TDC2019 Intel Software Day - Tecnicas de Programacao Paralela em Machine Learning

The document discusses a presentation given by Seth Schneider from Intel and Russ Glaeser from Cascade Game Foundry. It introduces Intel's Graphics Performance Analyzers (GPA) tool and demonstrates how it was used to optimize the game Infinite Scuba developed by Cascade Game Foundry. The presentation covered an overview of GPA, details about Infinite Scuba, and a live demo of using GPA to analyze and improve performance of the game.

Review state-of-the-art techniques that use neural networks to synthesize motion, such as mode-adaptive neural network and phase-functioned neural networks. See how next-generation CPUs with reinforcement learning can offer better performance.

Software AI Accelerators deliver orders of magnitude performance gain for AI across deep learning, classical machine learning, and graph analytics and are key to enabling AI Everywhere. Get started on your AI Developer Journey @ software.intel.com/ai.

The document describes Intel Graphics Performance Analyzers (Intel GPA), a free tool that allows users to optimize game performance on Windows, Android, and Ubuntu systems. Intel GPA includes tools like the System Analyzer for real-time in-game performance analysis, the Frame Analyzer for detailed frame-level analysis, and the Platform Analyzer to visualize CPU and GPU activity. It also allows experiments like changing graphics settings without code modifications to help identify performance bottlenecks.

Explore practical elements, such as performance profiling, debugging, and porting advice. Get an overview of advanced programming topics, like common design patterns, SIMD lane interoperability, data conversions, and more.

QuEST Global is a global engineering company that provides AI and digital transformation services using technologies like computer vision, machine learning, and deep learning. It has developed several AI solutions using Intel technologies like OpenVINO that provide accelerated inferencing on Intel CPUs. Some examples include a lung nodule detection solution to help detect early-stage lung cancer from CT scans and a vision analytics platform used for applications in retail, banking, and surveillance. The company leverages Intel's AI Builder program and ecosystem to develop, integrate, and deploy AI solutions globally.

This document discusses optimizations and new DirectX features for Intel graphics hardware. It begins with an introduction of Avalanche Studios, the developer of the game Just Cause 3. It then discusses the use of Intel's Graphics Performance Analyzers tools to analyze Just Cause 3 and identify optimization opportunities. The document outlines several low-level shader optimizations performed, including reworking math operations, rearranging variables, and reusing intermediate values. It also discusses leveraging new DirectX features pioneered by Intel. The goal of these optimizations is to improve performance for the large install base of gamers using Intel graphics.

See how Intel® Processor Graphics can accelerate machine learning and AI workloads to solve complex problems that were previously very difficult.

oneDNN Graph API extends oneDNN with a graph interface which reduces deep learning integration costs and maximizes compute efficiency across a variety of AI hardware including AI accelerators. Get started on your AI Developer Journey @ software.intel.com/ai.

Overview of the new Embree 3 ray tracing framework, including how to use the new API, supported geometry types, and ray intersection methods. Includes a look at new features like normal oriented curves, vertex grids, etc.

Learn how to write fast, efficient, and maintainable vector code for the CPU with examples of simple SIMD that use the Intel® ISPC.

With the growth of AI, machine learning, and data-centric applications, the industry needs a programming model that allows developers to take advantage of rapid innovation in processor architectures. TensorFlow supports the oneAPI industry initiative and its standards-based open specification. oneAPI complements TensorFlow’s modular design and provides increased choice of hardware vendor and processor architecture, and faster support of next-generation accelerators. TensorFlow uses oneAPI today on Xeon processors and we look forward to using oneAPI to run on future Intel architectures.

This document discusses Bodo Inc.'s product that aims to simplify and accelerate data science workflows. It highlights common problems in data science like complex and slow analytics, segregated development and production environments, and unused data. Bodo provides a unified development and production environment where the same code can run at any scale with automatic parallelization. It integrates an analytics engine and HPC architecture to optimize Python code for performance. Bodo is presented as offering more productive, accurate and cost-effective data science compared to traditional approaches.

Jeff Rous from Intel and Niklas Smedberg from Epic Games discussed optimizing the Unreal Engine 4 (UE4) game engine for Intel processors. They described measuring performance using Intel's Graphics Performance Analyzers, common pain points like memory bandwidth and dense geometry on Intel graphics, and shader optimizations. The presentation also covered optimizing UE4 for DirectX 12, adding support for Android x86/x64, and announcing fast ASTC texture compression support in UE4.

Integrated into Intel® Advisor, Cache-aware Roofline Modeling (CARM) provides insight into how an application behaves by helping to determine a) how optimally it works on a given hardware, b) the main factors that limit performance, c) if the workload is memory or compute-bound, and d) the right strategy to improve application performance.

Open Image Denoise is an open source library for denoising images rendered with ray tracing. It provides a deep learning based denoising filter that can run on any modern Intel CPU. The filter uses a convolutional neural network architecture and has been shown to improve image quality over other filters while maintaining interactive performance. The API is designed to be simple and easy to integrate into rendering applications. Future versions will include additional features like temporal coherence and support for more input buffers.

This talk focuses on the newest release in RenderMan* 22.5 and its adoption at Pixar Animation Studios* for rendering future movies. With native support for Intel® Advanced Vector Extensions, Intel® Advanced Vector Extensions 2, and Intel® Advanced Vector Extensions 512, it includes enhanced library features, debugging support, and an extensive test framework.

