Group Build at the Mason Dixon Hamfest

[wb4jfi]

Controlling an embroidering machine is a little different than receiving 
I/Q data at 192ks/s (each sample=32bits), or 6Mb/s, over a high-speed 
USB2 connection.  480k samples is even worse, with a data rate of over 
15Mb/s.

I'm not saying it won't work, just look out.  I have heard of other hams 
trying to run PowerSDR, a Windoze program, in a Windows VM under Linux, 
and it failed, due to USB compatibility problems.

USB!=USB between different OS's.  There can be a high amount of latency 
which on lower-speed (HID, CDC, even full-speed) USB devices does not 
cause problems.  But, given data rates of over 15Mb/s or more, the 
inefficiencies in all that API cross-talk/faking-it starts piling up.  
And, I think that there is a dramatic difference between usblib 
implementations between Linux and Windows.  What works fast in Linux, 
not so much in Windows.  It needs its own tweaking, at least that's what 
I've heard.

No first-hand knowledge on the above, as I've seen failures from people 
a lot smarter than myself.  Maybe just "there be dragons", but it's 
enough to keep me from the edge so far.  I have other things to waste my 
time on right now.

Anyone have any actual, real-world experience running a fast-data 
hardware device and software program for Linux on Windows, or vice 
versa?  A USB scope might be a good example, as the samples are coming 
in pretty fast...

Macs are a different issue., I think their implementations are 
easier/better.
Terry


On 10/26/2010 11:05 AM, Mike O'Dell wrote:
> Parallels on MacOS X also does a great job of forwarding USB devices
> to VMs. My sister-in-law runs her computer-controlled embroidery machine
> from inside Windoze-in-a-VM-baggie just fine.
>
> USB these days is pretty-much completely encapsulated by software APIs.
> even USB Class Drivers don't really see hardware except
> what's at the end of the "pipes" (a real USB construct).
> and since the USB transfers are highly-structured, only
> the payload, either async or isochronous, is what matters.
> yes, there are commands and responses, but the nature of
> USB constrains real-time events to isochronous transfers.
>
> from the standpoint of virtualizability, this is a huge
> improvement over the days when pins on a parallel port
> were used as GPIOs stroked in real-time by spin-loops
> in x86 ASM code. (see "bit-banging serial port")
>
>     -mo
>
>
> On 10/25/10 7:44 PM, Louis Mamakos wrote:
>>
>> On Oct 25, 2010, at 6:23 PM, wb4jfi wrote:
>>
>>> On 10/25/2010 5:44 PM, Louis Mamakos wrote:
>>>> On Oct 25, 2010, at 3:23 PM, Frank Gentges wrote:
>>>>
>>>>> In addition to assembling the boards, we had the Quisk HF radio 
>>>>> software up and running as a demonstration that also caught 
>>>>> peoples attention.  We are now getting a version of Ubuntu that 
>>>>> can be run that includes the files for Quisk and the Charleston 
>>>>> receiver.  We should be able to distribute DVD disks that will 
>>>>> bring up Charleston/Quisk/Ubuntu on a computer either as a live 
>>>>> load or as a hard drive install.  AMRAD member Rob Bowers has been 
>>>>> working on this and a big tip of the AMRAD hat to Rob.   Stay tuned.
>>>> What would be slick is if this could be also available as a VMware 
>>>> virtual machine image of an installation of Quisk on an Ubuntu 
>>>> operating system.  For those that don't run Linux on their desktop, 
>>>> it would be an easy way to use the software.  There's a free VMware 
>>>> Player<http://www.vmware.com/products/player/>   that could be used 
>>>> on Linux and Windows to run VMware appliance images.
>>>>
>>>> louie
>>>> WA3YMH
>>>>
>>>> _______________________________________________
>>>> Tacos mailing list
>>>> Tacos at amrad.org
>>>> http://www.amrad.org/mailman/listinfo/tacos
>>>>
>>>>
>>> I'm not sure that this would work, since the Charleston board uses a 
>>> high-speed USB connection to Quisk.  The software will probably 
>>> function, but the USB drivers may not play right.  I've heard of 
>>> other USB-based devices having trouble in virtual machines.  But, I 
>>> do not have any first-hand knowledge.
>>> Terry
>>> WB4JFI
>>>
>>
>> Good point, I don't know if the VMware Player exposes the USB 
>> peripherals.  I do know that VMware Fusion on the Mac does a great 
>> job of exposing USB peripherals.  I routinely run Windows XP inside a 
>> VM, and use USB serial devices as well as specialized CCD cameras for 
>> astrophotography, also connected via USB to the host (and into the 
>> VM).  At least in the environment, I'd expect it ought to work..
>>
>> louie
>> wa3ymh
>>
>> _______________________________________________
>> Tacos mailing list
>> Tacos at amrad.org
>> http://www.amrad.org/mailman/listinfo/tacos
>
>