Does FreeBSD 4.11-STABLE support the 8237R?

Danial Thom danial_thom at yahoo.com
Sun Jun 4 08:26:30 PDT 2006


I would guess the gigabit switch vs my cheapo
100Mb/s switch would make that sort of
difference. You have to do the test with the same
hardware, same server, same switches otherwise
you have no relative comparision that's valid.
The server could make a huge difference also. Ftp
servers are kind of kludgy in the way that they
decide on how to dispatch packets. I'd guess that
if you put an intel card into the box you'd get
similar relative differences in throughput.

Its not a good test anyway, but you have to
simplify things for jerry and the gang. A better
way to test is to generate a consistent load and
look at CPU usage. The efficiency of the server,
windowing, etc will all affect an FTP transfer
too much to use it as a complete test. The
seemingly tiny different between gigabit and
100Mb/s speeds could be the difference between
the window staying open or the process going to
sleep.
.

--- "Chad Leigh -- Shire.Net LLC"
<chad at shire.net> wrote:

> 
> On Jun 4, 2006, at 8:42 AM, Danial Thom wrote:
> >
> > How is informing someone that they're wasting
> > their money on a MB a waste of time? I'm sure
> > you've wasted thousands of your employers
> dollars
> > with your ignorant recommendations, Jerry. I
> can
> > get hours of entertainment just googling you.
> >
> > Ok, here's a test to illustrate my point. I
> have
> > a server with a big file (352MB). 2 client
> > machines running the same version of Freebsd:
> >
> > 1) AMD 1.8Ghz Opteron - onboard bge
> controller:
> >
> > Ftp results: 4MB/s
> 
> You have something wrong then.    I have 2 such
> machines, both with  
> the bge on a simple 32bit/33mhz pci bus, not on
> a 64 bit or a faster  
> pci-x bus (Tyan S2850 boards in both, both with
> Opteron 244  
> (1.8ghz)).  They are connected together with a
> low level (ie, less  
> expensive) gigabit switch with standard MTU
> size
> 
> bge1:
>
flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST>
> mtu 1500
>         
> options=1a<TXCSUM,VLAN_MTU,VLAN_HWTAGGING>
>          inet 192.168.2.129 netmask 0xffffff00
> broadcast 192.168.2.255
>          ether 00:e0:81:60:0c:f7
>          media: Ethernet autoselect (1000baseTX
> <full-duplex>)
>          status: active
> 
> bge1:
>
flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST>
> mtu 1500
>         
> options=1a<TXCSUM,VLAN_MTU,VLAN_HWTAGGING>
>          inet 192.168.2.110 netmask 0xffffff00
> broadcast 192.168.2.255
>          inet6 fe80::2e0:81ff:fe64:ae9d%bge1
> prefixlen 64 scopeid 0x2
>          inet 192.168.2.111 netmask 0xffffffff
> broadcast 192.168.2.111
>          inet 192.168.2.112 netmask 0xffffffff
> broadcast 192.168.2.112
>          inet 192.168.2.113 netmask 0xffffffff
> broadcast 192.168.2.113
>          inet 192.168.2.114 netmask 0xffffffff
> broadcast 192.168.2.114
>          ether 00:e0:81:64:ae:9d
>          media: Ethernet autoselect (1000baseTX
> <full-duplex>)
>          status: active
> 
> I ftp'ed a 3.9GB file between them using simple
> ftp protocol.
> 
> 4227530240 bytes sent in 05:06 (13.16 MB/s)
> 
> Both have dual bge ethernet ports and this was
> on the secondary port  
> on each, bge1.   The primary port, bge0, on
> each is hooked to a  
> 100BaseT switch and one server (origination)
> was serving a bunch of  
> http/php on bge0 and the other was serving
> clamav/spamassassin at the  
> same time.
> 
> On the origination side I did top and the load
> barely moved during  
> the ftp and system CPU time was a few % higher.
>  Not ideal but not a  
> deal breaker either.
> 
> Chad
> 
> >
> > 2) Intel 2.0Ghz Celeron 845 Chipset, onboard
> fxp
> > controller:
> >
> > ftp results: 11MB/s
> >
> > I think we'll all agree that a 1.8Ghz opteron
> is
> > substantially faster (and more expensive)
> then a
> > 2.0Ghz Celeron? (or will Jerry ask me to
> prove
> > this also)? Its not rocket science. What good
> is
> > the extra horsepower of the cpu doing you if
> > you're using a crap controller? Its mindless
> > stupidity; which is about what you'd expect
> from
> > a sys admin, and not an engineer. The problem
> > with this list is that its all sys admins, so
> > learning from other idiots just causes  you
> to be
> > just as stupid at your "teachers".
> >
> > DT
> >
> >
>
__________________________________________________
> > Do You Yahoo!?
> > Tired of spam?  Yahoo! Mail has the best spam
> protection around
> > http://mail.yahoo.com
> >
> _______________________________________________
> > freebsd-questions at freebsd.org mailing list
> >
>
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> > To unsubscribe, send any mail to
> "freebsd-questions- 
> > unsubscribe at freebsd.org"
> 
> ---
> Chad Leigh -- Shire.Net LLC
> Your Web App and Email hosting provider
> chad at shire.net
> 
> 
> 
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 


More information about the freebsd-questions mailing list