Home  
Login
Username:

Password:

Remember me



Lost Password?

Register now!
Sections
Who's Online
63 user(s) are online (58 user(s) are browsing Forums)

Members: 0
Guests: 63

more...
Support us!
Recent OS4 Files
OS4Depot.net



« 1 (2)


Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@tekmage

"Digital" mixer.. definitely do-able.
The Xena chip has no analog, but it can handle serial streams with amazing speed.
If the I/O for this digital mixer was S/PDIF optical, then it also is completely isolated electrically, which plays well into the "blow it all up" conversation as well.

The X-1000 has S/PDIF input AND output, one is just a header, but I'll bet it could be "activated" and populated fairly easily. so one channel of mix IN could be from AHI, and the final mix COULD be streamed, saved, further processed as well.

Quite interesting.


   Report Go to top

Re: The Xena Research Project Part 1
Just can't stay away
Joined:
2006/12/4 23:15
Posts: 1909
@tekmage

Digital mixer is the first sane usage suggestion I've seen (well except for the x logger concept)


_________________
Blender For OS4.x : Blues : Walker Broad
   Report Go to top

Re: The Xena Research Project Part 1
Home away from home
Joined:
2006/11/20 16:26
From Norway
Posts: 2643
@LyleHaze

But most sound cards has "Line In" this days.

_________________
(NutsAboutAmiga)

Basilisk II for AmigaOS4
AmigaInputAnywhere
Excalibur
and other tools and apps.
   Report Go to top

Re: The Xena Research Project Part 1
Home away from home
Joined:
2006/11/20 16:26
From Norway
Posts: 2643
I steal struggle to see what advantage the Xorro slot has over the PCIe slot.

_________________
(NutsAboutAmiga)

Basilisk II for AmigaOS4
AmigaInputAnywhere
Excalibur
and other tools and apps.
   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/12/2 11:14
From South Florida, USA
Posts: 196
@LiveForIt

That is like asking. What is the advantage of a wrench over a screwdriver?

Answer; Depends.

@ Thread

Digital mixer.......I like it.


_________________
A1-X1000
   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@mechanic
Choosing between a wrench and screwdriver doesn't make sense if you only want to get hammered. :)

@broadblues
Thanks. Sentinel X-Logger has some potential, and I REALLY hope someone tries the X5000 extensions for additional speed. That would take the project beyond just eliminating the need for a second computer.

@thread
I have a few ideas occasionally. One that is more "Fun and Flashy" would be a driver for "NeoPixels". briefly, these are serial chains of individually addressed, full color RGB LEDs. Less wiring, more fun. But the signal format used is very quick, and NOT tolerant of delays. Even using a PIC or Arduino can be tricky if they must pause to support comms with the master...
The "unusual" ability of Xena to use any single port pin for very fast, well timed output is a natural match. and if someone was foolish enough to want a LOT LOT of Neopixels, multiple strands should be as easy as
running the same program in multiple threads.

Getting all that data from the Amiga to Xena for imaging would be much simpler if there was some dual port RAM between them, which the X5000 has.

Much could be done on an X1000, MORE could be done on an X5000
with NeoPixels.

It could make some REALLY "flashy" stuff to show off, and an ARexx port could make the display useful too.

Yes, just flashy lights.. but what could you make with them?

   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/12/2 11:14
From South Florida, USA
Posts: 196
This is strange, or maybe I have forgotten everything in four years of non-use.

The programs I wrote in xc back then load and work with the xmos tools supplied with 4.1FE, but after downloading and installing Xtimecomposer x86-32bit 11.11.1 and 14.2.4 64bit, not even the old xc source will run. Seems to compile just fine.

I think it has something to do with the xn file as it dosen't even set the xcore frequency.

I think I need a tutorial. I remember changes had to be made to the xn(?) file, but don't recall what they were.

Using XS1-L16A-128-QF124-C10, and the old XS1-L02-QF124-C5 for the 32bit.

_________________
A1-X1000
   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@mechanic

Yes, we must add the clock frequency in the xn file, or else it will make bad assumptions.

.xn files are short, I have pasted one below.

If memory serves the name given must match the filename exactly..
A-EON_Xena Device

///////////////////File begins AFTER this line///////////////////////////////
<?xml version="1.0" encoding="UTF-8"?>
<Network xmlns="http://www.xmos.com"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.xmos.com http://www.xmos.com">
<Type>Device</Type>
<Name>A-EON_Xena Device</Name>

<Declarations>
<Declaration>tileref tile[2]</Declaration>
</Declarations>

<Packages>
<Package ID="0" Type="XS1-LnA-128-QF124">
<Nodes>
<Node Id="0" InPackageId="0" Type="XS1-L8A-64" SystemFrequency="500MHz">
<Tile Number="0" Reference="tile[0]"/>
</Node>
<Node Id="1" InPackageId="1" Type="XS1-L8A-64" SystemFrequency="500MHz">
<Tile Number="0" Reference="tile[1]"/>
</Node>
</Nodes>
</Package>
</Packages>

