IT Collaboration Leads to Unique Product Innovation

Intel IT is taking the concept of “IT shop” to a new level through our collaboration with Intel’s business units. We’ve previously written about how IT collaboration has enabled our product testing teams to build artificial intelligence (AI) into Intel’s validation processes, saving significant time and money. Recently, we pushed the envelope of IT collaboration by working closely with Intel’s Client Computing Group (CCG) to integrate AI into Intel products.

CCG defines, designs, and tests new generations of Intel® Core™ processors (as well as other Intel products). CCG engineers recently asked an intriguing question: Might it be possible to use AI and machine learning to dynamically adjust a processor’s power limits, based on the actual workload? Generally, changes to these power limits are based on fixed rules at runtime and machine learning algorithms are not involved. But if adjusting them on the fly were possible, in light of the actual workload currently running on the system (see the figure below), processor performance could theoretically be enhanced. Having heard of our previous collaboration successes, CCG sought help from Intel IT’s Advanced Analytics team to explore this idea. The end result? On select designs of an upcoming Intel Core processor generation, Intel® Dynamic Tuning Technology (Intel® DTT) will offer the first ever AI-based pre-trained algorithms to predict workloads, allow higher turbo burst when responsiveness is needed, and allow extended time in turbo for sustained workloads.

Successful Collaboration Is a Learning Process

CCG’s overall goal was to define a problem statement and what success would look like. Then the CCG and IT teams defined the experiments. It was a true collaboration, where CCG understood AI concepts, and the IT team understood CCG’s language, constraints, and requirements. By joining forces, we multiplied the value of each team’s expertise.

From ideation to a successful unit took about one year. Along the way, we discovered some significant challenges. One was making sure our experiments used data that represented real market data. Another major challenge was benchmarking and meeting strict approval requirements. When introducing AI into a product, you are obligated to meet the extremely strict existing quality. When working on an actual product (as opposed to an internal process like our previous work with test validation), it is imperative to ensure that you not only “do good” but that you also “do no harm” to the user experience (UX). Our experiments and models had to be designed to identify where the AI solution enhanced the UX and where it was a detriment—and to find solutions to the problem areas so that the overall UX is either much better or the same as without the enhancement.

The Future Is Bright for IT-Business Unit Collaboration at Intel

Many use cases will build on the success of this one. We proved that we could successfully introduce machine learning into an Intel product, and we earned the status of trusted collaborator with CCG for future machine learning projects. We plan to develop ideas for future enhancement of this technology and other use cases for introducing AI into Intel’s product line. In addition, with the optimized development and collaboration processes, we can now ideate with greater velocity. With three follow-on solutions in process with CCG, as well as several projects with design groups, our momentum is strong. Although these ideas will not all mature to production, the velocity of incoming ideas is staggering, and the business units are beginning to see IT as a valued and trusted partner to deliver solutions within process and time limits.

An important way for Intel to stay competitive is to continue to improve product performance; but as silicon complexity increases, it becomes more difficult to make significant performance gains. Therefore, any good idea is critical for Intel’s success. Intel IT is excited to continue working with Intel’s business units and design teams to put these ideas to work—both for Intel’s customers and for our own Intel product consumption as we provision the data centers and employee devices.

Learn more about what Intel IT is doing with AI at www.intel.com/IT.

Published on Categories Artificial IntelligenceTags , , , ,
Nufar Gaspar

About Nufar Gaspar

Nufar Gaspar joined Intel in 2012 after completing her Masters in Industrial Engineering in Ben Guryon University, with focus on optimization, statistics, and scheduling problems. Publications from her Thesis can be found at https://dblp.uni-trier.de/pers/hd/g/Gasper:Nufar. During 2012-2015, Nufar played different roles in the creation of Machine Learning solutions for Intel’s Design organizations. Since 2015, Nufar leads the Design Advanced Analytics team, which includes Data Scientists, Big Data SW Developers, Products Analysts, and Product Managers. Nufar and her team creates AI and Big Data tools to help revolutionizing how Intel designs and verifies its product. These capabilities include CLIFF (AI based test creation to uncover hidden bugs), ITEM (Intelligent Test Execution Mgmt. to optimize the test suit being used for validation), Gatekeeper Smart Filter (predict if a code submitted to GIT is likely to be buggy) and many others. These days Nufar and her team run multiple R&D effort utilizing state-of-the-art AI and optimization techniques as well as scaling the existing capabilities across the various design organizations at Intel. Examples include: Adaptive Testing: online testing manager agent that tunes its decisions in light of status vs. user-defined goals, and Aided debug: a collection of debug utilities that interact with human expert to speed up root cause analysis.