[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: linux vs. micro$oft



On Sun, 27 Jun 1999 14:39:17 +0300 (IDT), "Stanislav Malyshev a.k.a Frodo"
<frodo@sharat.co.il> wrote:

>MA>> Does anyone have anything to say about it ?
>MA>> Can micro$oft actually overcame (i truly hope not) Linux ? :-((
>
>NOT AGAIN! 
>We had a large flame already here on the subject. Slashdot had
No we did not. This was after the *first* benchmark. Now, the *2nd* benchmark
is out, which is more balanced this time, and no one denies it's results, only
their interpretations.

>termonuclear-explosion-sized flame on it. Every Linux top person wrote a
>long responce to it. Everybody is tired on it. Research the archives.

Summary of results (as far as I remember):

1. NT is faster (samba vs. NT server, apache vs. IIS), but not as it was seen
by the original benchmark. On a single processor box, it's only about 30%
slower (~1300 hits/s vs. ~1800 hits/s), while with a quad CPU box, NT has a
bigger advantage (~4000 hits/s vs. 1800 hits/s).

2. Everyone came to the conclusion that Linux TCP/IP is the problem, because
it's not reentrant (this will probably be improved by 2.3.x).

3. Someone is working on a static page module for apache, which will compete
with IIS speed.

4. This is irrelevant, because in order to get these speeds you must have
something like 10 T1 lines at peak bandwidth to actually see the slowdown.

If you want the references, look for them on slashdot.

Udi

To unsubscribe, send mail to linux-il-request@linux.org.il with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail linux-il-request@linux.org.il