Vampire Accelerator for NeXT hardware project!

[Rob Blessin Black Hole]

Hello NeXT Community:  I would like to announce the porting of the Vampire Accelerator for Amiga to NeXT hardware . Greg Casamento says he can probably  do it but would appreciate help porting to  fpgm  . It basically sucks energy right through the 68040 processor , this may increase our speed 300 to 400 % wow ..... who with us! http://www.apollo-accelerators.com/#top  After watching the videos holy moly I'm watching the doom demo on Amiga, you can change it to full screen from postage stamp and it is fluid on a friggin Amiga on NeXT it ought to be off the chain , changing our resolution to digital.... Best regards Rob Blessin http://wiki.apollo-accelerators.com/doku.php. I'm also advised by the Nitro developer for this to work the Vampire would need to have or implement  MMU support.

[rooprob]

Let's do it!

[Rob Blessin Black Hole]

Hello Rooprob: Heck yes, it has me so excited so we need to figure out how to add MMU support which NeXT Uses and the current Vampire does not. I'm trying to contact Vampire Developers to see what would carry over. I think for example with previous and some of the amazing hardware mods we have proven anything is possible in this forum .  The availability of these boards may also be an issue but they have schematics and I'm wondering with guys like Steve over at Bigmessofwires and Mike at codesrc there may be interest in a Macintosh version.   Also Darkmatter so let me see if we can get more support.

Also I found some discussion on this as well http://www.nextcomputers.org/forums/viewtopic.php?t=3565&sid=c4f7e24e66c9ff7df7409a71c9bd0b89    

And a version that has MMU http://linuxdevices.io/fpga-softcore-gains-mmu-runs-full-linux/

[wayne123]

I know there is a new version expected in the pipeline, version 4 is up for preorder.

 If the Vampire adaption does not pan out you might try to get in touch with Mike of the FPGA Arcade Replay board, maybe he could give you some ideas of what is possible. Mike has a 060 add on designed for his FPGA board but not too long ago he was discussing using a Raspberry Pi compute module for even better performance.

[Rob Blessin Black Hole]

I've been researching and I'm wondering if we would be able to create a DSP to FPGA accelerator dongle for NeXT hardware ...

https://www.altera.com/en_US/pdfs/literature/wp/wp_dsp_fpga.pdf

or a breadboard expansion board for the Cube  ..... I'm brainstorming may be it would be easier to have or develop a mini intel computer FPGA board that is programmed to run NeXTSTEP/ Openstep and small enough to fit under the hood of a slab or in a cube expansion port  ..... reality distortion field .

The Vampire boards literally piggy back on the 68K processor but currently don't have the MMU coding needed for a port , it is like I envision it ....
droping some modern horsepower and graphics and wifi onto our legacy NeXT boxes would rock  the technology on the hardware side now exists , it is a matter of programming a software bridge to make it happen , my guess is a Linux port would also be possible . https://www.bittware.com/fpga/intel/boards/
http://wonderfulengineering.com/10-best-fpga-boards-for-engineers-and-hobbyists/

Video on the Vampire board with demos! https://www.youtube.com/watch?v=F-egp3pWr1w

[Rob Blessin Black Hole]

Hello NeXT Community:   May be something like this https://www.bigmessowires.com/2014/11/04/breadboarding-the-68k/  I have a lot of NeXT Motherboards and am wondering if in reality distortion field theory we create a breadboard NeXT 68K test accelerator breadboard box and like the vampire it drops in on top of the 68040 processor or like the nitro is a daughter card the plugs into socketed NeXT motherboards.

I know this won't happen overnight would our best bet be a daughter card like the vampire  that interfaces directly with the cpu. A DSP Accelerator card using the DSP port like the original Daydream dongle or an intel expansion board for cubes  that runs NeXTstep / Openstep or

Would it be possible to do acceleration in software over ethernet .

[nextchef]

With the progress of emulation like Previous, I do not really see the benefit in these sort of upgrades to the original hardware.  I just think the limited amount of developer time would be better spent on Previous and possible some updated drivers for NS/OS intel to support newer hardware.

But that's just me..

[Rob Blessin Black Hole]

Hello: I'm hoping that with the experience and coding in  Previous would make porting the  software to the accelerator easier.  As a NeXT owner, who wouldn't want to drop a hemi in it he he he this ones for Steve Jobs .  Best regards Rob Blessin

[nextchef]

At this point I would rather drop in a complete replacement for the next internal hardware than trying to create some sort of accelerator card or adapter.  A system built on current hardware running Previous will end up being faster and probably give a better experience.  Add to this some of the adapters allowing us to use original peripherals over usb in order to get the full NeXT experience.

I could also see incorporating the previous emulation code with a GNUstep based distribution to give us the ability to run modern apps like firefox along side classic NeXT apps.

This seems like the better direction to go than to start modifying the dwindling supply of original hardware.

[Rob Blessin Black Hole]

Hello NeXT Chef: I have seen a lot of Cube and Station mods , in fact some amazing ones . What I like about the Vampire is the results they are getting on Amigas , it is mind blowing , they are getting 300 to 400 % increases in performance , DDR Ram and HDMI Video and Wifi , all whilst sucking the software through the original 68040 25MhZ processor  which they also boost to 68080 speeds..... apparently the 68080 fixed a lot of compatability issues that made the 68060 buggy.  I've been taking a look at the FPGA software so everything has to do with creating gates and we may be able to use the experience with previous code to port it over.  NeXT hardware is not dwindling
, a lot of the software pipes may be in place because of the Amiga work just a matter of linking up to them with NeXT protocols.  We have some of the most talented developers and engineers in the world IMHO including you on this forum as time goes on this will become easier to do . Also many are retired and puttering about looking for something fun to do , looking to get back on the horse that brung them so to speak , this may be the project . I think it has parallels to what is happening in restoring cars, their are the purests  and their are the modders those that want to push the envelop  that is what is cool about NeXT . We have had the  honor of being way out in front  from Steve Jobs Vision so had he access to FPGA back in the hey day of NeXT  in our reality distortion time machine , we can almost hear him now cracking the whip , we'll get her done , we are falling behind in hardware performance to the Amiga guys .    Yes a prime example is previous and all of the amazing hardware projects Brian Archer and others are doing!

[nextchef]

I totally understand the desire to "scratch an itch" so to speak, and if this excites someone and becomes their labor of love then more power to them!

It's not the direction I would go in, but of course I will still be interested to see how it progresses.

[bheron]

I totally understand the desire to "scratch an itch" so to speak, and if this excites someone and becomes their labor of love then more power to them!

It's not the direction I would go in, but of course I will still be interested to see how it progresses.

For me it's more of a labor of love.  What i want to start with is simply replacing the CPU with a daughterboard and an FPGA with a 680x0 core loaded onto it.  What i would do from there is to progressively have the board take over more and more functionality from the original NeXT hardware as needed.

At this point I know Previous is doing a pretty good job at emulation, but, for me, I really love the original hardware.   I suppose what might eventually result from this is a system on a chip or a mini re-implementation of the entire system (probably not, but it's possible)... which I, personally, would find really interesting.

Mostly this is because I want to exercise my hardware muscles and learn a few things along the way.

[bheron]

I totally understand the desire to "scratch an itch" so to speak, and if this excites someone and becomes their labor of love then more power to them!

It's not the direction I would go in, but of course I will still be interested to see how it progresses.

I have come back to this after a number of years.  I have been busy with ${DAYJOB}, GNUstep as well as other things.   Furthermore, I am thinking about doing an FPGA SOC.   Sort of a NeXT system on a chip, but able to interface with newer hardware.   As the lead of GNUstep, It's important to me to have working NeXT systems, but they are disappearing.   So this seems like a great challenge as well as something to benefit the community.

[spitfire]

So I'd love to help. But unfortunately I'm a SW guy. Yes I've done Linux kernel work, but never done any HW beyond blinking LED lights. But if there's any way I can help I will.

That said here's my thoughts. MisTER is the right place to start imho. It's already a well supported platform with 680x0 systems emulated. None of them use MMU though, so whatever you do you'll be blazing a trail there.

However, Most of the NeXT system was commodity parts. NCR 5390 SCSI, commodity serial, floppy controller, etc.
With a bit of glue for linking together the 68K. A number of the other systems running on MiSTER actually implement the same SCSI/floppy/serial controllers NeXT had. So at least on the peripheral chip side of things we can reuse some of the work already been done.

One nice thing with NeXT is that it is a BSD Unix. So its not completely weird and esoteric.

So how do we start? with one of the M68K cores trying to boot a kernel image?

[zombie]

This 68080 just came out:

http://www.apollo-core.com/features.html