Solve SoC Bottlenecks with Smart Local Memory in AI/ML Subsystems
In today’s disaggregated electronics supply chain the (1) application software developer, (2) the ML model developer, (3) the device maker, (4) the SoC design team and (5) the NPU IP vendor often work for as many as five different companies. It can be difficult or impossible for the SoC team to know or predict actual AI/ML workloads and full system behaviors as many as two or three years in advance of the actual deployment. But then how can that SoC team make good choices provisioning compute engines and adequate memory resources for the unknown future without defaulting to “Max TOPS / Min Area”?
There has to be a smarter way to eliminate bottlenecks while determining the optimum local memory for AI/ML subsystems.
To read the full article, click here
Related Semiconductor IP
- eUSB2V2.0 Controller + PHY IP
- I/O Library with LVDS in SkyWater 90nm
- 50G PON LDPC Encoder/Decoder
- UALink Controller
- RISC-V Debug & Trace IP
Related Blogs
- High Speed Memory in Smart Phones: MIPI UniPro v1.8 for JEDEC UFS v3.0
- Boost SoC Flexibility: 4 Design Tips for Memory Subsystems with Combo DDR3/4 Interfaces
- Faster Embedded Smartphone & Tablet Memory Is On The Way
- Building Smarter, Faster: How Arm Compute Subsystems Accelerate the Future of Chip Design
Latest Blogs
- Deploying StrongSwan on an Embedded FPGA Platform, IPsec/IKEv2 on Arty Z7 with PetaLinux and PQC
- The Rise of Physical AI: When Intelligence Enters the Real World
- Can Open-Source ISAs Catalyze Smart Manufacturing?
- The Future of AI is Modular: Why the SiFive-NVIDIA Milestone Matters
- Heterogeneous Multicore using Cadence IP