Question: How powerful of a graphics card will work in a beige Power Macintosh G3?

[DearthnVader]

Isn't any display card "Parent" to an connected display?

I think for the purposes of OF device tree, no, but OS X does build it that way for the drivers and the IO NDRV.

So if you look at the device tree via the IOReg, it will have an entry @0 for the display node, but in OS 9 in the IODispalyNameRegisrty I think it only really matters that the EDID is passed from OF.

 

[LightBulbFun]

Good point!

...so i tried. And yes! Flashed back to the PC ROM, the card shows up in the device list. Sitting in A1 in ths case.

good to see it does show up! I would shuffle it about the slots and make sure its all consistent to be safe

Isn't any display card "Parent" to an connected display?

you only get parent/child nodes on Graphics cards in OpenFirmware if its a multiple display capable graphics card, since OS X and the frame buffer personalities and drivers needs to have the multiple display outputs mapped out

as the Rage 128 Pro is not a dual display capable card, everything is contained within the one node


if you plop the Radeon 9000 PCI you have back in, you should see the multiple display output nodes

and its well known that the ATI Radeon 9200 works well in all old world PCI Macs, and thats a dual display card, so we know Old World OpenFirmware can handle the multiple nodes of an ATI fCode ROM at least

 

[flyproductions]

you only get parent/child nodes on Graphics cards in OpenFirmware if its a multiple display capable graphics card, since OS X and the frame buffer personalities and drivers needs to have the multiple display outputs mapped out

as the Rage 128 Pro is not a dual display capable card, everything is contained within the one node

if you plop the Radeon 9000 PCI you have back in, you should see the multiple display output nodes

Thanks for this explanation!

But after all, i keep asking myself, why it just dissapears in the "real" G3, when beeing flashed with the fCode. And why the log just quits after the first 4k. Would be a difficult thing to do, to get to a working ROM, if the simulation's behaviour differs so much from the real scenario.

 

[flyproductions]

if you plop the Radeon 9000 PCI you have back in, you should see the multiple display output nodes

...and yes, you are perfectly right!

FF855560:         /keyboard@0,0
FF855F80:         /mouse@1,0
FF856240:       /pram@0,0
FF8562F0:       /rtc@0,0               [rtc]
FF856820:       /power-mgt@0,0         [power-mgt]
FF8572D8:   /ATY,Radeon9000iParent@E
FF87CE08:     /ATY,Radeon9000i_A@0     [ATY,Radeon9000i]
FF87D5A8:     /ATY,Radeon9000i_B@1     [ATY,Radeon9000i]
FF856B80: /perch@0                     [Whisper]

Also naming is a bit different from what the Sawtooth does: "/ATY,Radeon9000i_A" would be "/ATY,Radeon9000i,Display-A" in the new world device list, i'd suppose.

 

[LightBulbFun]

...and yes, you are perfectly right!

FF855560:         /keyboard@0,0
FF855F80:         /mouse@1,0
FF856240:       /pram@0,0
FF8562F0:       /rtc@0,0               [rtc]
FF856820:       /power-mgt@0,0         [power-mgt]
FF8572D8:   /ATY,Radeon9000iParent@E
FF87CE08:     /ATY,Radeon9000i_A@0     [ATY,Radeon9000i]
FF87D5A8:     /ATY,Radeon9000i_B@1     [ATY,Radeon9000i]
FF856B80: /perch@0                     [Whisper]

Also naming is a bit different from what the Sawtooth does: "/ATY,Radeon9000i_A" would be "/ATY,Radeon9000i,Display-A" in the new world device list, i'd suppose.

a bit off topic but an interesting thing I noticed about your ATI Radeon 9000 Pro

is its frame buffer Personality is not one I have seen before ATY,Radeon9000i

the one im used to for the Radeon 9000 Pro, is ATY,Pheonix which is what all the Radeon 9000 Pro's I have from my FW800 MDDs display

ill have to jab one back in and compare the ROM revision and board/card ID's

in the meantime I dont suppose you could post a screen shot from the "profile" tab in the ATI Displays control panel/app?



its interesting to note that my Radeon 9000 Pro is 99702 while yours is 99703

https://forums.macrumors.com/threads/powermac-g4-fun-part-2.2246307/?post=28702025#post-28702025

im pretty sure ATI did sell the Radeon 9000 Pro as a retail card so I wonder if they did actually use a different ROM for that card?

