@tonyw

That's either a lot of ignorance there, or outright lies/history revisionism...

Quote:

To be fair, neither camp understood the hardware bugs.

The very opposite is true, bPlan/Genesi discovered the bugs, tracked them down, and suggested fixes to MAI, who OTOH refused to acknowlegde them. Only when Genesi were about to drag them to court about it, they accepted to make a new revision of the chip. But there were still bugs. So bPlan created their own patch that sat between the chip and the motherboard, the April fix. Then the April 2 fix. That brought the chip to an acceptable level according to the high bPlan standards. But only with the G3, and only if run at 100MHz. ("There is no MAI (May) without April", proved true within a few years).

No-one knew the problems of the Articia S better than bPlan/Genesi, not even MAI themselves. Eyetech? Well, Alan seems to have made up his own "knowledge" himself ("it's flawless"), with the agenda to sell as many overpriced motherboards as possible, using tight information control (moderators working with the same agenda, secret forums, etc) as a tool. I see people in this very thread who helped "spreading the truth and fighting the FUD" back then. You know who you are. You share the blame and responsibility.


Quote:

The Articia didn't even try to implement coherence, although its detractors called it a bug at the time. Cache coherence was intended to be added to a later model, which never happened.

Wrong, wrong, wrong!

The Articia S was supposed to have cache coherency with DMA transfers. Documentation and marketing material say this clearly. Only late in the game, this was CHANGED, when they realized they couldn't market the thing as such. I have the documents somewhere around if you want to see them. Claiming anything else is history revisionism! They went bankrupt shortly after...


Quote:

the MorphOS camp tried to do it in hardware with their "April" chip. Neither solution worked very well.

The April worked as it should, for what it was. It wasn't able to bring fast G4's as an option on a broad scale though (only a handful machines made). That's why they designed a new motherboard, the Pegasos 2, and offered this at major discounts to Pegasos 1 owners as an upgrade path.

@TRIPOS

Thank you very much for correcting the fanboy's response.
I am glad he is happy with his development board, but only bPlan and Genesi bothered to pursue corrections for the Artica S' flaws with changes to hardware.
And the Peg2, quite rightly, dropped that component entirely.

Thank God for Varisys and A-eon since their designs are so much better engineered.

@m0lebrain

Oh my, you're getting old! :P

Last edited by pixie on 09-Jun-2016 at 02:24 PM.

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@m0lebrain

You have made me look at my registration date and what a coincidence; exactly today 12 years ago I registered

_________________
No PowerPC, No Fun
Make Amiga Great Again

@m0lebrain

Attending College classes this Summer for the first time since I left the University of Delaware in 2002.

OMG that was a long time old.
Man do I feel old.

@TRIPOS

There's no need to leap to anyone's defence. I wasn't attacking anyone. merely stating my opinions of the resulting machines - that is, that although both worked about as well as they could, neither the hardware nor the software solution was an outstanding success.

As I recall (and it was well over ten years ago now), the disk-memory speeds on a Peg I and a standard 800 MHz XE were about the same. I'm certainly not going to argue about it.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

@TRIPOS

Quote:

The boards were crap but OTOH they were never really meant for end-users. Aside from any design flaws on the motherboards, they had that notorious FUBAR northbridge that never worked (not one single revision of it) as it should.

Despite this, it turned out to be a nice usable NG Amiga system for many users such as myself.

Quote:

Articia S

Sorry. I realised the error I made late last night and didn't have time to correct it. The Articia P was the later evolution of the Northbridge that was supposed to bring AGP X4, PCIx and DDR.

Quote:

sii, not sil

Are you sure about that?

@tonyw

Quote:

I'm certainly not going to argue about it

Except that you just did.
I wonder how well an XE would compare to a Peg2?

THAT was the final hardware fix, dumping the crappy Artica S.

Best Regs from an XE. Please call me if anything happens in the OS4 front. Yeah, that we have been waiting from the 2002...

I hope the expensive and also expensively repaired machine will start in the next few years to actually use it.

@PR

Don't worry little XE, we have alternative OS' if the whole OS4 thing doesn't work out.

@iggy

Quote:

I wonder how well an XE would compare to a Peg2?

