Is Windows 7 Actually Faster Than Ubuntu 10.04?

For many years past, it has been a problem for the Linux Community to gain access to closed hardware technology. Hardware OEM's have gone with the market-share front runner in the past, and true even today, that for whatever reason(s) is Microsoft.

When talking about the "latest" and the "greatest", more and more this means closed patented technology. Even with the drivers developments that have made using Linux with a larger selection of hardware much less of a task then just five years ago, there are those technology areas remain out of reach to Linux developers.

Speed, that has come when a given hardware OEM, has created a LInux just for it, and for the application of the hardware. Or, the same given OEM shares the technology to allow outside development and gain speed(s) equal and beyond. It is this last approach, where an OEM shares information that not only Linux, but other as well can make gains when an OS is running on the OEM's hardware.

Test, no matter how equal the equipment, they are done (no intent), with the advantage to the closed technology. Advantage because the hardware, again, may be technology that Linux developers (not just Ubuntu), don't have access. So while gamming and other number crunching apps tax hardware for more speed; speed comes down to a close fit of hardware and OS.



Just my opinions and takes on these things. Nothing set in stone ?

This whole thing just compares graphics speeds. Graphics cards are notorious for not having good linux support, especially latest and greatest.

It does not do a general purpose sort of benchmark, like lmbench (http://lmbench.sourceforge.net/) did/does.

Lmbench seems unix-a-like specific, but it would be interesting to run it under multiple OSes on the same hardware.

Without something like lmbench, I personally would hesitate to say "Windows 7 is faster" than J. Random Linux Distro.

They did mention at the end that this is part one and that non-gaming benchmarks will be coming up.

Yup. I'll be waiting for the non-gaming benchmarks.

What's interesting about these tests is that it clearly shows that gaming on Linux can be pretty much on-par with Windows. At least that's one myth dispelled.

Good point Sander. I would like to see the follow on performance with more than ext4 in *buntu. Phoronix has listed articles quoting ext3 generally out performing ext4.

It's basically a benchmark between the super duper proprietary intel driver for ms windows and the FOSS intel driver for linux. Now we know which one they've invested more effort into.

In contrast, the nvidia drivers, which are available for a number of platforms including linux, solaris, freebsd, and ms windoze, yield similar perfomance on windows vs linux - probably due to the fact that all supported platforms use what is essentially the same nvidia binary blob, differing only in the platform specific "glue code" or "wrapper" for the binary blob.