Some Really Good PC Performance Testing Software!

After my rant the other week about video software that didn’t work, I have been very pleasantly surprised tonight by a number of benchmarking programs, so I felt some praise was in order.

I’ve noticed that one of my desktops, is considerably faster than the other one, even though it has much less RAM.  So, I wanted to stress test the machines, and get some real numbers.

My first recommendation is a website that has reviewed many free benchmarking tools:

I only looked in depth at the CPU section. But it has 18 total sections worthy to look at including such topics as RAM, and hard disks.  A number of the products were rather old, going back to the 2002 era.  If you are a benchmarking geek, they also have forums for you.  In either English, or Russian.

From that page, I was able to find the program Masmo PrimeBench.

Masmo PrimeBench

PrimeBench calculates prime numbers to many decimal places. By measuring the time, you get a quick and dirty way to test how fast the CPU is working.   And, it makes a cool graph of the prime numbers.

Another cool program is Max Multicore PI, that calculates PI to many, many decimal places.

MaxPi PI Calculator

The interface is in German.  But you choose how many decimal places of PI that you want.  From 8k, up to 1 million.  Accurate enough?  And then let it calculate.  Again, the time it takes will give you a rough idea how fast your CPU is.

Then I discovered that the writer, M. Bicak, has gone on to much bigger and better things.

He has expanded on the original PI program, and has been making a number of tools to benchmark your single, or multicore machine.



PI calculator—preview-multi.php

Memory Test:—preview.php

MIPS Test:—preview.php

FLOPS Test:—preview.php

And others.

The software installs perfectly.  In fact, it is just a zipped up file.  You unzip it, and run the resulting executable.  No other software to install.  Thank you!

He has also made available the results from the field, so you can compare your machine to the rest of the world.  Gentlemen, start your computers.—mips.php

I highly recommend these programs.  Do check them out!

The most comprehensive program I found was the Passmark PerformanceTest Software, free to try for 30 days.

Passmark PerformanceTest Software

It did want me to install Direct X 9.  Surprise.  But it still worked without it.  Thank you.  But as in my other post, why was the Direct X 9 desired requirement not stated somewhere in the pre-reqs?

Passmark puts your machine through many tests:  CPU, memory, graphics, hard drives, even the DVD drives, etc.  Then gives real number counts as to the performance of your machine(s).  It also creates a composite number based on all the different tests.  I found it to be pretty accurate, based on my experience with my three machines.

Passmark also has an extensive online database of real world results that you can compare your components too. So, if you are thinking of upgrading your machine, you can use the webpage as a starting point as to what hardware you want to use.

Great programs guys!  Thank you for writing some good software, and pleasantly surprising me.


What I’m thinking now with these tools, is to go down to the computer store with the programs on a USB stick, and test out some new machines.  No sense buying something if it’s not much faster than what I already have.

Note:  I should point out that all these programs only seem to work on Windows.  I didn’t notice anything about Linux, Unix, or Mac.  That’s also something that should be mentioned on their  webpages.


It’s interesting that so many performance testing programs worked perfectly the first time.  Do you think that if you are often looking at raw performance numbers, that your thinking changes, and you begin to write better software?  A number is objective, not subjective. There is only one test (the software) and one independent variable (the machine). The results are simple.  Not complex.

This is so in contrast to so many shops I’ve worked at, where there was always more and more complexity.  And more opinions.  But no testing, numbers, or proving.


Now here is another question.  Why can’t I find similar composite and detailed numbers for off the shelf systems, say from Dell, HP, or Alienware?  There are lots of editor reviews of machines, as  on CNET. They have their editors picks, opinions, and lots of words.  But the reviews don’t really have simple numbers to make easy comparisons.

Passmark has listed the fastest machines they have ever discovered in the real world, using the exact same software tests on every machine:

But which company makes that system?  Where can I buy it please?   All I’ve got is the CPU, the video card, RAM, and the operating system.

This is yet another observation of mine that we are getting lots of -data- these days.  Much of it of a certain entertainment value, of little consequence.  But so little in what I’m reading online has much useful -information-.

Happy stress testing everyone!

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: