26-Oct-2021 04:40 GMT.
[News] Elbox posts view on G-RexANN.lu
Posted on 29-Mar-2001 23:20 GMT by Christian Kemp12 comments
View flat
View list
Darrin wrote: I just read this on the Mediator mailing list. It makes VERY interesting reading...
Subj: [Amiga-Mediator] Watch out: Insider's comments on G-Rex

Date: 3/29/01 3:02:54 PM Central Standard Time

From: support@elbox.com (ELBOX)

Reply-to: Amiga-Mediator@yahoogroups.com

To: Amiga-Mediator@yahoogroups.com, G-REX@yahoogroups.com


We have received a number of inquiries about whether the SharkPPC G3/G4 card could be used in the G-REX busboard.

Here is the response to clear up these doubts:

The SharkPPC G3/G4 card is a PCI-standard card which meets all the requirements of the PCI 2.2 specification. We are using PCI CPC710 bridge chips by IBM (http://www.chips.ibm.com/products/powerpc/chips/) in SharkPPC cards.

Designing the Mediator busboard, we have foreseen mechanisms to enable full use of the functions offered by the MPC710 interface, which operates as the Mediator PCI busarbiter. This required, firstly, implementing the full system for busmastering management in the PCI slots of Mediator and the full implementation of interruption system specified by the PCI standard.

The G-REX busboard design (called PREDATOR in some countries) has been made available for some weeks now. The G-REX design is based on the single-slot simplified bridge designed only with the aim of supporting a graphic card; this bridge was previously composed within the BVision card. Applying this bridge in the G-REX busboard resulted in a number of limitations. The most important of them are:

1. Complete lack of support for the busmastering (DMA) between PCI slots (the lines controlling busmastering - GNT#, REQ# are not physically connected to the 2nd, 3rd, 4th and 5th G-REX slot).

This makes the operation of PCI cards impossible, which cannot use their mutual resources without overloading the processor. Therefore, the G-REX busboard is not capable of offering what Elbox presented in the WOA2000 fair, where a TV card in the PCI slot was writing as busmaster data to Voodoo3 graphic card at full speed without intermediation of the processor.

2. Lack of a parity control system (the PAR parity line is not physically connected to the control logics of the busboard).

It means that many standard popular PCI cards will not run in this busboard.

3. Reduction of the interruption system from 4 to 1 - the interruption lines (INTB#, INTC#, INTD#) are not physically connected to PCI slots.

The processor card could be then installed only in the 1st slot, the one to which GNT and REQ signals are connected (which enable taking over controlling address/data buses).

However, as the G-REX busboard does not support (DMA) busmastering between slots, co-operation of the PCI processor cards with any cards working in this mode is impossible. All the currently produced TV, USB, FastEthernet and more advanced music cards (e.g. Sound Blaster Live) work solely in the busmastering mode.

These limitations in G-REX make installation of any advanced processor cards in it useless and ineffective.

I am really embarrassed to read letters from people who -- having no idea about hardware -- give their opinions on various products or designs. Elbox is a company present in the electronics for many years. Elbox employs high-standard engineers electronicians and programmers.

We are not in the habit of commenting devices or software, but here we make an exception, as the level of misinformation in the ML like this one reached its critical level.

Best regards,

Darek Dulian

ELBOX COMPUTER, Support Department

List of all comments to this article
Sorted by date, most recent at bottom
Comment 1Amifan29-Mar-2001 22:00 GMT
Comment 2Kay Are Ulvestad29-Mar-2001 22:00 GMT
Comment 3Graham29-Mar-2001 22:00 GMT
Comment 4Darrin29-Mar-2001 22:00 GMT
Comment 5Alkis Tsapanidis29-Mar-2001 22:00 GMT
Comment 6Darrin29-Mar-2001 22:00 GMT
Comment 7Ralph30-Mar-2001 22:00 GMT
Comment 8Darrin30-Mar-2001 22:00 GMT
Comment 9Johan Rönnblom30-Mar-2001 22:00 GMT
Comment 10Alkis Tsapanidis30-Mar-2001 22:00 GMT
Comment 11Darrin30-Mar-2001 22:00 GMT
Comment 12Alkis Tsapanidis30-Mar-2001 22:00 GMT
Back to Top