Mac OS X at Odds With Linux

by ServerWatch Staff

Benchmarks and dueling e-mails pit Mac OS X directly against Linux.

Mac OS X's BSD-based kernel puts it at the Unix end of the OS spectrum and thus ripe for comparison with Linux. And this is precisely what's been happening as of late. Some in the IT industry have called OS X on the carpet for perceived problems compared to Linux.

Dr. Jasjeet Sekhon fired the opening salvo against OS X. Sekhon, associate professor, Travers Department of Political Science Survey Research Center at UC Berkeley, penned a study titled "Linux versus Mac OS X and Windows XP on Intel Dual Core." In that article, Sekhon tested the three operating systems on a new Intel-based MacBookPro and found "Linux is found to be much faster than Apple's OS X for statistical computing. And although Linux is 5 to 10 percent faster than Windows XP, both are markedly faster than OS X. For example, in one benchmark both Linux and Windows XP are more than twice as fast as OS X."

That certainly got the attention of the Apple crowd. Some pro-Mac pundits have gone so far as to defend against the notion that Linux should replace the Darwin kernel for OS X.

Other responses blast the Sekhon study for using benchmarks that would deliberately highlight differences in the way OS X handles certain tasks — such as memory allocation — compared to Linux and Windows. While running statistics on R, for instance, the OS X test used the default memory allocator tool, but both Linux and Windows had a better allocator installed. The fault, according to counter-arguments, is not in the Darwin kernel but rather the choice of tools surrounding the kernel. Using the same Lea memory allocation tool found in the other two operating systems, OS X's benchmarks suddenly matched the Windows XP times.

Benchmark tests are bound to miss things like this. Still, the notion in the media that something is not well in OS X-on-Intel-Land is catching on strongly.

This was evidenced by the next salvo, a Tom Yager opinion piece on InfoWorld that accused Apple of preparing to close the source code for the Darwin kernel on the Intel platform. Actually, Yager wrote, the deed was already done and the OS X kernel (at least on Intel boxes) was already closed. Yager seems to have good sources on this, or is just plain overconfident. We had to wonder which, since Apple shifted into denial mode.

Apple's product manager, Ernest Prabhakar, said: "Just to be clear, Tom Yager was speculating about why we have — so far — not released the source code of the kernel for Intel-based Macs. We continue to release all the Darwin sources for our PowerPC systems, and so far have released all the non-kernel Darwin sources for Intel."

This article was originally published on Friday May 26th 2006
Mobile Site | Full Site