Sonnet 7200 PCI programming - Page 3

strim · 17 March 2015, 12:13

Quote:

Originally Posted by jbenam

Amazing! How does the CyberPI result compare to one from a BlizzardPPC/CyberstormPPC?

A person having access to one of these boards could run it to calculate 10000 decimals and post the results

. I do have them but can't do it now.

Hedeon · 17 March 2015, 12:48

Numbers floating around on internet (for 100000) are:

50 minutes on an 25 MHz 68030
30 minutes on a 25 MHz 68040
15 minutes on a 50 MHz 68060
9 minutes on a 150 MHz 603e (Prototype board)
4.5 minutes on a 200 MHz 604e

And doing this on the Sonnet is 1.5 minutes.

But it's still all academic.

Promilus · 17 March 2015, 18:35

It's nice on synthetic benchmarks but won't be just as fast in real usage. On the other hand it still just might be faster than any PPC accelerator Amiga ever had. BTW the same project just might open the door for even more PPC WOS-compatible expansions on more or less exotic CPUs (I myself have several MPC8245-266 MCUs for anyone interested, BGA though).

strim · 18 March 2015, 10:06

Cybermand now also working on Sonnet. Completely blows away 68060:

Locutus · 18 March 2015, 15:26

This is some pretty cool Hackery.

It says a lot that Elbox never released this (even as a WarpOS/PowerUP only solution) while you guys seem to be doing fine :-)

alexh · 19 March 2015, 00:02

I imagine that the PCI libraries are in a lot better state than they were a few years ago?

strim · 19 March 2015, 00:07

The current state of PCI libraries is laughable at best. Special hacks are currently necessary for sonnet.library to function properly.

See this issue on GitHub for explanation on how this impacts our development...

jbenam · 19 March 2015, 11:02

Damn, you guys are even working on a pci.library reimplementation?

I should totally get a Mediator now

strim · 19 March 2015, 11:06

At this moment I wouldn't recommend anyone getting anything yet

...

matthey · 21 March 2015, 18:25

Quote:

Originally Posted by strim

The current state of PCI libraries is laughable at best. Special hacks are currently necessary for sonnet.library to function properly.

See this issue on GitHub for explanation on how this impacts our development...

Why not fix up and use the AROS 68k pci.library? There has already been some work done although I don't think it's fully working.

strim · 21 March 2015, 18:37

Quote:

Why not fix up and use the AROS 68k pci.library? There has already been some work done although I don't think it's fully working.

AROS 68k PCI library is actually based on what I previously wrote for NetBSD... and as you noticed it does not work. While at least my NetBSD drivers are (mostly) working.

The problem (as usual in such hobby projects) is lack of necessary time. Although Hedeon is still making steady progress with WarpOS reimplementation!

Locutus · 21 March 2015, 23:21

How complex in itself is WarpOS anyway? Seeing as there is almost no I/O i kinda expect it to be relatively simple, but thats always a dangerous assumption (-:

strim · 21 March 2015, 23:40

It is surprisingly complex, at this point sonnet.library is around 150kB of hand-crafted 68k and PPC assembly (and it's still nowhere near complete yet). Think of all necessary 68k-PPC synchronisation routines, message passing functions, scheduler etc.

API is quite well explained in documentation of original WarpOS, check it out.

Locutus · 22 March 2015, 10:01

Oh, interesting!

I grabbed the WarpUP 41 package from H&P's website (of which i was suprised its still up!) and it seems to include full documentation.

Something to read for over the boring work train commute :-)

jack-3d · 23 March 2015, 01:42

Wow, this is unbelievable progress guys! Nice project, please keep it up, this would be great if you achieve it working like PowerPCs cooperation like it is now on our Blizzards and Cyberstorms under WarpOS.

Hedeon · 23 March 2015, 23:31

Sorry for the shakes in the movie linked below, but here is where the software is standing at the moment (well, actually a week ago):

https://dl.dropboxusercontent.com/u/...etInAction.mp4

Primary focus is now to get a debugger going.

strim · 23 March 2015, 23:49

I've re-uploaded to YT. For me it is unplayable from Dropbox (too slow).

[ Show youtube player ]

musashi5150 · 24 March 2015, 08:22

Fantastic stuff

Cowcat · 24 March 2015, 12:35

This is awesome stuff.
What would happened if as Warpos.guide tells:

Quote:

"AMIGA-OS runs the first time completely on the PPC, using the 68K emulator. This emulator runs as task inside WarpOS, which has been made 100% 68K-independent."

Where this stuff went ??

Hedeon · 24 March 2015, 14:02

I could ask Sam Jordan :-P

Anyway:

https://groups.google.com/forum/#!to...S3s%5B51-75%5D

The fabled emul68k.library

Quote:

Originally Posted by Cowcat

This is awesome stuff.
What would happened if as Warpos.guide tells:

Where this stuff went ??

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
WANTED: PCI Sound Card VORTEX2 SuperQuard / SQ2500 or Turtlebeach SantaCruz pci card	keropi	MarketPlace	12	15 March 2017 02:48
Wanted:Sonnet Technologies QuadDoubler 50 (Doubler 4000)	kronuz	MarketPlace	0	16 June 2013 00:29
eBay : Sonnet Quaddoubler	alexh	MarketPlace	3	07 May 2013 00:50
Overlay problem on GeForce 7200	Leandro Jardim	support.WinUAE	1	10 August 2010 15:50
Prometheus PCI & Voodoo 3 PCI GFX Card	Slayer	support.Hardware	21	05 September 2006 10:57

17 March 2015, 12:48	#42
Hedeon Semi-Retired Join Date: Mar 2012 Location: Leiden / The Netherlands Posts: 1,993	Numbers floating around on internet (for 100000) are: 50 minutes on an 25 MHz 68030 30 minutes on a 25 MHz 68040 15 minutes on a 50 MHz 68060 9 minutes on a 150 MHz 603e (Prototype board) 4.5 minutes on a 200 MHz 604e And doing this on the Sonnet is 1.5 minutes. But it's still all academic.

17 March 2015, 18:35	#43
Promilus Registered User Join Date: Sep 2013 Location: Poland Posts: 806	It's nice on synthetic benchmarks but won't be just as fast in real usage. On the other hand it still just might be faster than any PPC accelerator Amiga ever had. BTW the same project just might open the door for even more PPC WOS-compatible expansions on more or less exotic CPUs (I myself have several MPC8245-266 MCUs for anyone interested, BGA though).

18 March 2015, 10:06	#44
strim NetBSD developer Join Date: May 2012 Location: Warsaw, Poland Posts: 411	Cybermand now also working on Sonnet. Completely blows away 68060:

18 March 2015, 15:26	#45
Locutus Registered User Join Date: Jul 2014 Location: Finland Posts: 1,175	This is some pretty cool Hackery. It says a lot that Elbox never released this (even as a WarpOS/PowerUP only solution) while you guys seem to be doing fine :-)

19 March 2015, 00:02	#46
alexh Thalion Webshrine Join Date: Jan 2004 Location: Oxford Posts: 14,330	I imagine that the PCI libraries are in a lot better state than they were a few years ago?

19 March 2015, 00:07	#47
strim NetBSD developer Join Date: May 2012 Location: Warsaw, Poland Posts: 411	The current state of PCI libraries is laughable at best. Special hacks are currently necessary for sonnet.library to function properly. See this issue on GitHub for explanation on how this impacts our development...

19 March 2015, 11:02	#48
jbenam Italian Amiga Zealot Join Date: Jan 2009 Location: Italy Age: 36 Posts: 1,910	Damn, you guys are even working on a pci.library reimplementation? I should totally get a Mediator now

19 March 2015, 11:06	#49
strim NetBSD developer Join Date: May 2012 Location: Warsaw, Poland Posts: 411	At this moment I wouldn't recommend anyone getting anything yet ...

21 March 2015, 23:21	#52
Locutus Registered User Join Date: Jul 2014 Location: Finland Posts: 1,175	How complex in itself is WarpOS anyway? Seeing as there is almost no I/O i kinda expect it to be relatively simple, but thats always a dangerous assumption (-:

21 March 2015, 23:40	#53
strim NetBSD developer Join Date: May 2012 Location: Warsaw, Poland Posts: 411	It is surprisingly complex, at this point sonnet.library is around 150kB of hand-crafted 68k and PPC assembly (and it's still nowhere near complete yet). Think of all necessary 68k-PPC synchronisation routines, message passing functions, scheduler etc. API is quite well explained in documentation of original WarpOS, check it out.

22 March 2015, 10:01	#54
Locutus Registered User Join Date: Jul 2014 Location: Finland Posts: 1,175	Oh, interesting! I grabbed the WarpUP 41 package from H&P's website (of which i was suprised its still up!) and it seems to include full documentation. Something to read for over the boring work train commute :-)

23 March 2015, 01:42	#55
jack-3d kLiker Join Date: Mar 2011 Location: Brno / Czech Republic Posts: 371	Wow, this is unbelievable progress guys! Nice project, please keep it up, this would be great if you achieve it working like PowerPCs cooperation like it is now on our Blizzards and Cyberstorms under WarpOS.

23 March 2015, 23:31	#56
Hedeon Semi-Retired Join Date: Mar 2012 Location: Leiden / The Netherlands Posts: 1,993	Sorry for the shakes in the movie linked below, but here is where the software is standing at the moment (well, actually a week ago): https://dl.dropboxusercontent.com/u/...etInAction.mp4 Primary focus is now to get a debugger going.

23 March 2015, 23:49	#57
strim NetBSD developer Join Date: May 2012 Location: Warsaw, Poland Posts: 411	I've re-uploaded to YT. For me it is unplayable from Dropbox (too slow). [ Show youtube player ]

24 March 2015, 08:22	#58
musashi5150 move.w #$4489,$dff07e Join Date: Sep 2005 Location: Norfolk, UK Age: 42 Posts: 2,351	Fantastic stuff

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)