---
title: "kedar joshi (@kedar-joshi) | Hashnode"
description: "Physics enthusiast"
url: "https://hashnode.com/@kedar-joshi"
---

@kedar-joshi

# kedar joshi

@kedar-joshi · San Francisco, USA · Joined September 2026

### [Profiling Add and Load Instructions on Dual-Issue Cortex-M7 Core](https://kedar-joshi.hashnode.dev/profiling-add-and-load-instructions-on-dual-issue-cortex-m7-core?utm_source=hashnode&utm_medium=feed)

I was working on benchmarking perf for CRC32 calculation. I started looking into the instructions that the compiler generated and tried to explain the measured cycles for the various methods to calcul

8 min read · [kedar-joshi.hashnode.dev](https://kedar-joshi.hashnode.dev/profiling-add-and-load-instructions-on-dual-issue-cortex-m7-core?utm_source=hashnode&utm_medium=feed)

[1d ago](https://hashnode.com/posts/profiling-add-and-load-instructions-on-dual-issue-cortex-m7-core/6af336b76a3cdaafcd9dc8fd)

### [Closing the loop: performance benchmarking a control problem on ARM Cortex-M7](https://kedar-joshi.hashnode.dev/closing-the-loop-performance-benchmarking-a-control-problem-on-arm-cortex-m7?utm_source=hashnode&utm_medium=feed)

Picking up where I left off My last post benchmarked a simulated buck-converter control loop on a Nucleo-F767ZI. I talked about memory placement, killing off divisions, clock speed, I/D-cache. Last be

9 min read · [kedar-joshi.hashnode.dev](https://kedar-joshi.hashnode.dev/closing-the-loop-performance-benchmarking-a-control-problem-on-arm-cortex-m7?utm_source=hashnode&utm_medium=feed)

[Sep 11](https://hashnode.com/posts/closing-the-loop-performance-benchmarking-a-control-problem-on-arm-cortex-m7/6aa41907731f060c4d2aeabe)

### [Benchmarking a Buck-Converter Loop on Bare-Metal ARM Cortex-M7](https://kedar-joshi.hashnode.dev/benchmarking-a-buck-converter-loop-on-bare-metal-arm-cortex-m7?utm_source=hashnode&utm_medium=feed)

TL;DR: I set out to understand how fast a simple control loop can generate duty-cycle values to feed to a MOSFET gate driver. I benchmarked how placement of data in memory/cache, clock frequency, and

14 min read · [kedar-joshi.hashnode.dev](https://kedar-joshi.hashnode.dev/benchmarking-a-buck-converter-loop-on-bare-metal-arm-cortex-m7?utm_source=hashnode&utm_medium=feed)

[Sep 7](https://hashnode.com/posts/benchmarking-a-buck-converter-loop-on-bare-metal-arm-cortex-m7/6a9ef800135d92a612552606)