Re: [ql-developers] HTTP server speed test

2004-02-17 Thread Peter Graf
Richard wrote:

> With the Q60/QLwIP as server I currently get average 854 kBytes/sec HTTP
> throughput, measured with Ethereal, transferring a file larger than 10MB.
> It would be interesting how far from the optimal value this is still away.
>
if it is really reading all 10 MB of data from disk than thats very
fast, the disk access wont allow for much faster operation.
No, it's from buffer, as I wrote in another posting. I have to use Qubide.

Otherwise pure networking speed over TCP gets very close to the theoretical
1.25 MB/s between 2 Linux computers
Interesting info.

Measuring with Ethereal is an overkill,
Since the overkill just costs a few mouseclicks I don't care :-)

better measure is

ifconfig eth0; wget ; ifconfig eth0

... returns raw and usable speed. I get this, 172.16.0.2 is the
Q60 running Linux:
ifconfig eth0; wget -O /dev/null http://172.16.0.2/file.xxx; ifconfig eth0
eth0  Link encap:Ethernet  HWaddr 00:E0:7D:84:3B:FB
  inet addr:172.16.0.1  Bcast:172.16.0.255  Mask:255.255.255.0
  UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
  RX packets:8278 errors:0 dropped:51 overruns:0 frame:8
  TX packets:4756 errors:4 dropped:0 overruns:0 carrier:8
  collisions:68 txqueuelen:1000
  RX bytes:11069431 (10.5 Mb)  TX bytes:398565 (389.2 Kb)
  Interrupt:12 Base address:0x2000
--18:22:46--  http://172.16.0.2/file.xxx
   => `/dev/null'
Connecting to 172.16.0.2:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10,485,760 [text/plain]
100%[>] 10,485,760   982.07K/sETA 
00:00

18:22:56 (982.07 KB/s) - `/dev/null' saved [10485760/10485760]

eth0  Link encap:Ethernet  HWaddr 00:E0:7D:84:3B:FB
  inet addr:172.16.0.1  Bcast:172.16.0.255  Mask:255.255.255.0
  UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
  RX packets:15841 errors:0 dropped:51 overruns:0 frame:8
  TX packets:8563 errors:4 dropped:0 overruns:0 carrier:8
  collisions:68 txqueuelen:1000
  RX bytes:22054706 (21.0 Mb)  TX bytes:649940 (634.7 Kb)
  Interrupt:12 Base address:0x2000
So raw and usable speed are about 5% appart including HTTP overhead,
982 KB/s and 1028KB/s.
Thanks! How have you calculated raw speed?

(Delta ifconfig RX bytes / wget bytes) * wget speed ?

Note that the timing may depend on the binary pattern of the file,
eg a file consisting of all 0 should theoretically need considerably
more time than a file with random patterns
:-)

All the best
Peter



Re: [ql-developers] HTTP server speed test

2004-02-17 Thread Richard Zidlicky

On Sun, Feb 15, 2004 at 06:05:51PM +0100, Peter Graf wrote:
> 
> Hi,
> 
> is somebody out there who owns two networked PCs, one with a 10 Mbit/sec 
> ISA NE2000 (or clone) card, and could do a speed test using this PC as HTTP 
> server?
> 
> With the Q60/QLwIP as server I currently get average 854 kBytes/sec HTTP 
> throughput, measured with Ethereal, transferring a file larger than 10MB. 
> It would be interesting how far from the optimal value this is still away.
> 

if it is really reading all 10 MB of data from disk than thats very
fast, the disk access wont allow for much faster operation.
Otherwise pure networking speed over TCP gets very close to the theoretical
1.25 MB/s between 2 Linux computers