I have never seen a Retail 9000 Mac edition so I dont know if they where DVI ADC like an OEM card or DVI VGA like a PC card, if the later that would require a different ROM for the different port layout and a different port layout would mean a different frame buffer personality

 

[flyproductions]

in the meantime I dont suppose you could post a screen shot from the "profile" tab in the ATI Displays control panel/app?

Even beeing a bit off-topic, sure i can! 😉

its interesting to note that my Radeon 9000 Pro is 99702 while yours is 99703

The card is some little exotic and today quite impossible to find breed from Hercules.

As you can see in the profile, it is flashed with the (unmodified) "reduced" 136 ROM of the 9000pro OEM.

Both ports work. But i don't know, if i ever tried them at once. Also would be interesting to see, how the card performs, compared to the 9200. At least on AGP in older OpenGL-versions the 9000pro is stronger than the 6200. So all that's done here might be a bit pointless anyway...

...if not to see true core image hardware accelleration on a Beige.

 

[LightBulbFun]

very interesting!

As you can see in the profile, it is flashed with the (unmodified) "reduced" ROM 136 ROM of the 9000pro OEM.

I dont think that ROM is actually an OEM ROM, I think its a Retail ROM, since all my OEM 9000 Pro's have been of the ATY,Pheonix personality

however I noticed that the card number is the same which would indicate the physical boards are the same

and indeed doing some googling brought up the User manual which shows this to be true, the Mac Edition is an DVI ADC identical to the apple OEM version

https://drivers.amd.com/manuals/RADEON9000ProMacEdUserGuide.pdf

(one for @Amethyst1 note the special note talking about the IBM Big Bertha )


so its interesting how ATI still went out of their way to make separate retail and OEM ROMs for what was the same card hardware wise

but I guess they wanted to differentiate between the 2 for software reasons, so a retail update would not mess with an OEM card and vice versa, but its interesting how the retail version just predates named frame-buffer personalities



for reference heres the Radeon 9000 Pro from one of my FW800 MDD's

and a ROM dump, incase you want to try out the Pheonix ROM

 

[domii]

As you can see in the profile, it is flashed with the (unmodified) "reduced" ROM 136 ROM of the 9000pro OEM.

The 9000 has a slightly higher core clock I think, 275MHz vs 250MHz (on the 9200) so that helps a bit. Could probably overclock it a little as well by improving the cooling if one wanted to go down that path.

Curious, did you have to do any resistor mods to get it working? Themacelite page talked about a bank of resistors, and having to move one for a "mac switch" and possibly a few others for amtel vs st eeproms on their radeon 9000 page.

Also a little update on the EVGA 6200, I had messed up. In my hunt for newer nvflash to support the 2mb NOR flash I had mixed up my rom files and flashed it with another PC rom. I did not discover that until I spent a few hours saturday using hexedit making my own Mac rom as I wanted to make sure I had the correct softraps/timing..

When I attempted to flash it I got the "Unsupported code type: 0x01. Fcode support depreciated" error. So I tried a few of the prebuilt mac roms, including the one you sent me all with the same error(s). Ugh, can't use the older nvflash tools as this card has a new Nor flash, and the newer nvflash version(s) have a check to prevent flashing of a non-pc rom. I tried going down the rabbit hole of some modified nvflash tools with checks disabled but nothing so far. So I'll put that one aside for now and try a different older 6200 PCI instead.

Kind of too bad, as that EVGA does have a 2mb flash instead of 64kb (and 512mb onboard memory).

 

[flyproductions]

I dont think that ROM is actually an OEM ROM, I think its a Retail ROM, since all my OEM 9000 Pro's have been of the ATY,Pheonix personality

Ah, sure you are right again! 136 is not OEM, but Retail. The ROM, as far as i remember, is a reduced one from the old WIKI.

and a ROM dump, incase you want to try out the Pheonix ROM

Thanks, but i think, i'd prefer to leave it like it is. Never change a running system! ...and clockrates, if ever needed, can be adjusted with ATIcellerator. But also this i prefer to leave as it is for this passively cooled card, for which a replacement wouldn't be that easy to get.

Sidenote, which leads back to the thread's context: Even this card is not allways recognized. Not even with the onboard video disabled. Works fine on most boots after the Mac has been off for some time. But preferably on direct reboots it disappears. And in this cases also is not seen by OF.