作者：王宗業 【視覺進化論】AI智慧視覺運算技術論壇 ►活動日期：2018/9/26(三) 13:30-16:30 ►活動網址：https://makerpro.cc/events/180926_smart_vision/ ►聯繫Sertek：www.sertek.com.tw / 886-2-2696-0055

This session was held by Vladimir Brenner, Partner Account Manager, Disruptors & AI, Intel AI at the Dive into H2O: London training on June 17, 2019. Please find the recording here: https://youtu.be/60o3eyG5OLM

In this talk, Tong will start with the current landscape and typical use cases of Artificial Intelligence applications in the Telco domain. Then, she will introduce Intel’s strategy and products for Network AI, including our focus areas, our hardware portfolio, software stacks, roadmaps and some case studies. Speaker: Tong Zhang, Principal Engineer and Chief Architect for AI and Analytics of the Network Platforms Group, Intel

This document contains several legal notices and disclaimers from Intel regarding their products. No license is granted to any intellectual property and Intel assumes no liability relating to the sale and use of their products. Intel products are not intended for medical or life critical applications. Specifications and descriptions are subject to change without notice.

Preprocess, visualize, and Build AI Faster at-Scale on Intel Architecture. Develop end-to-end AI pipelines for inferencing including data ingestion, preprocessing, and model inferencing with tabular, NLP, RecSys, video and image using Intel oneAPI AI Analytics Toolkit and other optimized libraries. Build at-scale performant pipelines with Databricks and end-to-end Xeon optimizations. Learn how to visualize with the OmniSci Immerse Platform and experience a live demonstration of the Intel Distribution of Modin and OmniSci.

This session presents performance data for deep learning training for image recognition that achieves greater than 24 times speedup performance with a single Intel® Xeon Phi™ processor 7250 when compared to Caffe*. In addition, we present performance data that shows training time is further reduced by 40 times the speedup with a 128-node Intel® Xeon Phi™ processor cluster over Intel® Omni-Path Architecture (Intel® OPA).

Out of the box usability, scaling for HPC and big data, optimized for Intel processors. If you like what you read be sure you ♥ it below. Thank you!

This document discusses scaling Python performance in production environments. It introduces the Intel Distribution for Python, which provides optimized versions of NumPy, SciPy, and Scikit-Learn using Intel MKL to accelerate linear algebra and machine learning algorithms. It also supports parallelism through MPI, TBB for multithreading, and integration with big data frameworks. Profiling tools like Intel VTune Amplifier help optimize mixed-language Python applications for Intel architectures. The goal is to make Python usable for high performance computing and big data workloads while maintaining its ease of use.

Universal Scene Description* (USD) is an open source initiative developed by Pixar for fast, large scale, and universal asset management across multiple programs including Maya, Houdini, and others.

1. The document introduces the Intel Xeon Scalable platform, which provides the foundation for data center innovation with a 1.65x average performance boost over previous generations. 2. It highlights key advantages of the platform including scalable performance, agility in rapid service delivery, and hardware-enhanced security with near-zero performance overhead. 3. Various workload-optimized solutions are discussed that leverage the platform's performance to accelerate insights from analytics, deploy cloud infrastructure more quickly, and transform networks.

The capacity of data grows rapidly in big data area, more and more memory are consumed either in the computation or holding the intermediate data for analytic jobs. For those memory intensive workloads, end-point users have to scale out the computation cluster or extend memory with storage like HDD or SSD to meet the requirement of computing tasks. For scaling out the cluster, the extra cost from cluster management, operation and maintenance will increase the total cost if the extra CPU resources are not fully utilized. To address the shortcoming above, Intel Optane DC persistent memory (Optane DCPM) breaks the traditional memory/storage hierarchy and scale up the computing server with higher capacity persistent memory. Also it brings higher bandwidth & lower latency than storage like SSD or HDD. And Apache Spark is widely used in the analytics like SQL and Machine Learning on the cloud environment. For cloud environment, low performance of remote data access is typical a stop gap for users especially for some I/O intensive queries. For the ML workload, it's an iterative model which I/O bandwidth is the key to the end-2-end performance. In this talk, we will introduce how to accelerate Spark SQL with OAP (https://github.com/Intel-bigdata/OAP) to accelerate SQL performance on Cloud to archive 8X performance gain and RDD cache to improve K-means performance with 2.5X performance gain leveraging Intel Optane DCPM. Also we will have a deep dive how Optane DCPM for these performance gains. Speakers: Cheng Xu, Piotr Balcer

Apache CarbonData & Spark meetup "QATCodec: past, present and future" if from INTEL Apache Spark™ is a unified analytics engine for large-scale data processing. CarbonData is a high-performance data solution that supports various data analytic scenarios, including BI analysis, ad-hoc SQL query, fast filter lookup on detail record, streaming analytics, and so on. CarbonData has been deployed in many enterprise production environments, in one of the largest scenario it supports queries on single table with 3PB data (more than 5 trillion records) with response time less than 3 seconds!

This document discusses optimizing deep learning inference on Intel processor graphics using the OpenVINOTM toolkit. Some key points include: - Running inference on client devices provides advantages over cloud like privacy, bandwidth savings, and responsiveness. - OpenVINOTM provides tools to optimize models for Intel hardware and achieve 5-10x speedups on Intel GPUs compared to CPU baselines. - A case study demonstrates optimizing a deep image matting model, reducing inference time from 2.35 seconds to 291 milliseconds on Intel GPU using OpenVINOTM. - Emerging technologies like federated learning are discussed which could improve privacy for on-device inference.