Multicore processor integrates neural processing unit -

Multicore processor integrates neural processing unit


NXP- i.MX-8M-Plus-application-processor-smallNXP Semiconductors N.V. expands its EdgeVerse portfolio with the launch of its i.MX 8M Plus application processor. This is the first device in the i.MX family to integrate a dedicated neural processing unit (NPU) for advanced machine learning inference at the industrial and IoT edge. It also packages an independent real-time subsystem, dual camera ISP, a high-performance DSP, and 3D GPU for edge applications.

The i.MX 8M Plus combines a NPU delivering 2.3 tera operations per second (TOPS) with a Quad-core Arm Cortex-A53 sub-system running at up to 2 GHz, an independent real-time sub-system with an 800-MHz Cortex-M7, a 800-MHz audio DSP for voice and natural language processing, dual camera image signal processors (ISPs), and a 3D GPU for rich graphics rendering.

The NXP i.MX 8M Plus can execute multiple, highly-complex neural networks simultaneously, such as multi-object identification, speech recognition of 40,000+ English words, and medical imaging. Developers can off-load machine learning inference functions to the NPU, said NXP, allowing the high-performance Cortex-A and Cortex-M cores, DSP, and GPUs to execute other system-level or user applications tasks.

click for larger image

i.MX 8M Plus architecture combines multiple cores with a neural processing unit for machine learning acceleration. (Source: NXP)

The dual integrated ISPs support two high-definition cameras for real-time stereo vision or a single 12-megapixel (MP) resolution camera and includes high dynamic range (HDR) and fisheye lens correction. These features enable real-time image processing applications such as surveillance, smart retail applications, robot vision, and home health monitors.

For voice applications, the i.MX 8M Plus integrates a high-performance HiFi 4 DSP that enhances natural language processing with pre- and post-processing of voice streams. The Cortex-M7 domain can be used to run real-time response systems, said NXP, while the applications processor domain executes complex non-real-time applications to reduce the overall system-level power consumption. It performs this task by turning off the application processor domain while keeping only the Cortex-M domain alive for wake word detection.

In addition, for multimedia and video processing, the i.MX 8M Plus can compress multiple video feeds using the H.265 or H.264 HD video encoder and decoder for cloud streaming or local storage and a rich user experience, which is enabled by 3D/2D graphics, and Immersiv3D audio with Dolby Atmos and DTS:X.

The i.MX 8M Plus also is well suited for industrial IoT applications such as industrial productivity and automation with machines that can inspect, measure, precisely identify objects, and enable predictive maintenance by accurately detecting anomalies in machine operation. The chip supports Industry 4.0 IT/OT convergence, integrating Gigabit Ethernet with Time-Sensitive Networking (TSN).

Other features that make the i.MX 8M Plus industrial-ready are error correction code (ECC) for internal memories and the DDR interface, an industrial temperature range of -40°C to 105°C, and a power-on-profile (100 percent power on). It also will be part of NXP’s industry-best longevity commitment (15 years).

Samples of the i.MX 8M Plus processor will be available in Q1 2020.

>> An earlier version of this article was originally published on our sister site, Electronic Products.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.