On Tue, 8 Feb 2011, Kent A. Reed wrote: > Date: Tue, 08 Feb 2011 10:51:18 -0500 > From: Kent A. Reed <knbr...@erols.com> > Reply-To: "Enhanced Machine Controller (EMC)" > <emc-users@lists.sourceforge.net> > To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net> > Subject: Re: [Emc-users] EMC2 and Ethernet [AT5NM10-I] > > On 2/7/2011 6:42 PM, Steve Blackmore wrote: >> On Mon, 7 Feb 2011 15:53:34 +0000, you wrote: >> >>> On 7 February 2011 15:42, Steve Blackmore<st...@pilotltd.net> wrote: >>> >>>> Try the Asus version - AT5NM10-I >>> That also appears to have the P-port as a socket rather than a header. >>> However, I wonder if it has the same chipset etc? If not, then the >>> latency performance might not be the same. >> Intel NM10 chipset. Integrated Intel Graphics, Atom D525. >> >> Same chipset AFAIK? >> >> Steve Blackmore >> -- >> > Gentle persons: > > Since I had one on hand, I just ran the latency test overnight on an > AT5NM10-I board, 2GB RAM, 20GB SSD, onboard graphics, keyboard/mouse > connected to PS2 ports, a fully patched EMC2.4.6, kernel > 2.4.6-32-122-rtai, default bios settings. > > dmesg reports the cpu is an Intel Atom D510 (not D525). > > dmesg reports the parallel port is in EPP mode at 0x378. (Truth in > advertising: I haven't hooked up to a CNC controller yet so I have no > way of knowing if there are any slips "twixt the fork and the lip" as my > grandmother liked to say.) > > I simultaneously ran several copies of glxgears, firefox, vlc movie > player, and took directory listings from attached usb hard drives. > > The relevant test numbers are: > > 1ms thread: max interval 1009084 / max jitter 12603 > 25us thread: max interval 29248 / max jitter 14321 > > This appears to me to be an acceptable motherboard for EMC2 use, even > for software generation of stepper timing. > > Regards, > Kent > > > PS - I'll post these numbers to the wiki in a while.
Could you try disabling Hyperthreading (a bios setup option) and setting the isolcpus=1 grub option, then running the latency test again? I think all the really good results from the Atom MBs have been from systems that have the Hyperthreading off and the isolcpus option Heres a link that shows the improvement on a Atom 330: http://www.anderswallin.net/2010/11/emc2-upgrade/ Peter Wallace Mesa Electronics ------------------------------------------------------------------------------ The ultimate all-in-one performance toolkit: Intel(R) Parallel Studio XE: Pinpoint memory and threading errors before they happen. Find and fix more than 250 security defects in the development cycle. Locate bottlenecks in serial and parallel code that limit performance. http://p.sf.net/sfu/intel-dev2devfeb _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