Measuring with Ethereal is an overkill, better measure is

ifconfig eth0; wget ; ifconfig eth0

... returns raw and usable speed. I get this, 172.16.0.2 is the
Q60 running Linux:


ifconfig eth0; wget -O /dev/null http://172.16.0.2/file.xxx; ifconfig eth0
eth0  Link encap:Ethernet  HWaddr 00:E0:7D:84:3B:FB  
  inet addr:172.16.0.1  Bcast:172.16.0.255  Mask:255.255.255.0
  UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
  RX packets:8278 errors:0 dropped:51 overruns:0 frame:8
  TX packets:4756 errors:4 dropped:0 overruns:0 carrier:8
  collisions:68 txqueuelen:1000 
  RX bytes:11069431 (10.5 Mb)  TX bytes:398565 (389.2 Kb)
  Interrupt:12 Base address:0x2000 

--18:22:46--  http://172.16.0.2/file.xxx
   => `/dev/null'
Connecting to 172.16.0.2:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 10,485,760 [text/plain]

100%[>] 10,485,760   982.07K/sETA 00:00

18:22:56 (982.07 KB/s) - `/dev/null' saved [10485760/10485760]

eth0  Link encap:Ethernet  HWaddr 00:E0:7D:84:3B:FB  
  inet addr:172.16.0.1  Bcast:172.16.0.255  Mask:255.255.255.0
  UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
  RX packets:15841 errors:0 dropped:51 overruns:0 frame:8
  TX packets:8563 errors:4 dropped:0 overruns:0 carrier:8
  collisions:68 txqueuelen:1000 
  RX bytes:22054706 (21.0 Mb)  TX bytes:649940 (634.7 Kb)
  Interrupt:12 Base address:0x2000 

So raw and usable speed are about 5% appart including HTTP overhead,
982 KB/s and 1028KB/s.

Note that the timing may depend on the binary pattern of the file,
eg a file consisting of all 0 should theoretically need considerably
more time than a file with random patterns

Richard


Re: [ql-developers] HTTP server speed test

2004-02-16 Thread Peter Graf
Tony Firshman wrote:

I have not managed to get my debian system to install a mouse
successfully, so only have command line.
Is there a command line ethereal?
Not that I know. If you only have a commandline, you could transfer the 
file with curl, which I remember to have a useful speed display. But it 
displays the payload rather than the raw data, so it will not directly 
compare to my figure. If it works I can measure with curl for comparison.

Peter




Re: [ql-developers] HTTP server speed test

2004-02-16 Thread Peter Graf
Tony Firshman wrote:

>> >Unfortunately I have no second PC so I can only measure from PC to 
Q60 and
>> >vice versa.
>> I am sure I have such a card.
>> I will fit it in one of the nine PCs here tomorrow (5 run 24/7!) and
>> test.   It may not be until the evening.
>
>Many thanks!
I did have an NE2000.
However I am haivng trouble under W98.
It is not pnp but W98 found and installed it with NE2000 compat driver.
It though did not show up on my switch.
I connected the 'default' jumper, and it now show the 10mbps conneciton
LED, which flashed reassuringly on startup.
It also show up as working under device mgr.
However it is not finding my network.

Any ideas?  I have played with a few IRQs.
Default was 3, and it is now:
IRQ 9
I/O 0300-031F
All was fine with the existing pci card, which I removed.
Maybe selecting TCP/IP protocol under network bindings? However tinkering 
with your Windows settings looks like more trouble than this test is worth. 
I don't want you to waste too much precious time. Many thanks,

Peter




Re: [ql-developers] HTTP server speed test

2004-02-16 Thread Tony Firshman

I have not managed to get my debian system to install a mouse
successfully, so only have command line.