<Links>
<Link Encoding="5wire" Delays="0,1">
<LinkEndpoint NodeId="0" Link="XLG"/>
<LinkEndpoint NodeId="1" Link="XLF"/>
</Link>
<Link Encoding="5wire" Delays="0,1">
<LinkEndpoint NodeId="0" Link="XLH"/>
<LinkEndpoint NodeId="1" Link="XLE"/>
</Link>
<Link Encoding="5wire" Delays="0,1">
<LinkEndpoint NodeId="0" Link="XLE"/>
<LinkEndpoint NodeId="1" Link="XLH"/>
</Link>
<Link Encoding="5wire" Delays="0,1">
<LinkEndpoint NodeId="0" Link="XLF"/>
<LinkEndpoint NodeId="1" Link="XLG"/>
</Link>
</Links>

<JTAGChain>
<JTAGDevice NodeId="0"/>
<JTAGDevice NodeId="1"/>
</JTAGChain>

</Network>

   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/12/2 11:14
From South Florida, USA
Posts: 196
@LyleHaze
Thanks Lyle, I'll give it a try tomorrow.

In xde 11.11.1 the C5 device uses 'tile[n]' and the C4 device is still using stdcore[n], strange.

Do you recall what changes we made to the makefile? Was it just the optimization?

Everything I had went away when my power supply went bonkers and took out both hard drives. I only have some stuff left on USB sticks.

Len


_________________
A1-X1000
   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@mechanic

The change from core to tile happened just before AmiWest 2012, I believe.
I did have copies of BOTH types of device file, so the one I posted may be either.

The only modification I remember was to add the clock information to the device file, so if the format has changed again try adding the clock information in a similar place..??

And if you modify an existing device file, you will want to rename it, so it doesn't later get confused with an unmodified file.
One benefit to beginng mine with "A-EON" is that we get promoted to the head of the alphabetised list.


   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/12/2 11:14
From South Florida, USA
Posts: 196
@LyleHaze

The modified .xn file seems to be AOK.

Something has changed in the xcc builder. What used to build just fine on the earlier tools now produces errors. However, the code builds just fine with no problems indicated while in xtimecomposer, and produces the proper waves in the simulator, but looking at the .xe file with a hex editor (near the end) are notations of errors in the .xc code.

I do not know if this is the reason for the xe file not even setting the frequency when loaded with XRunxe, something has changed.

Do you still have an X1000?

_________________
A1-X1000
   Report Go to top

Re: The Xena Research Project Part 1
Not too shy to talk
Joined:
2006/11/27 16:55
Posts: 266
@LyleHaze

Quote:
"Digital" mixer.. definitely do-able. The Xena chip has no analog, but it can handle serial streams with amazing speed.


Would it be possible to use the XMOS chip for mixing/switching digital video too? That could serve a number of different applications.

Would it be possible to slave another computer to the XMOS and control it directly with the keyboard and mouse or other input devices from within OS4.

Having the Vampire hardware on a Xorro card would be really nice if you could use the host X1000 or X5000's input devices to control it and switch or mix the video output of the the 2 systems.

   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@mechanic

Yes, I still have an X1000. My Xorro is away at the moment though.


@Rob
I don't know enough about digital video to give a good answer.
_both_ incoming VIdeo signals would have to be wired to the Xorro card, as would the resulting output.
The two video inputs would have to be synchronized, just like the TBC's needed on toasters, for the same reason.

I just don't know.. interesting idea though.

   Report Go to top

Re: The Xena Research Project Part 1
Not too shy to talk
Joined:
2008/5/2 12:14
From Sweden
Posts: 477
@daveyw

To emulate Altivec (at any useful speed), Xena would have to be about 4 times faster than the PPC CPU itself, which it isn't. More like 1/20..? It would also be able to transparantly intercept any Altivec instruction the CPU 'accidently' consumes. The CPU itslef would be MUCH better at emulating missing Altivec, just like the FPU, or other missing instructions, can be emulated on 68EC-machines.

Video mixing is probably also out of reach considering the data rate of 3Gbit/s just to maintain a 24-bit fullHD image on screen.

S/PDIF digital audio on the other hand should be perfectly doable.

_________________
Software developer for Amiga OS3 and OS4.
Develops for OnyxSoft and the Amiga using E and C and occasionally C++
   Report Go to top

Re: The Xena Research Project Part 1
Not too shy to talk
Joined:
2008/12/20 17:26
Posts: 258
I summed up some ideas for SAM's FPGA and X's XENA.
http://amigaworld.net/modules/newbb/v ... _id=27379&forum=33#463638


Cleanup for xena....

