of AI, deep learning and machine learning projects, and today’s announcement represents a significant milestone toward achieving this vision,” said LF Deep Learning Technical Advisory Council Chair Ofer Hermoni of Amdocs.

Unifying Factor

The Acumos release is significant for the advancement of AI, DL and ML innovation, according to Edgar Radjabli, managing partner of
Apis Capital Management.

The AI industry is very fragmented, with virtually no standardization.

“Companies building technology are usually required to write most from scratch or pay for expensive licensed cloud AI solutions,” Radjabli told LinuxInsider. “Acumos can help bring a base (protocol) layer standard to the industry, in the same way that HTTP did for the Internet and Linux itself did for application development.”

LF Deep Learning members are inspired and energized by the progress of the Acumos AI Project, noted Mazin Gilbert, vice president of advanced technology and systems at AT&T and the governing board chair of LF Deep Learning.

“Athena is the next step in harmonizing the AI community, furthering adoption and accelerating innovation,” he said.

Open Source More Suitable

Given the challenges of growing new technologies, open source models are better suited to the development process than those of commercial software firms. Open source base layer software is ideal. It allows greater adoption and interoperability between diverse projects from established players and startups, said Radjabli.

“I believe that Acumos will be used both by other open source projects building second-layer applications, as well as commercial applications,” he said.

Today, the same situation exists in other software development. Open source base layer protocols are used across the industry, both by other open source/nonprofit projects and commercial operations, he explained.

“Athena clearly is geared to an open source environment, given that it already has about 70 or more contributors,” said Future Wealth’s Srivats.

Benefits for Business and Consumers

The benefits to be gained from AI, DL and ML are very significant. Companies across the industry have been making progress in the development of unique applications for AI/DL/MO. More growth in this space will result from Acumos, according to Radjabli.

One example involves a company that uses neural networks for predictive healthcare analytics. This system allows it to diagnose breast cancer with zero percent false negatives simply from patient data correlation analysis. This does not involve any invasive testing or imaging, according to Radjabli.

“The correlation is comprised of over 40 variables, which means it would have never been found through traditional medical research data analysis and was only made possible through the use of convolutional and recurrent neural networks working in combination,” he said.

AI, DL and ML are all geared toward businesses understanding and predicting consumer behavior, added Srivatsa.

“Both will benefit,” he said.

What’s Next for Acumos AI

The developer community for Acumos AI already is working on the next release. The company expects it to be available in mid-2019.

The next release will introduce convenient model training, as well as data extraction pipelines to make models more flexible.

Additionally, the next release will include updates to assist closed-source model developers, such as secure and reliable licensing components to provide execution control and performance feedback across the community.

Jack M. Germain has been an ECT News Network reporter since 2003. His main areas of focus are enterprise IT, Linux and open source technologies. He has written numerous reviews of Linux distros and other open source software.
Email Jack.

Source link

GNOME 3.31.2 Desktop Released – Phoronix


GNOME 3.31.2 is out this Friday as the latest development release in the trek towards next March’s GNOME 3.32 release.

Highlights for the GNOME 3.31.2 development milestone include:

– The Epiphany web-browser has added preview widgets to its file choosers.

– Support for XPS files within the Flatpak version of the Evince document viewer. Meson is also now the default build system for the Flatpak version of Evince.

– GNOME Boxes virtualization client now sets the default machine type to the Intel Q35 model.

– Crash fixes for the Nautilus file manager.

– Sushi has been ported to the Meson build system.

– Various application icons were updated.

– Performance work and fixes for GNOME Shell and Mutter.

More details on the GNOME 3.31.2 development release via the mailing list announcement.

Red Hat Reports $823 Million in Revenue for Sec… » Linux Magazine

Red Hat has evolved beyond its original role as a Linux vendor and is now positioned as a cloud player that offers complete solutions to enterprise customers. The company has been expanding its product portfolio to help customers embark on their cloud native and digital transformation journey.

Red Hat’s aggressive repositioning is reflected in its revenue. The company earned the total revenue of $823 million, up 14% year-over-year, in the second quarter of the fiscal year 2019.

“The expansion of our technology portfolio has increased our strategic importance with customers, which is evidenced by the number of deals over five million dollars in the second quarter more than doubling year-over-year,” said Jim Whitehurst, President and Chief Executive Officer of Red Hat. “Customers continue to prioritize their digital transformation initiatives, and they are adopting Red Hat’s hybrid cloud enabling technologies to modernize their applications and drive greater efficiency and effectiveness in their business.”

Which technologies segments are growing within Red Hat is apparent from the breakout of the revenue. Subscription revenue remains the largest earnings at $527 million, but it registered a mere 8% in year-over-year growth. On the other hand, revenue from emerging technologies (read cloud and containers) was $196 million for over 31% year-over-year growth.

If Red Hat keeps up this pace, it might touch the $4 billion annual revenue mark in 2019.

Source link

Linux 4.20 Showing Some Performance Slowdowns


Being well past the Linux 4.20 merge window I have moved onto benchmarking more of this development version of the Linux kernel. Unfortunately, there are some clear performance regressions.

This week I got to firing off some Linux 4.20 kernel benchmarks… I started with the AMD Ryzen Threadripper 2990WX and Intel Core i9 7980XE for being the interesting HEDT CPUs in my possession at the moment. On the 7980XE I spotted several performance regressions with this Linux 4.20 development kernel compared to Linux 4.19 and 4.18, so then I fired up the completely separate Intel Core i9 7960X box to carry out the same tests. Sure enough, with that different hardware, there is further confirmation of slowdowns with Linux 4.20.

The common trait of these systems was Ubuntu 18.10 x86_64 and using the Linux 4.18.18, 4.19.1, and 4.20 Git kernel packages provided by the Ubuntu Mainline Kernel PPA. With the differing hardware the intention is not to compare the performance between the systems but in looking at the direction of the Linux kernel performance.

Given the measurable slowdowns with Linux 4.20 in several workloads and being able to reproduce it on the 7960X box, I’ll be testing a more diverse selection of hardware over the days ahead to see if it’s an Intel specific issue (perhaps a change in their P-State code or other vendor-specific code paths) or what to more specifically isolate the regression. If it comes to it, thanks to the Phoronix Test Suite, it can be automatically bisected as well — especially with regressing on these high core count systems where kernel builds are a breeze.

As one of the few plus notes, the Vega GPU on the 2990WX box is indicating better Vulkan performance thanks to AMDGPU DRM changes at least for the synthetic VKMark benchmark…

Rodinia is where the Core i9 performance started drawing concern…

And then moving to the Java-based DaCapo benchmarks raised more concern with clear performance setbacks when using the current Linux 4.20 Git kernel while the AMD Threadripper performance was unchanged.

Code compilation speeds also are slower now under the Linux 4.20 kernel on the Intel HEDT boxes.

The PostgreSQL performance was lower across the board.

The synthetic Stress-NG kernel benchmarks were at least showing an improvement in socket activity performance across the three tested systems thus far.

Context switching performance was also improved.

Blender was yet another multi-threading workload slower on the Intel CPUs tested.

With at least some workloads, on the Core i9 7960X and Core i9 7980XE were clear performance regressions using the Linux 4.20 kernel. At least from the initial benchmarks run on Linux 4.20, most of the regressions appear to be with multi-threaded workloads. I’ll be firing up Linux 4.20 tests on more hardware over the days ahead and bisecting if the issue isn’t resolved upstream sooner than that. But so far from trying on two completely different Core i9 systems, there are some clear slowdowns.