Compare this with Apple's ads about a G3 toasting a Pentium II. In that case, they measured core speed using special instructions, and with the basis of those results, stated that the G3 is more than twice as fast.
Computer architects do not like these benchmark; they are more biased towards SPECint and SPECfp, since those compares system performance using real applications and not core speed using a toy program.
In the same way, the Mac OS Server vs. Linux test should be run on the same hardware, and configured in a correct way - equivalence down to compile options of the web server, if necessary.
Given Apple's tendency to 'twist' the numbers, I wouldn't take those benchmarks too seriously, although I find Mac OS Server quite exciting.
Slashdot Mirror
Compare this with Apple's ads about a G3 toasting
a Pentium II. In that case, they measured core speed using special instructions, and with the basis of those results, stated that the G3 is more than twice as fast.
Computer architects do not like these benchmark; they are more biased towards SPECint and SPECfp, since those compares system performance using real applications and not core speed using a toy program.
In the same way, the Mac OS Server vs. Linux test should be run on the same hardware, and configured in a correct way - equivalence down to compile options of the web server, if necessary.
Given Apple's tendency to 'twist' the numbers, I wouldn't take those benchmarks too seriously, although I find Mac OS Server quite exciting.
regards,