Is there a command line ethereal?
-- 
 QBBS (QL fido BBS 2:252/67) +44(0)1442-828255
 tony@.co.uk  http://www.firshman.co.uk
   Voice: +44(0)1442-828254   Fax: +44(0)1442-828255
TF Services, 29 Longfield Road, TRING, Herts, HP23 4DG



Re: [ql-developers] HTTP server speed test

2004-02-16 Thread Tony Firshman

On  Mon, 16 Feb 2004 at 11:32:25,  wrote:
(ref: <[EMAIL PROTECTED]>)

>
>On 15 Feb 2004 at 22:15, Tony Firshman wrote:

>> >Unfortunately I have no second PC so I can only measure from PC to Q60 and
>> >vice versa.
>> I am sure I have such a card.
>> I will fit it in one of the nine PCs here tomorrow (5 run 24/7!) and
>> test.   It may not be until the evening.
>
>Many thanks!
I did have an NE2000.
However I am haivng trouble under W98.
It is not pnp but W98 found and installed it with NE2000 compat driver.
It though did not show up on my switch.
I connected the 'default' jumper, and it now show the 10mbps conneciton
LED, which flashed reassuringly on startup.
It also show up as working under device mgr.
However it is not finding my network.

Any ideas?  I have played with a few IRQs.
Default was 3, and it is now:

IRQ 9
I/O 0300-031F

All was fine with the existing pci card, which I removed.
>
>> What is Ethereal?  Will a stop watch do?
>
>It will be difficult with a stop watch. Ethereal is a fine network
>protocol analyzer. If you have one of the usual Linux distros installed,
>you're likely to already have it. BTW it's free software, and easily
>ourperforms commercial tools that cost > 1000 EUR.
>
>http://www.ethereal.com
It wasn't (under debian) but I have now installed it (using apt-get)
>
>If you want to try it, just say "Capture/Start", choose the interface and
>"OK". Stop when you're finished and look at "Tools/Summary". The best is,
>if you capture only during the transfer, so the result is not influenced
>by human interaction with the browser.



-- 
 QBBS (QL fido BBS 2:252/67) +44(0)1442-828255
 tony@.co.uk  http://www.firshman.co.uk
   Voice: +44(0)1442-828254   Fax: +44(0)1442-828255
TF Services, 29 Longfield Road, TRING, Herts, HP23 4DG



Re: [ql-developers] HTTP server speed test

2004-02-16 Thread pgraf

On 15 Feb 2004 at 22:15, Tony Firshman wrote:

> 
> On  Sun, 15 Feb 2004 at 18:05:51, Peter Graf wrote:
> (ref: <[EMAIL PROTECTED]>)
> 
> >
> >Hi,
> >
> >is somebody out there who owns two networked PCs, one with a 10
> >Mbit/sec ISA NE2000 (or clone) card, and could do a speed test using
> >this PC as HTTP server?
> >
> >With the Q60/QLwIP as server I currently get average 854 kBytes/sec
> >HTTP throughput, measured with Ethereal, transferring a file larger
> >than 10MB. It would be interesting how far from the optimal value this is
> >still away.
> >
> >My test setup:
> >Server: Q60/80, Realtek RTL8019AS ISA, QDOS Classic 3.25 Beta R, QLwIP
> >Client: Athlon 1.2 GHz, Realtek RTL8029AS PCI, RedHat Linux 9.0, Netscape
> >browser
> >
> >Unfortunately I have no second PC so I can only measure from PC to Q60 and
> >vice versa.
> I am sure I have such a card.
> I will fit it in one of the nine PCs here tomorrow (5 run 24/7!) and
> test.   It may not be until the evening.

Many thanks!

> What is Ethereal?  Will a stop watch do?

It will be difficult with a stop watch. Ethereal is a fine network 
protocol analyzer. If you have one of the usual Linux distros installed, 
you're likely to already have it. BTW it's free software, and easily 
ourperforms commercial tools that cost > 1000 EUR.

http://www.ethereal.com

If you want to try it, just say "Capture/Start", choose the interface and 
"OK". Stop when you're finished and look at "Tools/Summary". The best is, 
if you capture only during the transfer, so the result is not influenced 
by human interaction with the browser.

> Have you played with the I2C stuff yet?

Not yet. Anyway I wanted to purchase Minerva, because you and Lau have 
kindly placed it under the GPL, which deserves some recognition.

