[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-kernel
Subject:    Re: Performance Comparision
From:       Phillip Ezolt <ezolt () perf ! zko ! dec ! com>
Date:       1999-04-30 20:31:04
[Download RAW message or body]

Otto,
	Can you do latency tests with lmbench to compare the results between
FreeBSD and Linux on the same machine?  I have seen big differences in 
memory latency between Tru64 and Alpha/Linux on identical Alpha hardware.

I think that it might be a combination of cache-sloshing and lack of page 
coloring, but I haven't had time to track it down more.  

This might be what FreeBSD does better than Linux, and you are seeing the
results with the HINT benchmark.

--Phil

Digital/Compaq:                     HPSD/Benchmark Performance Engineering
Phillip.Ezolt@compaq.com                            ezolt@perf.zko.dec.com


On Fri, 30 Apr 1999, Otto Solares wrote:

> Steve Willer wrote:
> 
> > On Fri, 30 Apr 1999, Phillip Ezolt wrote:
> >
> > > While that is true for "scientific" analysis, a user who picks up BSD will
> > > see the performance difference, and not care whather it was a result of
> > > compilers, linker, etc.
> >
> > I can tell you that personally, I don't spend that much time compiling. I
> > worry about interactive X speed, database speed, networking speed, and
> > speed running Apache. For these purposes, I certainly wouldn't trust a
> > compile-time test that doesn't make all variables equal except for the OS.
> 
> This is where this benchmark gets in, not in compile time, again has nothing
> to do with it, all the problem begins when HINT (a 9kb executable) is
> running and it tells you how much the hardware can do in intervals of
> time but the point is that in the same machine FreeBSD has a better
> performance than linux, HINT overloads tha machine to its maximum
> running floating point operations and it fill the L1 cache then the L2 cache
> then RAM and then swap, then it finished telling you how much your
> machine do. If somebody though about the poor linux performance in this
> area i think we can achieve a better in unbeatable OS.
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/
> 
> 


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

[prev in list] [next in list] [prev in thread] [next in thread]