On Mon, 17 Sep 2012, Gene Heskett wrote:
> Date: Mon, 17 Sep 2012 11:39:57 -0400
> From: Gene Heskett <[email protected]>
> Reply-To: "Enhanced Machine Controller (EMC)"
> <[email protected]>
> To: [email protected]
> Subject: Re: [Emc-users] AMD based Mini-ITX Mobo, Good option?
>
> On Monday 17 September 2012 11:19:30 Peter C. Wallace did opine:
>
>> On Mon, 17 Sep 2012, Joe Spanier wrote:
>>> Date: Mon, 17 Sep 2012 07:47:53 -0700 (PDT)
>>> From: Joe Spanier <[email protected]>
>>> To: "[email protected]"
>>> <[email protected]> Subject: [Emc-users] AMD based
>>> Mini-ITX Mobo, Good option?
>>>
>>> Hey Guys! I found another motherboard for a great price and I was
>>> curious on your opinions. Heres a link
>>>
>>> : http://www.newegg.com/Product/Product.aspx?SID=0cJiSgDVEeKnXe4ky_MqH
>>> :A0_Ic8B3_0_0_0&AID=10440897&PID=1225267&nm_mc=AFC-C8Junction&cm_mmc=A
>>> :FC-C8Junction-_-cables-_-na-_-na&Item=N82E16813128561
>>>
>>> Its AMD based with the new APU tech. Has anyone out here tried one of
>>> these yet? Or have any ideas whether it would be a good replacement
>>> for the D525 boards?
>>>
>>>
>>> BTW, Is there a good way to search these emails or the archive? I Cant
>>> for the life of me find a search function in the archive and there
>>> seems to be a alot of great stuff hidden in these email lists.
>>>
>>> Thanks,
>>> Joe
>>
>> I got a ASUS E35M1-M for testing with the same on card CPU/APU and its
>> fine (~20 usec latencies)
>
> I would not call ~20 u-sec latencies comparable to the pair of D525MW's I
> have. They are showing circa 3 u-secs, with a very occasional, separated
> by hours, exception that will throw a realtime error. This is at 23 u-sec
> BASE_PERIOD. Reseting BASE_PERIOD to 25 u-secs absolutely stops that.
> Considering that 25 is 20 better than what I had to start with using the
> older xP1600 athlon, I'm ecstatic.
If you actually test latencies with I/O going on, I think you will find the
that D525s are nowhere near 3 usec real latency. The latency test does no
I/O so misses things that block hardware access. I've measured in the
neighborhood of 75 Usec jitter (with a 'scope) on the servo thread with
D525 that reported ~12 uSec on the latency test...
This also shows up with the inability of a D525 to run with much more than a
1.5 KHz servo thread without occasional real time faults. The APU boards are
better in this respect that the D525 (less servo threadjitter).
All this to say, the latency test is not a realistic test of LinuxCNC
operating latency, it will detect gross problems and give you a feel for the
machince latency performance but thats about it.
>
> However, since the D525MW is now mostly out of the pipeline, I agree that a
> suitable 'next-gen' board needs be found.
>
> However, I am also concerned that we are running out of time with the
> 10-04.4 LTS distribution, particularly since we seem to be locked into a
> kernel whose uptime on this machine has only exceeded 3 weeks of uptime
> once in close to a year. Its scheduler for normal desktop usage sucks dead
> toads thru soda straws, and it leaks memory, often into swap on a 4 core,
> 4Gb machine in an hour after the reboot. This morning seems to be an
> exception, no swap at 5 hours uptime, but when I rebooted, it took 2
> reboots as usual.
>
> I have been tempted to add a 6 hour cycle to the root crontab that does a
> swappoff -a, sleeps 2 secs, and re-enables swap.
>
> Has anyone else been tempted to control the leakage by this means?
>
>> Seems a little faster overall than the D525
>> at the expense of somewhat higher power consumption (the ASUS E35M1-M
>> has passive cooling and the heastsink gets pretty hot)
>>
>> Peter Wallace
>> Mesa Electronics
>>
>> (\__/)
>> (='.'=) This is Bunny. Copy and paste bunny into your
>> (")_(") signature to help him gain world domination.
>
>
> Cheers, Gene
> --
> "There are four boxes to be used in defense of liberty:
> soap, ballot, jury, and ammo. Please use in that order."
> -Ed Howdershelt (Author)
> My web page: <http://coyoteden.dyndns-free.com:85/gene> is up!
> A day without sunshine is like night.
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> Emc-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/emc-users
>
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
