In article <dl3p58$1ugl$1@xxxxxxxxxxxxxxxxx>,
Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> wrote:
> Microsoft finally *give* the fact, but hide them under a thick blanket (PDF)
> that would better not leak to the world:
(PDF is a thick blanket???)
> <quote>
> What's noteworthy about it is that Microsoft compared Singularity to FreeBSD
> and Linux as well as Windows/XP - and almost every result shows Windows
> losing to the two Unix variants. For example, they show the number of CPU
> cycles needed to "create and start a process" as 1,032,000 for FreeBSD,
> 719,000 for Linux, and 5,376,000 for Windows/XP.
> </quote>
(Note: I'm not commenting on whether or not the benchmarks are accurate,
but rather on the accuracy of the reporting)
Well, mathematically, he is correct, in the sense that to a
mathematician, "almost every" means "all but a countable number of
exceptions". However, to most people, he's got a weird definition of
"almost every".
In that table of basic operations, there are 6 comparisons"
Cost (CPU Cycles)
Singularity FreeBSD Linux Windows
Read cycle counter 8 6 6 2
ABI call 87 878 437 627
Thread yield 394 911 906 753
2 thread wait-set ping pong 1,207 4,707 4,041 1,658
2 message ping pong 1,452 13,304 5,797 6,344
Create and start process 300,000 1,032,000 719,000 5,376,000
Windows loses to "the two Unix variants" in 1 of 6, beats both in 2 of
6, and is between Linux and FreeBSD in 3 of 6.
In the I/O tests, Windows lost to both slightly on the random read and
write tests, and was between or won slightly on most of the sequential
tests.
--
--Tim Smith
|
|