All the best
Peter



Re: [ql-developers] HTTP server speed test

2004-02-16 Thread pgraf

On 15 Feb 2004 at 19:30, Fabrizio Diversi wrote:

> I have a lot of different configurations, but the only 10 Mb card I
> have is used into the Q60 and Q40. I have to think about how to
> arrange it. BTW 854 kb/sec I think is near the maximum you can reach
> (10% less than the nominal value of 1000) and is far from what i
> thought was possible to obtain with QLwIP . Good results, according
> the fact that you told us QLwIP was slow ..

A bit of misunderstanding I guess. QLwIP with polled ethernet driver never 
was slow by itself, but I deliberately reduced speed, because for some 
(yet unknown) reason, a massive usage of scheduler calls under QDOS (and 
SMSQ/E likewise) could make a job stay suspended although there had been a 
release call. It is yet unclear wether this is a problem of QLwIP or the 
OS. Neither QDOS nor SMSQ/E have ever been tested with something like 
1 scheduler calls per second, from several different jobs, partially 
triggered by asynchronous events. So I have no reference.

What you might be referring to, is the principal shortcoming of QDOS if an 
ISR-driven structure is used. With such a clean driver, _every_ 
multitasking TCP implementation that uses jobs for the higher protocol 
levels is indeed doomed to a limited throughput, i.e. "slow".

So the figure of 854 kB/sec stems from a polled driver. I've been able to 
reduce the amount of scheduler calls at the cost of some more polling. 
This version is still experimental and not 100% stable.

I'm not sure if 854 kB/sec is a good result. BTW I only get this if the 
file is already buffered. Otherwise the efficiency of harddisk plus driver 
(or the lack thereof) is measured.

All the best
Peter



Re: [ql-developers] HTTP server speed test

2004-02-15 Thread Tony Firshman

On  Sun, 15 Feb 2004 at 18:05:51, Peter Graf wrote:
(ref: <[EMAIL PROTECTED]>)

>
>Hi,
>
>is somebody out there who owns two networked PCs, one with a 10
>Mbit/sec ISA NE2000 (or clone) card, and could do a speed test using
>this PC as HTTP server?
>
>With the Q60/QLwIP as server I currently get average 854 kBytes/sec
>HTTP throughput, measured with Ethereal, transferring a file larger
>than 10MB. It would be interesting how far from the optimal value this
>is still away.
>
>My test setup:
>Server: Q60/80, Realtek RTL8019AS ISA, QDOS Classic 3.25 Beta R, QLwIP
>Client: Athlon 1.2 GHz, Realtek RTL8029AS PCI, RedHat Linux 9.0,
>Netscape browser
>
>Unfortunately I have no second PC so I can only measure from PC to Q60
>and vice versa.
I am sure I have such a card.
I will fit it in one of the nine PCs here tomorrow (5 run 24/7!) and
test.   It may not be until the evening.

What is Ethereal?  Will a stop watch do?

Have you played with the I2C stuff yet?
-- 
 QBBS (QL fido BBS 2:252/67) +44(0)1442-828255
 tony@.co.uk  http://www.firshman.co.uk
   Voice: +44(0)1442-828254   Fax: +44(0)1442-828255
TF Services, 29 Longfield Road, TRING, Herts, HP23 4DG



[ql-developers] HTTP server speed test

2004-02-15 Thread Peter Graf
Hi,

is somebody out there who owns two networked PCs, one with a 10 Mbit/sec 
ISA NE2000 (or clone) card, and could do a speed test using this PC as HTTP 
server?

With the Q60/QLwIP as server I currently get average 854 kBytes/sec HTTP 
throughput, measured with Ethereal, transferring a file larger than 10MB. 
It would be interesting how far from the optimal value this is still away.

My test setup:
Server: Q60/80, Realtek RTL8019AS ISA, QDOS Classic 3.25 Beta R, QLwIP
Client: Athlon 1.2 GHz, Realtek RTL8029AS PCI, RedHat Linux 9.0, Netscape 
browser

Unfortunately I have no second PC so I can only measure from PC to Q60 and 
vice versa.

Thanks,
Peter