Links| Forums| Comments| Report news
Chat| Polls| Newsticker| Archive

[Login] [Register] [Forgot your password??]

< Next messagePrior message >

Interview with Frank Mariak

Author: Team
Interviewpartner: Frank Mariak from Vision Factory Developments
Translator: Martina Jacobs (Interview in German) If we understood right, there are two different kinds of drivers under development: CGX VIRGE V3 driver, and CGX Voodoo driver. What is the technical difference between those, and which driver is meant for which product.

Frank Mariak: Those drivers are designed for two PCI boards completely different from each other. On the one side for the VIRGE/DX chip set, which was already utilized on the CyberVision64/3D, but in the meanwhile became somehow anterior. This still appears to be a good and cheap compromise for Amiga considering 2D performance. The Voodoo driver of course is meant for the Voodoo3 chip set by 3dfx. This one offers remarkable enhanced 3D performance in comparison to the CVisionPPC/BVisionPPC boards (if used along with a PPC processor), and even more 2D performance, than the CyberStorm/Blizzard graphics accelerators with Permedia2 chip set. What will support of 3D capabilities be like?

Frank Mariak: There will be no special support from CyberGraphX. At all events for MorphOS there will be a modulated 3D extension basing on Rave for the Voodoo boards. The Warp3D team has said to be interested in modulating the 3D hardware, but unfortunately I don't know very much about the 3D development state. Please, illustrate in some more detail the (alleged?) technical problems with developing Voodoo drivers for the Mediator board.

Frank Mariak: I have to admit that initially I programmed the Voodoo driver on the PCI interface extension for my CyberStorm PPC board and then just tried to make it running on the Mediator board. Originally I used that PCI extension for developing Permedia2 CVision/BVisionPPC drivers more than two years ago, since the prototype board of these boards was a PCI card for Apple PowerMac. Due to this the current Voodoo driver worked with this PCI extension port at first, the modulation for the Mediator board then was done supplementary. The announced G-REX extension rather is just a reworked 4 slot variant of this 1 slot test board. Since I don't have that extended development system on the ELBOX case as I have on my A4000T, I always only tested functionality of the driver with the Mediator board. The technical problems in the meantime have been discussed extensive on miscellaneous nets. Basically one may say that the mystified "bank-switching problem" doesn't cause any trouble, as long as only one CPU (e.g. 68k) is concerned. Also a PowerPC-only operating system (which emulates 68k-code) should cause no problems. Does the cooperation agreement with Elbox still exist, and if not, due to what reason and by whom was it canceled?

Frank Mariak: I definite did not cancel the agreement. After all one may not forget that ELBOX products are covering a much bigger target group than extensions like G-REX do, which are designed only for the special LocalBus on the Blizzard/CyberStrom accelerator boards. But I agree that the timing of the G-REX announcement was quite disadvantageous. I'm already in contact with DCE since more than one year, and since then to bring a successor graphics board for the current Permedia2 products was looked ahead. DCE and me, we set the seal on developing Voodoo3 driver signing a contract now, so Thomas Dellert has planning reliability, and also I would not come out of with with empty hands. There never has been such a contract with Elbox, there has just been some kind of declaration of intent on cooperation with an option of a possible Voodoo driver, since in these days it was not to foresee if such a driver really would work. This was meant to be that I would create free limited CGX3 drivers which then would be delivered along with the Mediator board. I then should be responsible of "bringing in" my input myself by selling the CGX4 CDs. In the meanwhile I ask myself, how I could agree with this. In the view of stagnating software sells (CGX 4 CD), software pirates, and continuous requests from ELBOX when at last a somehow working Voodoo driver for the Mediator board would be available, I felt extremely pressurized. In the end and moreover one should not forget that at this time Mediator boards are alowed to be delivered along with a free CGX3 VIRGE driver, and will be in the future. Even after DCE's release there was an update for this driver, so there should not be any talking about canceled cooperation. Why did you decide an exclusive agreement?

Frank Mariak: Only this way a really satisfying solution for customers can be gained. Only licensing makes it possible to support a product in an optimal way. In my notion that was never eligible with ELBOX. Does the exclusiveness of the Voodoo driver concern GGX3 or CGX4 or even both of these?

Frank Mariak: I don't understand that panic about the current Voodoo driver, at all. Is the Mediator board a slot for a Voodoo graphics board or a 4x PCI slot adapter? People might better ask themselves where are the announced drivers for the Yamaha sound card, the 100MBit network card or the MPG decoder card. In place of focusing on completion of this driver a PCI PPC G3/G4 extension was announced wholehearted which virtually automatically converts the Mediator board to a data-thruput-"high-achiever". The Amiga world has seen themselves how long it took come to some state of being error free on 603/604e processors with PowerUP or the subsequent competitor WarpOS. A G3/G4 software solution shall suddenly appear for Amiga within two months? I don't want to further comment this, but I entertain some doubt.... I just can wish the ELBOX company much success with that. They at least developed a 68030 accelerator board or have this under their present offer. Does this agreement also mean that there will be no Voodoo CGX drivers for Predator by Eyetech? (Which seems to be quite similar to the G-REX.)

Frank Mariak: No, as a matter of principle nothing avoids respective license for the drivers. Can the drivers be used with every standard board or only with tested boards (compatibility list provided)?

Frank Mariak: In the case of e.g. Voodoo3 the boards available are substantially of the same assembly. But basically it is recommended to use only tested boards since else function of the drivers cannot be guaranteed. This is what I feel to be a significant problem with the at this time available/released PCI port extensions. When I purchase a PCI card for a PC, I will get a driver adapted to this card along with this card. An Amiga user doesn't have such opportunity if he purchases a PCI card in some shop. Many people simply underestimate the software problem. Even on a WindowsPC this is often solved dissatisfying and you get put off with half done drivers. When I purchased for a ZorroII/III extension I was able to call the manufacturer to account, if hardware/software did not run as it should do, and the manufacturer was able to respond if necessary. Thus also now only tested products should be used or the respective board should be ordered from an Amiga stockist right away to at least reasonably play safe. Does CGX v5 for MophOS represent a complete rewrite of version 4 with enhanced routines or is this a 1:1 port?

Frank Mariak: The drawing routines as well as the drivers where almost entirely newly developed, since CGX4 still uses assemlber optimesed 68k routines which of course can be used emulated only with a PPC. Can CGX be further adopted to WarpOS or PowerUp or can it be further optimised for those? And if, is there hope for this to happen?

Frank Mariak: No, there will be no further adoption to or optimisation for PowerUp or WarpOS. Will there be a 68k version of CGX v5?

Frank Mariak: No, the advantages/enhancements of version 5 only make sense along with a PPC OS e.g. like MorphOS. The 68k variant will stay to be version 4, and independent from V5 will be further developed/enhanced. CGX v5 in conjunction with the current MorphOS test version will be developed and enhanced continuously. Frank, thank you very much for this interview and for your extended answers. (unk) (Translation: unk)

[News message: 18. Oct. 2000, 01:27] [Comments: 0]
[Send via e-mail]  [Print version]  [ASCII version]
< Next messagePrior message >

Masthead | Privacy policy | Netiquette | Advertising | Contact
Copyright © 1998-2021 by - all rights reserved.