According to a thread on Amiga-news.de, there's talk that PCI riser cards won't work in µA1-C/I, because the PCI Slot has no INT (Interrupt) signals for a riser card?

If this is true than the PCI bus makes no sense on µA1?

Or has anybody with µA1 a PCI-riser card that works?

Please EYETECH, play with open cards and tell us about the hardware issues of all A1 (A1SE / 7 A1XE / µA1-C and µA1-I)

Thanx.

Last edited by mbilla on 06-Jan-2005 at 10:23 PM.

_________________
A computerworld without MS products and Windows!
Connect your Amigas ...
...The Red ONE-A1XE G4 - A3000T- A3000 - A4000 - A2500- A1000 - A600 - CDTV - CD32...
and your PDAs and laptops ...
Psion 5mx Pro - Psion NetBook - Apple iPhone - MacBook Pro

@mbilla

According to the original announcement, you should be able to use pci risers, but I don't know of anyone who has tried it yet.

@Herewegoagain

I hope they work, because the uA1 is a less than attractive purchase if you can't fit any PCI cards to it.

Imagine being stuck with a Radeon 7000 for the lifespan of the machine

_________________
Who do you serve, and who do you trust? - Galen

@mbilla

I can gladly report that PCI card and riser card works quite happily in my Micro A1 !!!!

Regards
AnythingAmiga

@anythingamiga

Apparently the PCI bus in the PCI riser slots seems to be NON-DMA!!!
But on the other hand www.amiga-news.de threads seem to be a lot of misinformations.

I would like to see PCI riser cards to work in DMA mode as I will buy a micro-A1-I if available.

How many slots do you have? 2 or 3?

_________________
A computerworld without MS products and Windows!
Connect your Amigas ...
...The Red ONE-A1XE G4 - A3000T- A3000 - A4000 - A2500- A1000 - A600 - CDTV - CD32...
and your PDAs and laptops ...
Psion 5mx Pro - Psion NetBook - Apple iPhone - MacBook Pro

Please, can someone confirm which of
these assertions are true e which not??

Thank u in advance for clarification
alex

_________________
uA1 750Gx AmigaOS 4.1 u6
512 MB
SIL0680
HD 120GB + HD 20GB
DVDRom + DVDWriter
WNCE2001 wifi adapter
-----------------------------
AlexDran.net

@anythingamiga

What kind of riseer card do you have?

rgds,
Stefan

@mbilla

I surely hope this ain't true, because it throw a could wrench in our AF design specs.

Last edited by AmigaMac on 07-Jan-2005 at 08:35 PM.

_________________

@mbilla

INT is provided by an onboard chip if it's a one-to-several port riser.
I have no idea if it requires drivers, or if it's automagic.

@AmigaMac

Since you only plan to offer one pci slot, you'll have no problems.
You only need an 90 deg. extender.

Last edited by Gopal on 07-Jan-2005 at 08:55 PM.

_________________

@AmigaMac

I can't see any reason why a single-slot riser would not work, as it is just a simple adaptor that passes through the signals from the motherboard PCI slot unchanged. Multi-slot ones are more complex.

It would be nice if some obliging uA1-C owner could stick a 2- or 3-slot riser into their machine and see how it gets on when filled with cards, including at least one high-bandwidth DMA device like a Sil680.

But I've a horrible sinking feeling that it won't work, not the 3-slot riser anyway. Mai's product sheet for the ArticiaS states that it can handle five busmasters on the 33MHz PCI bus (bus 0) and two on the 66MHz bus (bus 1). On an SE or XE that is physically implemented like this:

BUS0 : PCI33 -- PCI33 -- PCI33 -- VIA 686B -- 3c920.
BUS1 : AGP -- PCI66.

But on the uA1-C there are five devices on the motherboard already. I'm presuming they are mapped like this :

BUS0 : VIA 686B -- CIM8738 -- 3c920 -- PCI33.
BUS1 : Radeon 7000.

