GDDR6 Beyond Graphics : Memory for AI,VR, and Autonomous Driving
Modern computer applications rely heavily on graphics processing and rendering which involve a lot of simultaneous mathematical calculations. A typical CPU is not suitable for jobs that require simultaneous processing, which is why the concept of a dedicated Graphics Processing Unit (GPU) was introduced. The GPU has found its scope not only in graphics processing but also several emerging applications like AI, machine learning, VR, autonomous driving, and network routing.
GPU’s require memory which can offer much higher throughput than conventional memories like DDR, since it processes massive chunks of data all at once. The memory must also be capable of providing minimal latency, along with the possibility of simultaneous write/read. As a result, Graphics Double Data Rate (GDDR) memory, a dedicated type of SGRAM for the GPU, came into the picture.
To read the full article, click here
Related Semiconductor IP
- Ultra-Low-Power LPDDR3/LPDDR2/DDR3L Combo Subsystem
- Parameterizable compact BCH codec
- 1G BASE-T Ethernet Verification IP
- Network-on-Chip (NoC)
- Microsecond Channel (MSC/MSC-Plus) Controller
Related Blogs
- Autonomous Vehicles: Memory Requirements & Deep Neural Net Limitations
- What’s on the Horizon for NAND and DRAM?
- DDR3/DDR2 price crossover reached
- Apple iPad: no LPDDR2?
Latest Blogs
- What Does a GPU Have to Do With Automotive Security?
- Physical AI at the Edge: A New Chapter in Device Intelligence
- Rivian’s autonomy breakthrough built with Arm: the compute foundation for the rise of physical AI
- AV1 Image File Format Specification Gets an Upgrade with AVIF v1.2.0
- Industry’s First End-to-End eUSB2V2 Demo for Edge AI and AI PCs at CES