A new golden age for computer architecture by Prof. David Patterson

In last 50 years, there are 3 lessons that we can draw. First – software advances can inspire architecture innovations. Second – when we raise the hardware/software interface, it creates opportunities for architecture innovation. Third – in our field, the way we settle these debates, isn’t by just arguing in a bar, rather people spent/invest billions of dollars to investigate their ideas and marketplace settles these debates

Continue reading

Coverage Driven Functional Verification on RISC-V Cores

Design Verification is critical to proving functional correctness and establishing confidence in a design. Several studies from industry and academia, particularly over the course of the last two decades, have explored various verification methodologies that fall somewhere between dynamic or purely static formal approaches.

Random stimulus methods perform better because they eventually cover many cases. Most new ideas in dynamic verification over the last two decades have largely been towards semi formal verification methodologies such as coverage driven verification and constrained test generation. In this paper, we explore an approach to dynamic functional verification that we use at the RISE lab, IIT Madras for the verification of the RISC-V based Shakti cores.

Continue reading

Rapid Physical IC Implementation

“Rapid Physical IC Implementation and Integration using Efabless Platform”. This paper describe a rapid backend process flow (synthesis, placement, STA, routing) and top level integration to implement a small RTL IP into a tapeout ready chip using the Efabless online platform

Continue reading

Symposium VI – Standard cell layout/characterization

Symposium VI – Standard cell layout/characterization, ECSM puts its number in the same arc as NLDM. The numbers you see in above image, below the cell_rise, cell_fall, rise_transition is all NLDM information. Under rise_transition, you will have ecsm_waveform and ecsm_capacitance. Now this is only one waveform, because we gave it only one load and one slope, just like we have one value under “rise_transition”. If we had 3×3 under rise_transition, then you would have had ecsm_waveform(“0”), ecsm_waveform(“1”), till ecsm_waveform(“8”), essentially 9 waveforms. And same thing with capacitance

Continue reading

Symposium V – Machine Intelligence in EDA/CAD applications

Symposium V – Machine Intelligence in EDA/CAD applications- Let’s investigate a simple Wire Resistance Estimate (WiRE) model
This is common design automation problem which is used for estimating timing and power characteristics for analysis and implementation for many steps in ASIC flow. We will restrict our scope to physical implementation only, where known quantity is “length” of wire and resistance is predicted.

Continue reading

Symposium I – Front-end open-source EDA tool flows for IC design and verification

Question – Who doesn’t want a 3.5X improvement in their code size? I guess everyone wants efficient and effective improvement. Now these are just few tips to have the easy implementation of pipe-line. You are free to implement your ideas in TL-verilog, compile, simulate and see the improvements on your own. For few more tips, you might want to check out below course on “VSD – Pipelining RISC-
V with TL-verilog”

Continue reading