So that would mean plugging in a 2-slot riser would take bus 0 to it's full complement of 5 busmasters. Of course, my presumption may not be correct. There are implementations (such as clocking bus 1 to 33MHz and sticking the 8738 or 9c920 on it) that would permit a 3-slot riser to work.

It'd be nice to know for sure.

_________________
Who do you serve, and who do you trust? - Galen

@DrBombcrater

Quote:

It'd be nice to know for sure.

You are getting very close to the patch on the XE board now
There is a way to connect the VIA chip to Artica without using the any of the BUS1
arbitration inputs. This will leave 3 slots available.

rgds,
Stefan

@DrBombcrater:

Your knowledge of PC hardware is better than mine.

Surely there is only a single Req/Grant pair on a PCI slot? Thus a 3-slot riser would have to connect the mobo slot to share the same R/G on all three new slots.

Don't know much about PC hardware, but I suggest the same would be true of the DMA R/G.

Whatever is wired to the single mobo slot must be shared amongst all remote slots that you run from it. In other words, the maximum throughput for the riser is going to be the same as for a single slot.

tony

_________________
cheers
tony

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

@DrBombcrater

Quote:

Imagine being stuck with a Radeon 7000 for the lifespan of the machine

We had AGA for several years....

_________________
Site admins are people too..pooff!

@mbilla


Will you please post original discussion for me to miikka@domnik.net . Does not matter, if it is german only. Oops, I was cabable to open discussion forum. Just link is enought then :) Thank you.


Before you start to be jumpy, make couple of questions for yourself ...

1. Do these people really own uA1 ?

2. Do they have schematics of uA1. Not likely! But if someone says, that uA1 don't have INT line installed, it is better to search very very carefully. Days are gone, when we had only compatible hardware. Probs to install 3 slot riser can be fault of riser itself also.

3. Are they tested situation properly ?


Lot of bull###t is around. Partially because of untrust, partially because people talk a lot without knowing (maybe they don't even own uA1... :), partially because someone will us fail (really...)

Before it is fact, it is rumour... :)

Before it is yet another hw-flaw, it could be also other situation. More likely.

Miikka

Last edited by mlehto on 08-Jan-2005 at 07:27 AM.
Last edited by mlehto on 08-Jan-2005 at 07:27 AM.
Last edited by mlehto on 08-Jan-2005 at 07:08 AM.

@tonyw

>Whatever is wired to the single mobo slot must be shared amongst all remote slots that you run from it. In other words, the maximum throughput for the riser is going to be the same as for a single slot.


Is it allways same in PCI-slots, wich are in same channel ?? Then rest is handled by busmastering/DMA/INT ??

This is more or less same than ZIII, shared throughput.

I mean, that you cannot get more than 133 MBs per one channel. This throughput is anyway allways shared between slots. Busmastering/DMA/INT keeps thing smooth and "multitasking" from wiev of cards...

So if you have 6 slots shared to 2 channels (3+3 slots) your maximum throughput is 266 MBs total. In theory after all. Of course you have other on-board stuff allso in PCI.

Just for clarification, that uA1 is not any worse than other setups with PCI ...

Please correct, if I'm wrong... :)

Miikka

@tonyw

Since the maximum throughput for a single slot is the maximum throughput for the whole bus, you basically just said that the maximum throughput on the riser card is the same as for a normal PCI bus

Good thing you said it, though

About R/G lines, this is why the first slot on the Peg has more pins than the other two. A riser placed here acts just like one of the other slots, because the extra R/G signals are routed from the extra pins. Unless I'm VERY mistaken. If I'm just a LITTLE mistaken, don't bother correcting me

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@mbilla

Checked that discussion, I didn't find nothing real alarming from there :)

Thank you about link anyway :)

If someone want check (german only...):

http://www.amiga-news.de/de/news/comments/151518.html


Miikka

Last edited by mlehto on 09-Jan-2005 at 05:08 AM.
Last edited by mlehto on 09-Jan-2005 at 05:08 AM.