Quote:

No, just include the post and the thread.

Or you could just scroll down (or up, depending on your prefs) a bit.........



On the list of the most pointless targets for a MorphOS port (that will never happen anyways) a Mediator+Sonnet comes in below:

X1000
SAM440
Yikes (B/W PowerMacG3 and early PowerMacG4)
G3 PowerBook
G3 iBook
Beige G3 PowerMac
603/604 Macs (those that the Sonnet was supposed to go in)
BeBox
Mac-Clones
TeronOne


Edith:
The Sonnet is supposed to go into 601 based Macs, feel free to list those below the Mediator+Sonnet combo


[ Edited by Kronos 29.05.2019 - 21:38 ]

> What we'd really need is Sonnet PPC support. I missed a chance
> to buy an 800 MHz board last year.

Sonnet PCI cards (i.e. Crescendo 7200) max out at 500 MHz. Faster cards supported by SonnetLibrary are not Sonnet cards but various PMC cards to be installed on a PCI carrier board.

Quote:

The card I was offered via eBay was listed as a Sonnet card pulled from a Mac clone.
Other Sonnet PPC cards were also available.

If the listing was incorrect, at least it made me aware of the availability of 800 MHz card. I you now state that there were 1 GHz cards.

Actually, the few G4 level cpus still available range around 1.3 GHz (unless you include the 7448).

>> Sonnet PCI cards (i.e. Crescendo 7200) max out at 500 MHz.
>> Faster cards supported by SonnetLibrary are not Sonnet cards
>> but various PMC cards to be installed on a PCI carrier board.

> The card I was offered via eBay was listed as a Sonnet card pulled from a Mac clone.

Only PCI cards (or non-PCI cards on a PCI carrier) are of use in a PCI-enabled classic Amiga. Faster than 500 MHz Sonnet cards are not PCI but for proprietary Apple connectors.

> If the listing was incorrect, at least it made me aware of the availability of 800 MHz card.

If the listing claimed it was a PCI card, then it definitely was incorrect. Overall, non-PCI Sonnet cards go up to 1.8 GHz.

> I you now state that there were 1 GHz cards.

As I wrote, these are not Sonnet PCI cards but various (non-Sonnet) PMC cards to be installed on a PCI carrier board. The support library is called SonnetLibrary simply because it all started with Sonnet PCI cards. Most cards currently supported by SonnetLibrary are non-Sonnet cards, though.

> the few G4 level cpus still available range around 1.3 GHz (unless you include the 7448).

None of these Sonnet cards are PCI so irrelevant for PCI-enabled classic Amiga.

Quote:

So provide me some info about the PMC cards.
And the Mac compatible the card was listed along with Sonnet cards, and was listed as PCI, not a proprietary interface.

But again, could be wrong.
As to the 1.3 GHz cpus I was referring to, that's just the stock of processors NXP has left (outside of the 7448). Its surprisingly very few.

So, references for the PMC cards and the carriers?

No just provide the number of the post.

> the Mac compatible the card was listed along with Sonnet cards, and was listed as PCI,
> not a proprietary interface. But again, could be wrong.

Wrong for sure. The only Sonnet accelerators for the PCI slot are the Crescendo 7200 cards, which, as said, max out at 500 MHz. There's a Sonnet card series called Crescendo/PCI G4 available at 700, 800 and 1000 MHz, but despite the name, these are not PCI cards. Maybe it was one of these?

> As to the 1.3 GHz cpus I was referring to, that's just the stock of
> processors NXP has left (outside of the 7448). Its surprisingly very few.

This shouldn't be relevant as nobody will seriously want to restart production of any of the compatible cards anyway.

> provide me some info about the PMC cards. [...] references for the PMC cards and the carriers?

See there (info is somewhat cluttered):
http://aminet.net/package/driver/other/SonnetLibrary
https://github.com/Sakura-IT/SonnetAmiga
https://github.com/Sakura-IT/SonnetAmiga/wiki
http://eab.abime.net/showthread.php?t=76633 (start at comment #1447 on page 73)

Quote:

Most likely, I should have noted the make and model of the computer.
It's entirely like they implemented Apple's CPU slot.

As to the 7200s, reverse engineering them would be a chore. Although it should be possible to add the missing multipliers.

Still, from what I have seen, these things are a PITA to program.

Quote:

Considering that we are still supporting the 400MHz G3 level
Efika, does any of this matter?

Hey a Mediator equipped Amiga with a 500 MHz Sonnet card might be faster than the Efika.

Accelerated Amiga's hold some kind of weird form of nostalgia, but realistically I can't see using anything before the 1.42 GHz G4.

The fact that a 1.33 GHz accelerator could be built is mildly interesting, but I'd still rather have a late 2005 G5.

But hey, isn't the TeronOne almost the same as a Pegasos I?
And the only reason the X1000 has never gotten any love is because of Ben Hermans' involvement (because, let's face it, it has many advantages over an X5000).

[ Edited by Jim 07.06.2019 - 23:17 ]

Quote:

It might, as benchmarks proved that 330 MHz (overclocked by Stachu100) 630e on a BlizzardPPC seemed to be almost as fast as Efika in some cases.

It might be faster than an Efika, but I somehow doubt the Efika would be seen as a viable target if it was released today.

All examples I listed *could* get a MorphOS port, but none would yield a reasonable number of registrations and none is related to already supported HW closely enough to make it trivial (the TeronOne has a buggy SB on top of the buggy NB it shares with the Pegasos1).

The Sonnet/Mediator is way down the list cos:
- HW is even harder to get by than anything else on the list
- requires an support OS on the 68k side to access anything that isn't on the PCI bus

> G3 level Efika

It's G2 level even:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=11&topic_id=11137&start=744

> a Mediator equipped Amiga with a 500 MHz Sonnet card might be faster than the Efika.

A 500 MHz G3 trumps a 400 MHz G2, yes.

> The fact that a 1.33 GHz accelerator could be built is mildly interesting

Why would it be restricted to 1.33 GHz?

> isn't the TeronOne almost the same as a Pegasos I?

Yes, the latter is based on the former, with some enhancements like newer southbridge.

> the only reason the X1000 has never gotten any love is because of Ben Hermans' involvement

For some years, and as Hermans left A-Eon prior to the X1000 release, the fact that the PA6T and thus the X1000 are no longer and won't ever be available new might have been the primary reason.

The primary reason I've mentioned the 1.33 GHz CPU is that it's still available and would be the easiest to adapt to that design.

But it is a bit pointless, and as I've mentioned before,there are other platforms that make more sense.

> The primary reason I've mentioned the 1.33 GHz CPU is that it's still available

...as are 1.42, 1.6 and 1.7 GHz ones:

https://www.rocelec.com/part/FREFREMC7447AHX1420LB
https://www.rocelec.com/part/NXPFREMC7448VU1600LD
https://www.rocelec.com/part/NXPFREMC7448VU1700LD
https://www.arrow.com/en/products/mc7448vu1700ld/nxp-semiconductors
https://www2.mouser.com/ProductDetail/NXP-Freescale/MC7448HX1700LD?qs=%2Fha2pyFaduhE9Nfwsy6ij6rTF5HFgDV0kLLfe4p6HpYy74DI%2FPAkHQ%3D%3D

Question is why anyone would use an old e600-based CPU instead of a newer, faster and cheaper e6500-based CPU for newly created accelerators.

> and would be the easiest to adapt to that design.

To which design? Besides, it should be easier to adapt a board design to the CPU than the other way round ;-)

Quote:

From a design standpoint, I don't think an e6500 or e5500 cored CPU would be easier to implement.
The e600 cored cpus are much more similar to the cpus used on older PCI boards like those we have been discussing.
And the software would be the same. No need for separate development.

But, again, I'd still favor something that wasn't plugged into 25+ year old technology.

> From a design standpoint, I don't think an e6500 or e5500 cored CPU
> would be easier to implement.

Why not? The SoC already incorporates the memory controller and other northbridge functionality, and there'd be no need to route traces from pins of unneeded on-chip peripherals.

> The e600 cored cpus are much more similar to the cpus used on older
> PCI boards like those we have been discussing.

None of the designs of those older PCI boards are open, so they would have to be designed from scratch anyway.

> the software would be the same. No need for separate development.

The same as what? Not separate from what? It's not like MorphOS already supports older PCI-based accelerators.
And e6500 support shouldn't be that hard given that MorphOS already supports both AltiVec and e5500.

> New hardwware supported as iMac G5 iSight [...]? Maybe a PowerMac G5 Quad
> with PCI Express graphic card will be supported?

Well, the MorphOS 3.12 release notes state the following:

"Added basic PMac12,1/PPC970FX frequency switching support [...]
Basic sensor support for PMac11,2/PMac12,1 models [...]
Added dynamic CPU frequency switching (full/half speed as supported by 970FX) for PowerMac12,1/iMacG5 based on CPU load statistics"

None of these features would make sense being mentioned if the public MorphOS 3.12 release didn't run on those machines, I guess ;-)

I'm having trouble installing 3.12 from a USB stick on my 1.67GHz PowerBook. I have followed the instructions on the main MorphOS website for preparing the USB stick, then placed it into my Windows computer and copied the SD Card/USB Image to the empty USB stick, after renaming it morphos.iso, but when I place the USB stick in my PB and try to open the morphos.iso file by double clicking it (as instructions say), I get an error message saying "No hexview yet for Usb:morphos.iso"

I'm sure this is something simple I am over looking, or have forgotten, but if I'm following the instructions on the official site and running into this problem, perhaps there are others having the same problem?

Edit: I think I have figured it out, and I am now downloading the ISO image, not the USB image, even though I will be installing from a USB stick. Maybe someone can explain to me what the SD card/USB Image downloads are for, if installing from a USB stick uses the ISO images? Is it just that the instructions for installing from a USB stick need to be updated?

Edit2: Okay, now I have the MorphOS.iso downloaded and copied to the USB stick correctly (I think), but when I open it and copy the "boot.img" from the mac_ppc32 directory, I can't paste it to the USB stick's main partition. In fact, I can't paste the file to any hard drive, or partition on my MorphOS system. I just get an error message that says "Error while copying '@1:mac_ppc32/boot.img'". Any ideas why this file won't copy to my USB stick? I can copy and paste the readme file to the USB stick main partition.

[ Edited by amigadave 07.10.2019 - 05:34 ]

Quote:

Why copy the SD Card/USB Image to USB stick? Just use the normal LiveCD .iso file

Quote:

Because I am trying to install from a USB stick, so naturally, when I went to the Download section and saw a USB Image as one of the two choices for downloading, I chose the USB Image. Seems logical to me! Going back to the installation instructions for using a USB stick, I then saw that it says "download the ISO image" so I went back and downloaded the 3.12 .iso image.

Edit3: Finally got the installation going. I had to "Mount" the iso image, to be able to copy the "boot.img" file from the mac_ppc32 directory. Just double clicking the iso image as instructed on the page for USB installation, did not work for me.

[ Edited by amigadave 07.10.2019 - 05:47 ]

Quote:

The SD/USB image is used when booting on Sam 460 or Efika (as stated here). I guess the USB installation instructions in MorphOS Library article could be more precise.

Quote:

Installation completed, and I'm replying from my PowerBook now. I need to find the correct registration keyfile for this computer, as it appears that it has been deleted or corrupted along the way some months ago.

Quote:


I had exactly the same issue - the instructions could be a little better. I've installed to a 1.67 Powerbook too, the DLSD model - how are you finding it? I have frequent lockups and crashes.

That certainly does not sound normal. Care to share more details?