Kiwi was a collaborative project between the University of Cambridge Computer Laboratory and Microsoft Research Limited, headed by David Greaves (UoCCL) and Satnam Singh (MRL). From 2013 onwards, the Kiwi system was further developed at the Computer Laboratory and using a logic synthesis library called HPR-L/S.

Kiwi is developing a methodology for algorithm acceleration using parallel programming and the C# language. Specifically, Kiwi consists of a run-time library for hardware FPGA execution of algorithms expressed within C# and a compiler, KiwiC, that converts .NETbytecode into Verilog RTL for further compilation for FPGAexecution. In the future, custom domain-specific front ends that generate .NETbytecode can be used.

The Kiwi technology has many potential uses, but some of note are:

  1. Kiwi-HPC: High-performance computing or scientific acceleration.
  2. ASIC hard-core generation for standard algorithms that are to be implemented in silicon, such as MPEG compression.
  3. Routing logic for software-defined networking.
  4. Rapid transaction processing and hardware implementation of automated trading algorithms.

Compared with existing high-level synthesis tools, KiwiC supports a wider subset of standard programming language features. In particular, it supports multi-dimensional arrays, threading, file-server I/O, object management and limited recursion. Release 1 of KiwiC supports static heap management, where all memory structures are allocated at compile-time and permanently allocated to on-FPGA RAM or external DRAM. Release 2 of KiwiC, which has had some successful tests already, supports arbitrary heap-allocation at run time but does not implement garbage collection.

The Kiwi performance predictor is an important design space exploration tool. It enables HPC users to explore the expected speed up of their application as the modify it, without having to wait for multi-hour FPGA compilations in each development iteration.

The Kiwi compiler, KiwiC, itself consists of about 22 klocs (thousand lines of code) of F# (FSharp) code that is a front end to the HPR L/S logic synthesis library that is composed of another 60 or so klocs of F#. The code density for F#, like other dialects of ML, is perhaps (conservatively perhaps) 3 times higher than for common imperative languages like C++, Java and C#, so it is a significant project.

Note that the PDF version of this document tends to be more up-to-date than the HTML version.

David Greaves 2017-06-26