 

[flyproductions]

Curious, did you have to do any resistor mods to get it working? Themacelite page talked about a bank of resistors, and having to move one for a "mac switch" and possibly a few others for amtel vs st eeproms on their radeon 9000 page.

Wasn't that for the 9800pro and later found not even necessary, if a reduced ROM is used?

and the newer nvflash version(s) have a check to prevent flashing of a non-pc rom.

Aren't there options you can use to ignore/overwrite these checks? For 5.31 i also have to use -4 -5 -6 -j at the end of the command. Otherwise the app states different board-IDs for the file and the card and refuses to flash.

 

[DearthnVader]

When I attempted to flash it I got the "Unsupported code type: 0x01. Fcode support depreciated" error. So I tried a few of the prebuilt mac roms, including the one you sent me all with the same error(s). Ugh, can't use the older nvflash tools as this card has a new Nor flash, and the newer nvflash version(s) have a check to prevent flashing of a non-pc rom. I tried going down the rabbit hole of some modified nvflash tools with checks disabled but nothing so far.

I just got the same card with the same issue, note that it says "depreciated" suggesting that it has not been removed. We'll have to try different undocumented switches until we can find the magic one if it exists.

 

[DearthnVader]

I just got the same card with the same issue, note that it says "depreciated" suggesting that it has not been removed. We'll have to try different undocumented switches until we can find the magic one if it exists.

Sadly it wasn't an alphabetical switches a-z......

I wonder because we have a EEPROM size that would accommodate both PC and Mac images if we could stitch a PC ROM and an FCODE ROM into one and it may bypass this check if we put the PC image as the first image, @joevt?

Joe's debug ROM v5 for the PNY card goes to abort very early when I try and load it from file and the 6200 ROM Maker ROM only adds the Open and Close .words then seems to go to abort too.

Here are the PC and Mac ROMS for the EVGA 512MB PCI card:

 

[domii]

Wasn't that for the 9800pro and later found not even necessary, if a reduced ROM is used?

Aren't there options you can use to ignore/overwrite these checks? For 5.31 i also have to use -4 -5 -6 -j at the end of the command. Otherwise the app states different board-IDs for the file and the card and refuses to flash.

It was just based on what I read regarding the 9000 on this page: http://themacelite.wikidot.com/radeon-9000np-conversion

I tried all those switches, including the -j. From my understanding, on a rabbit hole I went down on another forum, is that newer versions of nvflash (64bit windows) put in some kind of checks including one for non-PC type of rom (and other types of checks also, but a fellow called joedirt released several modified nvflash versions to bypass many of those checks, but not one for the unsupported code type, think I read through 30+ pages on that thread).

I ended up there is the older nvflash versions don't support the nor type flash on that 2015 made EVGA 6200..
But do correct me if wrong, as I could have missed something!

 

[DearthnVader]

Joe's debug ROM v5 for the PNY card goes to abort very early when I try and load it from file and the 6200 ROM Maker ROM only adds the Open and Close .words then seems to go to abort too.

Escape character is '^]'.
 ok
0 > dev pci1/@2  ok
0 > load hd:5,\ppc\jvt5.rom load-size=ffc0 adler32=91bd8356
 ok
0 > 800000 1 byte-load [ NV43
here:ff9e14a8
 ok
0 > words

close                     
open  ok                 
0 > .properties
vendor-id               
                        000010de
device-id               00000221
revision-id             000000a1
class-code              00030000
interrupts              00000001
min-grant               00000005
max-latency             00000001
subsystem-vendor-id     00003842
subsystem-id            0000b402
devsel-speed            00000001
fast-back-to-back       
name                    pci3842,b402
compatible              pci3842,b402
                        pci10de,221
                        pciclass,030000
reg                     00061000 00000000 00000000  00000000 00000000
                        02061010 00000000 00000000  00000000 01000000
                        42061014 00000000 00000000  00000000 20000000
                        02061018 00000000 00000000  00000000 01000000
                        02061030 00000000 00000000  00000000 00020000
assigned-addresses      82061010 00000000 92000000  00000000 01000000
                        82061018 00000000 91000000  00000000 01000000
                        82061030 00000000 90000000  00000000 00020000

