From 250b9c60d3509b837a9dcd84672905059b8a776f Mon Sep 17 00:00:00 2001
From: jf3g19 <jf3g19@soton.ac.uk>
Date: Sun, 29 Nov 2020 22:09:06 +0000
Subject: [PATCH] Update README.md

---
 README.md | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

diff --git a/README.md b/README.md
index 99431b3..c0261ac 100644
--- a/README.md
+++ b/README.md
@@ -1 +1,36 @@
+**Contents**
+
 Implementation of a Basic Distributed Arithmetic Architechture for MAC
+
+**Operation**
+
+- The Distributed Arithmetic Architechture is based on the assumption that for a given Sum Of Products (SoP) between two arrays `C` and `X` that you know the value of `C` at compile time.
+
+- Based on this assumption we can generate a LUT which allows us to transform the SoP to a series of shifts and adds.
+
+- The File generate_LUT.cpp can be used to generate the LUT values for a given set of input coefficients. If you compile the file it can be run in the command line in the following format:
+
+`> generate_LUT [Multiplicand Bit-Width] [C[0]] [C[1]] [C[2]] [C[3]] .... [C[n]] `
+
+i.e
+
+`generate_LUT 16 -4 -3 1 9`
+
+would generate the LUT for the coefficients c[0] = -4, c[1] = -3, c[2] = 1, c[3] = 9, and a 16-bit Bit-Width for your Multiplicand.
+
+The generate multiplication coefficients can be stored as an unpacked array of packed values and should be connected directly to the `multiplication_coefficients` input port.
+
+
+**Example**
+
+An Example ModelSim output for the testbench provided, which displays the inner product of 
+
+`X = [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16]`
+`C = [ [-79 -136 312 654 -1244 -2280 4501 14655 14655 4501 -2280 -1244 654 312 -136 -79]`
+
+[output](https://imgur.com/WczvADF)
+
+We can see that after input ready goes high, 17 clock cycles later, output ready goes high [with the correct value 278,511](https://imgur.com/ctePVPN)
+
+
+
-- 
GitLab