More computing bang for your buck on Linux?

Question

I have both a Linux box (running Debian) and a powerful Windows desktop running XP. At work I have a Windows 8 machine.

I notice that my Linux box, even though it is just a little Zotac seems to outperform the XP machine, but I have not done quantitative tests. Given the same motherboard, do Linux boxes run computational applications faster than Windows? For example, if I were to take Mathematica, install on two identical machines, one running Linux, the other Windows XP, and then compute PI to a trillion places or something like that, would the Linux box be noticeably faster than the Windows box?

Answer 1

There will be little to no measurable difference for pure computational tasks, or the differences will be masked by other things such as disk access or other OS features taking precedence.

There's a nice set of benchmarks at TuxRadar: Benchmarked: Ubuntu vs Vista vs Windows 7

For the first few graphs the benchmarks are meaningless and a bit pointless, the Initial Install Size one is silly as we all know Windows likes to make use of hard disk space, while linux is "sleek" and expects you to install everything from the repos once you get it going.

The main benchmark relevant to your question is the processor intensive one:

Richards benchmark

Notes: This was done using the cross-platform Python port of Richards. For reference, Ubuntu 8.10 uses Python 2.5.2, Ubuntu 9.04 uses Python 2.5.4, and we used Python 2.5.4 on the Windows tests. Even though the 64-bit results for Linux and Windows don't look that far apart, we have to admit to being very impressed with the Windows tests - the deviation between tests was just 3ms on Vista, and 5ms on Windows 7, compared to 20ms on Linux.

_{Amount of time taken to execute the Python Richards benchmark. Measured in milliseconds; less is better.}

It's clear from that graph that having a 64-bit OS can make a real difference in compute-intensive tasks, but it's not too pleasing to see Windows pip Linux to the post in nearly all results.

Essentially the set of benchmarking shows that while the OS makes a difference for how disk is accessed and the general feeling of responsiveness, when you get down and actually put the thing to work the differences become very negligible and could even be attributed to measurement error. Linux does not "win" at this and, IMO, neither does Windows. The differences are too minor.

Further reading:

• Windows 10 vs. SteamOS: Benchmark Test Reveals Linux Lags Far Behind
• Windows 7 vs Ubuntu 9.10 (more focused on disk than actual raw CPU, no suprises that Linux can be more effective/responsive. No actual CPU usage tests.)