I was just looking at the Ragemem 0.37 results built into Sysmom. According to the scores the AmigaONE SE G3 @ 600Mhz beats the Pagasos 2 G3 @ 600mhz in everything but the tricky write to RAM. The margins are very small apart from reading and writing to RAM where the result are considerably better on the A1 SE. Of course that's just a one set of synthetic tests so we can't draw any real conclusions from it.

Whatever performance difference is, if there is much, the SE and XE do have one clear advantage over the Pegasos 2, they can use Radeon X and HD cards.

@Fransexy

Happy Birthday

_________________
____________________________
c64-2sids, A1000, A1200T-060@50(finally working!),A4000-CSMKIII
! My Master Miggies- Amiga 1000 & AmigaOne X1000 !
mancave-ramblings
X1000 I BELIEVE

@Raziel

10 Jan 2004 :)

_________________
Cross-developer on Windows, OS3, OS4, Linux; Current Projects:-
Nephele Cloud App OS4
UserProfile System OS4
AmigaOneXE OS4.1.6

TaoSoftwareBlog Youtube

@Rob

Really selective comparison.
Neglible bench difference when using a G3 (and apparently you didn't want to address the more common G4).
And do you really think that running a PCIe Radeon X or HD card in a PCI slot of a 600MHz G3 powered system is worthwhile?
It has no where near the cpu power needed to use those cards.

You use your G3 powered AmigaOne board with a Radeon X or HD card.
I would prefer a Peg2 G4 with a Radeon 9200 or 9800 (the latter having acceleration support under MorphOS).

No Artica S bugs and supported by more than one NG OS.

But to heck with either these days.
I want a P5040 based X5000.

@iggy

Quote:

Really selective comparison Neglible bench difference when using a G3 (and apparently you didn't want to address the more common G4).

600Mhz G3 was the only parity between Pegasos 2 and AmigaOne test results in database of my copy of Sysmon. Nothing I can do about that, sorry.
The only G4 results are an A1 at 1266mhz and a Peg2 at 1133Mhz . That would hardly be fair to the Peg2. Also, like I said, it's just one set of synthetic so it would be pointless to draw any conclusions.

Better to see some real tests where the XE and Peg have a G4 clocked at the same speed and have identical video cards.

Quote:

And do you really think that running a PCIe Radeon X or HD card in a PCI slot of a 600MHz G3 powered system is worthwhile? It has no where near the cpu power needed to use those cards.

You were the one who wondered how the 2 boards compare, I just gave the best answers I could based on what information was available to me. Warp3D Nova supports low end R5 (230, 235 and 235X) cards anyway so you don't need a beast of a video card to take advantage of the new shader based API.

Quote:

You use your G3 powered AmigaOne board with a Radeon X or HD card. I would prefer a Peg2 G4 with a Radeon 9200 or 9800 (the latter having acceleration support under MorphOS).

I've never owned any G3 based and have an X1000 anyway.

@Rob

Fair enough.
The argument about lower end HD cards is certainly valid.
And X series cards start with cards like the X1300 (or even lower).

The X1000 you own is a completely different beast and should blow everything out of the water (until the X5000 is released and then the differences should not be that great).

@iggy

Quote:

The X1000 you own is a completely different beast and should blow everything out of the water (until the X5000 is released and then the differences should not be that great).

Maybe not a really fast G4. Crisot replaced the CPU on his XE and i believe it is clocked and stable at 1466Mhz, although he put 1.4Ghz in his signature.

With the WarpUP for Sonnet project underway I wonder if anyone would build a similar card around the Tsi108 host bridge, imagine putting a 1.7Ghz 7448 on there. The idea of a 1200, 3000 or 4000 having more CPU grunt than the X1000 is kind of amusing.

Seems like guys went a bit off topic

_________________

@Rob

More grunt than an X1000? Not from the speeds you mentioned.
It would be close with a 1.8 or faster 7448.

And there is an unused second core on the Nemo board.
Also, video cards on the X1000 would have a proper PCIe card, the hard drives are faster SATA units ( although you could install an Silicon Image controller on the XE), and memory throughput is probably higher on the Nemo board.

There are a few 1400MHz G4 cpu cards for the Peg2, but I didn't bring that up as we would be comparing apples and oranges again.

And legacy hardware with a fast PPC?
It would be slowed down by the Amiga chipset.

@AmigaMac

Quote:

Seems like you guys went a bit off topic

Like that is something new.