Xena ideas:
-3D render node (or adapter to slave node(s))
-DSP + digital sound out+in (delfina emulation, paula emulation, AAA audio, etc.) (*)
-second CPU controller (Cell/SPE/intel Atom + memory/legacyPPC)
-adapter to chip like xcore86 for i386 emulation
-GPS receiver
-EDGE/GPRS/3G receiver
-GSM base station
-WLAN receiver
-ADSL receiver
-ethernet router & firewall + port
-bluetooth receiver
-radio receiver
-Watchdog
-HardPowerOff (to truly shut off ATX power)
-Oscilloscope (*)
-LogicAnalyzer
-DataAcuistion (industry, home automation, car automation) (*)
-Retro ROM module reader
-PATA IDE controller
-SCSI controller
-Firewire controller/bus
-yet another USB port
-yet another RS232 port
-Zorro2 controller/bus
-A1200 clock port(s)
-Amiga/PC floppydrive controller (*)
-Bridge to Amiga 1200 "local bus" (original AmigaInc&Eyetech kludge idea)
-adaptor for original Amiga keyboards (for example for A500kb or as a version of Keyrah)
-keyboard switcher
-Car computer connections
-Home automation control bus
-MIDI Ports
-Amiga/Atari joystic ports (eight simultaneous digital joystics, etc.) (*)
-Playstation/Xbox Joystick ports
-PC joystick ports
-Parallel ports
-IR port
-Discolights controller
-video out -port (320x256)
-adapter to use indivision AGA, NatamiAGA, OCS,...
-Display Controllers (sevensegment, matrix LED display)
-MP4 encoder/decoder (or adapter to MPEG4 chip)
-video capture/digitizer
-multichannel digiTV receiver (and decryption, cable, terrestial and satellite) (*)
-iPad/iPhone adapter
-other future port/bus

The best ones (IMO) are marked with (*).


Edited by KimmoK on 2017/5/22 10:01:07
Edited by KimmoK on 2017/5/22 10:01:57
_________________
- Kimmo
--------------------------PowerPC-Advantage------------------------
"PowerPC Operating Systems can use a microkernel architecture with all it�s advantages yet without the cost of slow context switches." - N. Blachford
   Report Go to top

Re: The Xena Research Project Part 1
Just can't stay away
Joined:
2007/2/23 13:49
From Finland, the land of Santa, sauna, sisu and salmiakki
Posts: 1064
* Some Xmos guys have created C64 SID emulation. Also there's that MidiBox product where SID's are connected to each other to modulate each other. I'm wondering how many logical threads one SID chip emulation takes, if a single Xena chip could emulate many of them like in the MidiBox. So I could have a MidiBox in an X1000 without needing external hardware. The last time I checked MidiBox'es were not available. I don't know how to "digitize" the sound and play it back through AHI in such an emulator.

* Add a laser and light show controller and an analog+digital mixer to the "soft-MidiBox" Xorro card.

* Where's Transputer turbo card with thousands of logical cores for the Xorro (+PCIe) slot ?! (GPU's have thousands of cores. I don't know which one of the two GPU and Xmos is harder or easier to program for. Add power consumption differences and so on.)

* Thinking of that you have multiple xorro cards and needing to swap them back and forth frequently is not a temping thought. Create an external user port on a Xorro card where single/multiple ports are on the back of the computer with hot swapping capability and a protection circuit. (Yes I know USB/Firewire/Thunderbolt exists). So people can connect multiple devices at once and even hot swapping them without needing to open the machine. Maybe add a A1200 clock port with modern speeds to such a card (like KimmoK suggested).

I guess a PC is needed for the xc compiler ?

_________________
Rock lobster bit me, so excuse me.
X1000 + AmigaOS 4.1 FE. Be authentic.
"Anyone can build a fast CPU. The trick is to build a fast system." - Seymour Cray
   Report Go to top

Re: The Xena Research Project Part 1
Just popping in
Joined:
2006/11/24 14:19
From Georgia, USA. Just north of the Florida border
Posts: 207
@TSK

I designed a digitally controlled mixer that maintains the audio signal completely in the analog domain. "MidiBox Mixer"

No pots, no sliders, no static at all.

Small World. :)



I have considered making a proper "digital" mixer that works the same.
Xena would be far better suited for that one.


   Report Go to top

Re: The Xena Research Project Part 1
Not too shy to talk
Joined:
2008/12/20 17:26
Posts: 258
@TSK

"I guess a PC is needed for the xc compiler ?"

Last time I checked, yes.

I think there was some study on how to use cloud/x86 node to compile...

(possibly cheap intel tablet or "sick" could also do the compiling, most likely dosbox is not enough)

_________________
- Kimmo
--------------------------PowerPC-Advantage------------------------
"PowerPC Operating Systems can use a microkernel architecture with all it�s advantages yet without the cost of slow context switches." - N. Blachford
   Report Go to top


« 1 (2)



[Advanced Search]


Powered by XOOPS 2.0 © 2001-2016 The XOOPS Project