Efficient Memory Programming

But! If you're looking for examples of practical use (and abuse) of memory to dramatically boost the performance of your application, the book holds little or no value for you. Starting from the fact that it only briefly describes the internals of PII processor and it has nothing on PIII, PIV and AMD Athlon principles of operating with memory and finishing with code examples which are very simplistic and impossible to take advantage of.

If you're into real-world memory "hacking" and not the academic stuff take a look at "Code optimization: effective memory usage" by Kris Kaspersky.


It contains some valuable insights, but lack the details   July 8, 1999
 4 out of 4 found this review helpful

This is my first book on efficient programming technique. The book offers brief description on how different kinds of memories are organized (specifically, cache and virtual memory). In addition, it breifly introduces data dependencies analysis (which is what I was looking for).

The book does not offer detail enough information on how to do the analysis. Also, the examples in the book are too simple to be applied to real-life problem.

I was especially disappointed by the content of chapter 10(the chapter on advance optimization). I was expecting something that I could used on my code, but all the chapters talked about was optimization that could only be done by compilers.

In conclusion, I think this book is only good for beginners (like me :) who can benefit from some basic ideas on how cache and virtual memory operates and on how the compiler performs some of the optimizations.


Required knowledge for high performance applications   May 13, 1999
 5 out of 5 found this review helpful

Having a good understanding of the impact of the memory hierarchy is discussed throughout the book, including CPU registers, CPU caches, virtual memory, and the hard disk. Cost to access data at each level is also discussed.

The concept locality of reference (temporal and spatial) is discussed and how that can affect performance. For example, while a linked list is a useful structure, it also has performance impacts because each node in the list can be in very different locations in memory, potentially causing a page fault for every node.

This book is for the C/C++ programmer that wants to manage their own memory for high performance. It is not as useful for Basic or other interpreted languages. The author stays pretty general, but gives short examples using the Pentium processor.

I have worked on creating a high-performance server and learned a few things from the book, as well as having some of my assumptions confirmed. The book is well worth your time if your trying to create a high performance application.


Way cool, and about time, too!   April 9, 1999
 1 out of 1 found this review helpful

I started programming with 4K of RAM, moved to 16K, then to 64K. I didn't think that I would need much more than that, since I DID have an eight-inch floppy drive or two to store stuff to...then, suddenly, today happened. Any and all developers should read this book, and make use of its guidance. It may seem a little passe to worry about a program's performance, these days, but you will write better code and create better products if you implement only a tenth of this book's concepts. Buy this and read it. Over and over.


What's the point of a fast CPU if your programs are slow?   March 7, 1999
This is the kind of stuff you'd normally have to spend a couple of years in computer science grad school to learn, and it's all in one book and accessible to any programmer who needs or wants it. I'm tired of having to upgrade my system every year just so I can run Microsoft Word; this is a book that should be required reading at Microsoft.

Loshin explains how memory and processing work, then explains how to use that knowledge to write more efficient programs. If you're a real programmer, you should have this book just to make sure you're not missing out on anything.