sponsored

Intel Xeon Scalable Processors for the Visual Cloud

New breakthroughs in performance and TCO

istock 598674382 1619x1080 blak
Intel

For years, service providers have capitalized on the advantages of industry-standard servers for media processing in the cloud. These industry-standard servers, typically optimized for Intel architecture, offer the versatility to handle diverse visual cloud workloads and the total cost of ownership (TCO) advantages of off-the-shelf Intel-based hardware.

Today, we are building on this foundation for media processing in the cloud with new breakthroughs in performance and TCO made possible by the launch of the new Intel® Xeon® Scalable processors. The successor to the Intel® Xeon® processor E5 and E7 families, the Intel Xeon Scalable processors deliver up to 28 cores, significant increases in memory and I/O bandwidth, and built-in accelerators like Intel® AVX-512, along with six DDR4 memory channels and 48 PCIe lanes. These characteristics make Intel Xeon Scalable processors a great choice for accelerating media processing in the cloud such as software-based transcoding and video stitching.

Already, some early users are realizing the benefits of the Intel Xeon Scalable processors in the visual cloud space. One of these companies is Ericsson. With Intel Xeon Scalable processors, Ericsson achieved a 1.5x performance increase for its Ericsson MediaFirst Video Processing Encoding Live software transcoder, compared to the previous-generation Intel Xeon platform.[1] This jump in performance — going from 47 frames per second (fps) to 71 fps — gives Ericsson the ability to deliver 10-bit HEVC high definition video in real time on a single server.

The keys to these performance gains in the transcoding space include the increased core count, higher frequencies, greater memory bandwidth, and on-chip performance boosters like Intel® AVX-512 or Intel® QuickAssist technology available with the Intel Xeon Scalable processors. Transcoding applications also benefit from the platform’s improved cache management of the processor.

Tencent, meanwhile, realized a similar jump in performance for its video stitching application, which allows the company’s customers to create immersive video content for virtual reality, 360 videos, and other applications. With Intel Xeon Scalable processors powering its video stitching application, Tencent reduced stitching latency by 72 percent and increased video stitching performance by more than 1.7x, compared to the previous-generation Intel Xeon platform.[2] This jump in performance enables a seamless, smoother, and more immersive 360-degree video experience for the user.

Tencent found that more cores and high memory bandwidth were key drivers of the overall performance gain for the video stitching application. The company is also benefitting from software optimizations that utilize Intel® Advanced Vector Extensions 512 (Intel® AVX-512). These optimizations enabled a 34 percent performance improvement over the use of Intel® AVX2 instructions, for a greatly enhanced stitching function.

Results like these illustrate the significant benefits that Intel Xeon Scalable processors bring to media processing. As I noted last year in a blog on the Visual Cloud Second Wave, Cloud computing is a major factor driving the growth of the visual cloud. Today, the arrival of Intel Xeon Scalable processors represents the next big step in the improvement of cloud computing, and we are already seeing the benefits to the visual cloud.

To dive down into the features and capabilities of the new Intel Xeon Scalable processors, visit intel.com/xeonscalable.

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations, and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For complete information visit http://www.intel.com/performance. *Other names and brands may be claimed as the property of others.

[1] Testing conducted on Ericsson* software comparing Intel® Xeon® Platinum 8168 processor to 2S Intel® Xeon® Processor E5-2699 v4. Ericsson MediaFirst Video Processing UHD HEVC transcoding workload. OS: CentOS Linux* 7.2 kernel 3.10.0. Testing by Ericsson in May 2017. BASELINE: 2S Intel® Xeon® processor E5-2699 v4, 2.2GHz, 22 cores, Turbo and HT on, BIOS 251.R01, 64GB total memory, 8 slots / 8GB / 2133 MT/s / DDR4 LRDIMM, CentOS Linux* 7.2 kernel 3.10.0.

NEW: Intel® Xeon® Platinum processor 8168, 2.7GHz, 24 cores, turbo and HT on, BIOS 412, 192GB total memory, 12 slots / 16GB / 2600 MT/s / DDR4 LRDIMM, CentOS Linux* 7.2 kernel 3.10.0.

[2] Testing conducted on Tencent* software comparing Intel® Xeon® Platinum 8180 processor to 2S Intel® Xeon® Processor E5-2699 v4. Configuration details: Tencent Business Analytics*: Video Stitching workload. OS: CentOS  7.3.1611 Linux kernel 4.9.8. Testing by Intel April 2017. BASELINE: 2S Intel® Xeon® processor E5-2699 v4, 2.2GHz, 22 cores, turbo and HT on, 128GB total memory, 8 slots / 16GB / 2400 MT/s / DDR4, SATA3 4TB. NEW: Intel® Xeon® Platinum processor 8180, 2.5GHz, 28 cores, turbo and HT on, 192GB total memory, 12 slots / 16GB / 2666 MT/s / DDR4, s3700 800G SSD.

Related: