We at vi-tek.dk are looking at selling complete systems only.
Targeting the danish market

_________________

@mx3

Add to that that there is a Level 2 cache in the 460ex which the 440EP lacked.

That's the problem with certain people, they don't do any basic research and then cover it up by preceding their erroneous posts with "AFAICT" or "AFAIK".

It is much more fun to create the impression of a rift between A-Cube and Hyperion unhindered by any factual knowledge.

_________________

@Hyperionmp

Now that the "460ex version is being held back for X1000 release" theory has failed the fantasists need something new to cling onto.

A while ago it was perceived (hoped) by the detractors, that the X1000 announcement was damaging to Sam440 sales, now the release of Sam460 is damaging future X1000 sales.

@Rob

Place is full of psychics. I still see a lot of posters saying they are holding out for the Amiga X1000.

_________________

Just woundering are we likly to see Aros on the system like we did the 440?

Plus it shows on there website Linux Debian running on the system, will that ship with system as well?

Plus http://www.sam4x0.com/images/Win95onSam460.png
how fast was windows 95 on it, just intresed really.

_________________
AmigaNG, YouTube, LeaveReality Studio

@amigang

Quote:

how fast was windows 95 on it, just intresed really.

DosBox window has 3000 CPU cycles = 80386DX performance.
I estimate that 460EX 1.15 GHz could be fast enough for 6000 CPU cycles.

Last edited by pavlor on 29-Jan-2011 at 01:48 PM.

Quote:

There are at least 2 technical reasons why the 440 kernel cannot work on the 460ex SoC: - the core used on the 460ex (440H6) has 36-bit physical addresses (none of the previous AOS4 kernels supported 36-bit)

The MMU in PPC440 core is defined as producing 36 bit physical addresses (see page 103 onwards of the PPC440 data sheet). While specific implementations may only physically expose a lesser number of address bits, the MMU-internal data structures are still defined for the full 36 bits. I doubt you want to suggest that Hyperion's contractors were abusing those four extra bits in Microsoftian style.... So no, not a convincing argument.

Also, you might want to check the 440EP datasheet, and compare it with the 460EX one. You will find that both mention "41-bit virtual address, 36-bit (64GB) physical address".

Quote:

- the Interrupt Controller is different

Not really, either. The 460EX has four "Universal Interrupt Controllers", at addresses 0xc0, 0xd0, 0xe0 and 0xf0. The 440EP has two of the very same Universal interrupt Controllers, at addresses 0xc0 and 0xd0. Thus, the interrupt hardware integrated into the 460EX is a strict superset of that found in the 440EP. And given that the two UICs on the 440EP provide 10 external interrupts, which was sufficient for the Sam440 line, it is quite possible that the extra 6 interrupts on the 460EX are unused. Or used for something which is not currently supported (sound comes to mind...).



But hey, let's be direct --- who made up the CD that's going to ship with these boards? Was it Hyperion, or was it Acube?

Quote:

Add to that that there is a Level 2 cache in the 460ex which the 440EP lacked.

And why in the world would an operating system care about a completely transparent L2 cache?

Quote:

That's the problem with certain people, they don't do any basic research and then cover it up by preceding their erroneous posts with "AFAICT" or "AFAIK".

So, Ben --- have *you* read the 460EX and 440EP data sheets? And I don't mean the first three pages only...

Last edited by umisef on 29-Jan-2011 at 02:11 PM.

@umisef

Statements from representatives of both companies aren´t enough for you?

@umisef

Quote:

But hey, let's be direct --- who made up the CD that's going to ship with these boards? Was it Hyperion, or was it Acube?

Post #12

Quote:

The AmigaOS 4.1 CD:s will be shipped to ACube in a week from now!

The poster seems quite sure of himself, no? Why would Acube be waiting for a CD if they were making it themselves?

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

Quote:

Statements from representatives of both companies aren´t enough for you?

Statements which are (a) incorrect and/or irrelevant from a technical point of view, and (b) appear to be carefully worded to imply, but not explicitly say something... Well, let's just say that I remember other statements like that, and in those cases, the implied-but-never-said something turned out to be not the case...

So, asking those representatives to make a simple, declarative statement on the issue (rather than talk technicalities which at least one of them has no hope of ever comprehending) seems like the reasonable thing to do.

Quote:

The poster seems quite sure of himself, no? Why would Acube be waiting for a CD if they were making it themselves?

The poster, however, does not have a whole lot of credentials.

Last edited by umisef on 29-Jan-2011 at 02:49 PM.

@umisef

Quote:

Not really, either. The 460EX has four "Universal Interrupt Controllers", at addresses 0xc0, 0xd0, 0xe0 and 0xf0. The 440EP has two of the very same Universal interrupt Controllers, at addresses 0xc0 and 0xd0. Thus, the interrupt hardware integrated into the 460EX is a strict superset of that found in the 440EP. And given that the two UICs on the 440EP provide 10 external interrupts, which was sufficient for the Sam440 line, it is quite possible that the extra 6 interrupts on the 460EX are unused. Or used for something which is not currently supported (sound comes to mind...).

Wrong, the extra two UICs handle the following IRQs:

- Gigaethernet (full working)
- USB OHCI (full working)
- SATA2 (which is still sperimental, but IRQ works...)
- PCI Express (full working)

About the L2 cache on the 460ex, please check your datasheets, it's not a "trasparent" cache as on other CPUS, it could be used as SRAM and need to be setup before use (not a big deal but anyway you need to init it)

_________________
Massimiliano Tretene, ACube Systems, Soft3

@umisef

Quote:

But hey, let's be direct --- who made up the CD that's going to ship with these boards? Was it Hyperion, or was it Acube?

The CD is produced with work done by the following people:

- Hans-Jorg and Thomas Frieden for the 460ex kernel
- AOS4 developers Steven, Renè, Jorg, Hans, Max, Ross for the specific Sam460ex drivers
- the others AOS4 developers (too long to name them all) for the AOS4.1 update 2 baseline
- AOS4 betatesters
- ACube for assembling the final ISO from the AOS4.1 update2 baseline, manuals and packaging as required and approved by Hyperion

Hope it's clear

_________________
Massimiliano Tretene, ACube Systems, Soft3

@umisef

When you are in a hole, stop digging.

Quote:

About the L2 cache on the 460ex, please check your datasheets, it's not a "trasparent" cache as on other CPUS, it could be used as SRAM and need to be setup before use (not a big deal but anyway you need to init it)

Sure --- but that's really not the job of the operating system, now is it? By the time the OS gets loaded, the cache should be well and truly set up, and be fully transparent.

Quote:

- ACube for assembling the final ISO from the AOS4.1 update2 baseline, manuals and packaging as required and approved by Hyperion

Thank you. So much for Nicsoft's assertion of
Quote:

The AmigaOS 4.1 CD:s will be shipped to ACube in a week from now! The Sam460ex motherboard has been ready for delivery for months, but Acube have had to wait for Hyperion to finish the Sam460ex version (of OS 4.1)...

Last edited by umisef on 29-Jan-2011 at 03:31 PM.

@umisef

Quote:

Sure --- but that's really not the job of the operating system, now is it? By the time the OS gets loaded, the cache should be well and truly set up, and be fully transparent.

Not sure here... U-Boot could use the L2 as SRAM, for example to boot the board from there, without need to have DDR ram installed.

The Linux kernel does the same, configure L2/SRAM as cache

_________________
Massimiliano Tretene, ACube Systems, Soft3

hmm, to sum up it really looks like it is as umisef suspects: sam460 os4.x cd has been assembled by acube from previous os4 release or update and the contributions of third party developers supporting os4 with their work. production of this cd by acube has been obviously (only) approved by hyperion, and the statement regarding the shipment has nothing to do with hyperion themselves.

Last edited by Nicsoft on 29-Jan-2011 at 11:53 PM.
Last edited by Nicsoft on 29-Jan-2011 at 05:38 PM.

@nicsoft
Quote:

The AmigaOS 4.1 CD:s will be shipped to ACube in a week from now! The Sam460ex motherboard has been ready for delivery for months, but Acube have had to wait for Hyperion to finish the Sam460ex version (of OS 4.1)... Porting AmigaOS 4.1 to X1000 is a much bigger task, so therefore you shouldn't be so eager to call it vaporware!

well ive just checked, and for me at least it is pretty clear that in the above post you are trying to suggest that the version of os4 to be supplied with the sam460 is an updated version of os4.1 update 2 programmed by hyperion themselves, and that it is going to be shipped to acube by hyperion soon.

Last edited by Nicsoft on 29-Jan-2011 at 11:53 PM.
Last edited by Nicsoft on 29-Jan-2011 at 06:45 PM.