[En-Nut-Discussion] Heave load will harm the system?
ivesworking
ivesworking at hotmail.com
Tue Feb 28 22:17:43 CET 2006
Just back from to office, i just want to update the report, the system just
hang not burnning :p,
I believe should be my software flaw, however when i do the test again, it
seem notthing wrong, no hang, no traffic drop.
everything seem fine, this maybe rare case happen once a year, anyway this
is my latest test
http://img371.imageshack.us/img371/6876/trafficcr7ur.jpg
http://img379.imageshack.us/img379/4971/traffic2cr4nc.jpg
this test seem it able to handle up to 20 concurrent user, (however only 4
httpd been set)
i believe last time the system hang cause by a suddent memory drop, and not
able to recover from that drop,
however when i redo the test, it seem everything just fine.
WFLOH
----- Original Message -----
From: "ivesworking" <ivesworking at hotmail.com>
To: "Ethernut User Chat (English)" <en-nut-discussion at egnite.de>
Sent: Tuesday, February 28, 2006 8:33 PM
Subject: Re: [En-Nut-Discussion] Heave load will harm the system?
> the quesion still go back to the original point, as u say atmega128 will
> still run at idle, my question is if the whole MCU are processing without
> idle thread.
> for my case, i run xml phrasing when a http request been send from the pc,
> i using some http test tools to load the mcu.
> i was infront the development board when i do the 1min load test, and i
> feel a bit warm the normal after i do the load test,
> i was continue doing the test after i left the office using remote
> desktop, the test is the same, but this time i load the MCU a bit longer,
> 7 min,
> it work fine on the first 3 min, the suddenly all the traffic drop to 0, i
> am not sure about the mcu burn out or just a system hang because i still
> remote controlling my pc,
> i will report later see is a system hang or the worse case.
> idle case and the load case, the mcu temperature is different.
>
>
> WFLOH
> PS: i like side track question.
>
> below are some of my test result using cgi which need little processing
> speed, maybe some one will interested
> Receiving speed (kbits/sec)
>
> Profile
> 0:00:00- 0:00:10
> 0:00:10- 0:00:20
> 0:00:20- 0:00:30
> 0:00:30- 0:00:40
> 0:00:40- 0:00:50
> 0:00:50- 0:01:00
> Total
>
> 20connection
> 130
> 201
> 237
> 184
> 152
> 175
> 180
>
> 4connection
> 252
> 206
> 245
> 272
> 264
> 242
> 247
>
> 2connection
> 307
> 276
> 305
> 259
> 306
> 285
> 289
>
>
>
>
> Response time, sec
>
> Name
> Time
> 0:00:00- 0:00:10
> 0:00:10- 0:00:20
> 0:00:20- 0:00:30
> 0:00:30- 0:00:40
> 0:00:40- 0:00:50
> 0:00:50- 0:01:00
>
> 20connection
> Min
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
>
> 4connection
> Min
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
>
> 2connection
> Min
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
> 0.02
>
> 20connection
> Avg
> 0.2
> 0.19
> 0.2
> 0.25
> 0.29
> 0.26
>
> 4connection
> Avg
> 0.15
> 0.2
> 0.2
> 0.14
> 0.16
> 0.17
>
> 2connection
> Avg
> 0.06
> 0.07
> 0.07
> 0.08
> 0.07
> 0.08
>
> 20connection
> Avg90
> 0.9
> 0.91
> 0.96
> 1.08
> 1.13
> 1.03
>
> 4connection
> Avg90
> 0.85
> 0.98
> 1.1
> 0.77
> 0.82
> 1.08
>
> 2connection
> Avg90
> 0.29
> 0.36
> 0.34
> 0.44
> 0.42
> 0.45
>
> 20connection
> Max
> 3.05
> 4.08
> 4.03
> 4.04
> 3.03
> 3.2
>
> 4connection
> Max
> 5.05
> 4.0
> 6.98
> 3.08
> 4.06
> 5.99
>
> 2connection
> Max
> 2.11
> 2.08
> 2.03
> 2.84
> 3.1
> 3.04
>
>
>
> this is all program generate
>
> ----- Original Message -----
> From: "Harald Kipp" <harald.kipp at egnite.de>
> To: "Ethernut User Chat (English)" <en-nut-discussion at egnite.de>
> Sent: Tuesday, February 28, 2006 7:38 PM
> Subject: Re: [En-Nut-Discussion] Heave load will harm the system?
>
>
>>
>> At 13:12 28.02.2006 +0000, you wrote:
>>>running a webserver on the atmega128, and i use performance tools to load
>>>the system, if i continue load the server, will the mcu burn for
>>>overheat? anyone have experience on this?
>>
>> The ATmega128 is running full speed all the time. If all
>> threads are waiting for an event, the idle thread will
>> run. This can last for years without problem or any
>> heating up.
>>
>> The Ethernet Controller is a different story, as the
>> PHY part contains an analog power circuit. It may heat
>> up a bit during heavy Ethernet traffic. But there is
>> no danger of overheating of the 10 MBit Controller used
>> on Ethernut 1. The 100 Mbit Controller on Ethernut 2
>> becomes really warm, even in idle state. The Controller
>> on Ethernut 3 shows, that this is not mandantory for all
>> 100 MBit PHYs. It stays relatively cool.
>>
>> Harald
>>
>> P.S. I liked this question. Wonder why it hadn't been asked
>> before.
>>
>>
>> _______________________________________________
>> En-Nut-Discussion mailing list
>> En-Nut-Discussion at egnite.de
>> http://www.egnite.de/mailman/listinfo.cgi/en-nut-discussion
>>
> _______________________________________________
> En-Nut-Discussion mailing list
> En-Nut-Discussion at egnite.de
> http://www.egnite.de/mailman/listinfo.cgi/en-nut-discussion
>
More information about the En-Nut-Discussion
mailing list