[Shootout-list] Mono timings correct for tcp* tests?
Isaac Gouy
igouy2@yahoo.com
Mon, 21 Mar 2005 12:36:20 -0800 (PST)
>From Dave's next experiment it seems like we're timing the server
process alone?
# time tcl tcpecho.tcl 10
real 0m3.760s
user 0m0.777s
sys 0m1.130s
replies: 64000 bytes: 4096000
Looks like I get about the same thing for TCL. If I insert a timer
inside the client block of the mono code (I'm not real familiar with
TCL so haven't done that), the internal timer reports back ~3.6 seconds
alot closer to the 'real' time than the user+sys time.
--- Dave <davejf@frontiernet.net> wrote:
> I just tried this with mono v1.1.4 (they recommend upgrading to this
> on
> their homepage) and it displays the same behaviour as 1.0.6 on my
> system, RH
> Fedora Core 3.
>
> - Dave
>
> ----- Original Message -----
> ># mono --version
> > Mono JIT compiler version 1.0.6, (C) 2002-2004 Novell, Inc and
> > Contributors. www.go-mono.com
> > TLS: normal
> > GC: Included Boehm (with typed GC)
> > SIGSEGV : normal
> > Globalization: ICU
> >
> > ----- Original Message -----
> > Sent: Monday, March 21, 2005 3:15 AM
> > Subject: Re: [Shootout-list] Mono timings correct for tcp* tests?
> >
> >
> >> Which version of Mono?
> >>
> >> --- Dave <davejf@frontiernet.net> wrote:
> >>>
> >>> If I run tcpecho on an otherwise idle machine, there is a pretty
> >>> large
> >>> difference between user+sys and real times.
> >>>
> >>> # time mono tcpecho.exe 10
> >>> replies: 64000 bytes: 4096000
> >>>
> >>> real 0m3.658s
> >>> user 0m0.559s
> >>> sys 0m1.016s
> >>>
> >>> I'm wondering which is accurate?
> >>>
> >>> If user+sys is wrong I'd guess it's because two seperate mono
> >>> processes are
> >>> being run - one for the Listener and one for the Client.
> >>>
> >>> Here's a process listing for tcpecho:
> >>> root 18615 0.2 1.4 34172 7624 pts/6 Sl+ 20:48 0:00
> mono
> >>> tcpecho.csharp_run 100
> >>> root 18616 0.0 0.5 18588 2792 ? Ss 20:48 0:00
> mono
> >>> tcpecho.csharp_run 100
> >>> root 18619 0.3 1.0 33932 5668 ? Sl 20:48 0:00
> >>> /usr/bin/mono
> >>> tcpecho.csharp_run
> >>>
> >>> Here's a process listing for 'heapsort':
> >>> root 18605 0.2 15.9 105204 82128 pts/6 Sl+ 20:47 0:00
> mono
> >>> heapsort.csharp_run 10000000
> >>> root 18606 0.0 0.4 17920 2236 ? Ss 20:47 0:00
> mono
> >>> heapsort.csharp_run 10000000
__________________________________
Do you Yahoo!?
Yahoo! Small Business - Try our new resources site!
http://smallbusiness.yahoo.com/resources/