Programming Systems for Analog Hardware
summary
Presents a concrete compiler that maps dynamical systems to a real CMOS analog differential-equation chip, producing a first-ever compilation toolchain that targets hcdc V2 and runs in microjoules of energy with per-device power in the micro- to milliwatt range. The flow synthesizes circuits, then applies a geometric programming scaling pass to meet port ranges (±2), ADC/DAC limits (±1), noise (~1–4% stddev) and process gains (~0.97–1.01), and showcases results on 12 benchmarks, 200 circuits each with much better fidelity after a co-designed calibration change.
AI-written summary · about
chapters (9)
- 0:00Intro
- 1:06A different computing paradigm
- 7:22Analog Computing
- 11:44What can you do with dynamical systems?
- 13:05How do you program analog computers?
- 16:08Challenges with programming analog hardware
- 26:02Matching Dynamics
- 31:54Running the Analog Device Configuration
- 38:33Applying the Scaling Transform
related talks
16mBuild your own Open Hardware CPU in 25 minutes or less
01h 02mTurning software into computer chips – Hastlayer
56m“Towards Heterogeneous Programming in C++"
01h 09mC++20 on Xilinx FPGA with SYCL for Vitis
01h 01mTake Advantage of All the MIPS - SYCL & C++
45mAutomated, Simulation-Based Flow for Low-Cost FPGA-Accelerated Devices with Zeph...
37mC++ for multi-accelerator, multi-vendor systems
32mRunning Zephyr RTOS on Cadence® Tensilica® HiFi 4 DSP