Discover ComputeRAM™
What is ComputeRAM?
- Ground-breaking performance: A ComputeRAM-enhanced embedded chip delivers 100 times better energy efficiency and latency, making it ideal for low-latency, energy-constrained applications.
- Seamless integration: It is fully CMOS, requires no foundry waivers and is compatible with any CMOS process and instruction set architecture (ARM, RISC-V, x86). ComputeRAM™ Software Development Kit (SDK) integrates with popular embedded development workflows.
- Developer-friendly: SDK includes optimized implementations of linear algebra subroutines, signal processing functions, and neural network layers. It can be easily extended to support new functions, making it an ideal solution for the evolving needs of embedded products.
A Minimal Matrix-Vector Product Benchmark
At the most basic level, ComputeRAM™ enables embedded microcontrollers to perform fast and efficient matrix-vector products. By offloading up to 99% of model operations to memory, reducing bus activity, and leveraging an advanced in-memory computing engine, an Arm Cortex-M0 enhanced with ComputeRAM™ delivers up to 139x faster processing and 158x better energy efficiency than an Arm Cortex-M0 with conventional SRAM memory.
Read more in our MVM Benchmarking article.
A New Era of Embedded AI Applications
Achieves up to 30x faster processing and 32x better energy efficiency on the MLPerf™ Tiny benchmarks compared to the best submission in the MLCommons database. ComputeRAM revolutionizes TinyML, eliminating the need for dedicated AI accelerators in wearable, robotics, and smart sensing applications, thereby lowering design complexity, time to market, and silicon cost.
Read more in our TinyML Benchmarking article.
Up to
30x
Faster
Up to
32x
More Energy Efficient
ComputeRAM™ Product Brief
The document presents a comprehensive overview of ComputeRAM™, highlighting its advanced in-memory computing capabilities, exceptional performance improvements, and broad applicability across various industries.