Performance

So it happens.

As the website states, it works by:

• Sending your page requests through Google machines dedicated to handling Google Web Accelerator traffic.
• Storing copies of frequently looked at pages to make them quickly accessible.
  Downloading only the updates if a web page has changed slightly since you last viewed it.
• Prefetching certain pages onto your computer in advance.
• Managing your Internet connection to reduce delays.
• Compressing data before sending it to your computer.

To quote him:

A GPL’ed (say, it’s not OpenSource, but the source code is available publicly) PHP accelerator has been evaluated at lab, which is called eAccelerator. I should adjust my claim that this sort of application is not useful.

During my test, the accelerator has improved the TPS by about 50%, and is quite impressive.

In order to install this on FreeBSd, you should have installed PHP from ports, and go to www/eaccelerator and ``make install’'.

A post on tech-kern has suggested that NetBSD has better performance on several aspects. I think FreeBSD should take some of their ideas. What’s more, I believe it’s more important to have something that is unique in FreeBSD, not doing what’s the equivalent because “Solaris have this” :-)

Finally NetBSD has released their 2.0-RELEASE, which is a significant progress of the NetBSD development.

NetBSD is widely known as the most portable operating system in the world. It currently supports fifty four different system architectures, all from a single source tree, and is always being ported to more. NetBSD 2.0 continues the long tradition with major improvements in file system and memory management performance, major security enhancements, and support for many new platforms and peripherals.

What is strict aliasing? To make the long thing short, it means that the compiler apply strictest aliasing rules applicable to the language and having these assumptions the optimizer can do further optimizations.

To quote gcc manual:

I have been suffered by the speed of GNOME because it usually gets “jitter”.

It seems that marcus@ have the same problem. I have tried HZ=100 but it seems that nothing was changed. Then I have disabled ACPI and it seems the issue was worked around.

What happend in ACPI, so? Still under investigate.

Today NetApplience folks has told me that Linux performance is twice better than FreeBSD when using em(4) driver, and they pointed out that the driver FreeBSD is using does not do a good job at this time. I have tried to tune some certain part of the kernel however the result is still not so good at the time I am writing.

Scott Long have committed this which will:

It turns out that we have to optimize our FS before benchmarking. I have read some articles and maillist archives so I got some impression:

1. Give correct parameters to newfs(1)
   avgfilesize, filesperdir, will control DIRPERF hashing.

2. Turn on noatime mounts.

3. etc.

Just make a note.