From: Erik D. R. <eri...@ce...> - 2006-05-12 18:12:21
|
Some library like you mentioned would be primo! My specific usage = involves all the floating point calculations in my neural network = (Including the iLab toolbox). I have been banging my head against the = wall trying to figure out how to modify the calculations to work with = fixed point. Alas, I am not a mathematician. Just a dumb little = programmer.=20 =20 You get that next-gen Gum out and I will be one of the first customers. = Until then, anyone have a library that would help on this? ________________________________ From: gum...@li... on behalf of Craig = Hughes Sent: Fri 5/12/2006 1:17 PM To: gum...@li... Subject: Re: [Gumstix-users] iLab Toolkit/Visual Processing We're working on next-gen gumstix stuff now; I think actually that=20 the main contenders at this point for next-gen CPU don't have FPUs on=20 them. As someone said to me the other day though "If you know what=20 you're doing, fixed point math will always be faster anyway though". =20 My reply was "Yeah, but show me the open source lib which saves me=20 having to code it myself, man!" I think the topic was actually MP3=20 encoders -- incredibly, there doesn't seem to be a decent fixed-point=20 open source MP3 encoder out there. C On May 11, 2006, at 10:19 AM, Erik D. Rodriguez wrote: > Ahh... yeah... You know, that stood out and poked me in the eye=20 > when I was considering the Gumstix. For some reason, I didnt pay a=20 > great deal of attention... Don' need no stinkin FPU!!! Argh,=20 > yeah... Hmmm. That is going to have a pretty big impact even if I=20 > dont use iLab. Most my neural network code is floating point. Any=20 > chance of a FPU enabled Gum? Either way, no HUGE loss as the Gum is=20 > more than capable of dealing with most of the primary functions. > > ________________________________ > > From: gum...@li... on behalf of Craig=20 > Hughes > Sent: Wed 5/10/2006 12:54 AM > To: gum...@li... > Subject: Re: [Gumstix-users] iLab Toolkit/Visual Processing > > > I haven't used it, but I notice that this is a floating-point=20 > library, and since there's no FPU on the gumstix, it's going to run=20 > EXTREMELY slowly. If you want to run code on the gumstix, you=20 > should probably look for a fixed-point library instead. > > C > > On May 9, 2006, at 10:40 AM, Erik D. Rodriguez wrote: > > > I am currently working through the object model of the iLab=20 > Toolkit (See http://ilab.usc.edu/toolkit/) classes trying to=20 > determine which classes I need for my project, and which would work=20 > on the Gumstix. Then it struck me, perhaps one of you have used=20 > this? (And would be willing to share information/Examples) If not,=20 > have any of you used any other visual processing software? I am=20 > trying to use an established toolkit to allow my robot a basic (Or=20 > better) visual understanding of its environment. iLab looks like a=20 > very useful and competent toolkit, but it also appears incredibly=20 > processor intensive (And very large!). > =20 > I am not really looking for any form of image recognition, but=20 > visual path and object detection are a concern. Other things such=20 > as object tracking, and path projection would be useful. > > > <winmail.dat> ------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, = security? Get stuff done quickly with pre-integrated technology to make your job = easier Download IBM WebSphere Application Server v.1.0.1 based on Apache = Geronimo http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=3D= 121642 _______________________________________________ gumstix-users mailing list gum...@li... https://lists.sourceforge.net/lists/listinfo/gumstix-users |