 ok
0 > words

close           open  ok 
0 > " pci1/@2" open-dev to my-self r_3 @00061010.l = 92000000
w_3 @00061010.l = ffffffff
r_3 @00061010.l = ff000000
w_3 @00061010.l = 92000000
r_3 @00061014.l = e0000008
w_3 @00061014.l = ffffffff
r_3 @00061014.l = e0000008
w_3 @00061014.l = e0000008
r_3 @00061018.l = 91000000
w_3 @00061018.l = ffffffff
r_3 @00061018.l = ff000000
w_3 @00061018.l = 91000000
r_3 @00061030.l = 90000000
w_3 @00061030.l = ffffffff
r_3 @00061030.l = fffe0001
w_3 @00061030.l = 90000000
r_3 @00061004.w = 4
w_3 @00061004.w = 6
w_0 @00000004.l = 1000000
 ok
0 > 800000 1 byte-load [ NV43
here:ffa04668
get subsystem-id
val93f=b402
r_3 @00061014.l = e0000008
r_3 @00061010.l = 92000000
w_3 @00061010.l = ffffffff
r_3 @00061010.l = ff000000
w_3 @00061010.l = 92000000
r_3 @00061014.l = e0000008
w_3 @00061014.l = ffffffff
r_3 @00061014.l = e0000008
w_3 @00061014.l = e0000008
r_3 @00061018.l = 91000000
w_3 @00061018.l = ffffffff
r_3 @00061018.l = ff000000
w_3 @00061018.l = 91000000
r_3 @00061030.l = 90000000
w_3 @00061030.l = ffffffff
r_3 @00061030.l = fffe0001
w_3 @00061030.l = 90000000
r_3 @00061004.w = 6
w_3 @00061004.w = 6
w_0 @00000004.l = 1000000
r_0 @00000000.l = a100a104
r_0 @00101000.l = 8f9ec523
ABORT:dd7#2
 ok
0 >

 

[flyproductions]

It was just based on what I read regarding the 9000 on this page: http://themacelite.wikidot.com/radeon-9000np-conversion

Thanks! I didn't even know such procedures existing for the 9000/9200 cards too.

No, i had to do none of that with the pci-card it could just be flashed with the Mac ROM.

 

[flyproductions]

Joe's debug ROM v5 for the PNY card goes to abort very early when I try and load it from file and the 6200 ROM Maker ROM only adds the Open and Close .words then seems to go to abort too.

Don't really get it. Are you working with a real card, flashed with a PC-ROM to get it recognized and then loading the fCode ROM from a file?

I don't even get to the point where i could test anything as OF doesn't even see the card.

 

[DearthnVader]

Don't really get it. Are you working with a real card, flashed with a PC-ROM to get it recognized and then loading the fCode ROM from a file?

I don't even get to the point where i could test anything as OF doesn't even see the card.

The point is to test the FCODE before we flash the card, and I can't flash the card anyway because the version of nvflash needed to support the EEPROM gives an error that FCODE ROM's are "depreciated" and refuses to flash. Look back over the last page of the thread.

OF doesn't see your card because you flashed an FCode ROM to it that has issues, those issues are making the card "invisible", that's why I said very early on that we should be working with a card with a PC VBIOS on it and loading the FCODE ROM from disk. BUT there are a few issues we have to work out, the only card I have successfully done this on are the GF4Ti 4600 and Quadro4 900 XGL, and OF on the Beige G3 seems to have issues as well, but those may only effect my Rev.1 ROM not your Rev.3 ROM.

When @joevt comes back maybe he can say why the GF5/GF6 FCODE ROMs goto abort when loaded from file and executed from byte-load?

 

[flyproductions]

OF doesn't see your card because you flashed an FCode ROM to it that has issues, those issues are making the card "invisible", that's why I said very early on that we should be working with a card with a PC VBIOS on it and loading the FCODE ROM from disk.

Yes, i remember.

So the way to go would be reflashing the card again. And assuming these are the matching commands...

Escape character is '^]'.
 ok
0 > dev pci1/@2  ok
0 > load hd:5,\ppc\jvt5.rom load-size=ffc0 adler32=91bd8356
 ok
0 > 800000 1 byte-load [ NV43
here:ff9e14a8
 ok

where "@2" is where the card sits and "\ppc\jvt5.rom" is the path to the ROM from top level of the disk and "load-size=ffc0" is the size. Has adler32 to calculated or is it anywhere in the file?

And when all that is done, i can again try to execute the commands from #191?

 

[DearthnVader]

Yes, i remember.

So the way to go would be reflashing the card again. And assuming these are the matching commands...

Escape character is '^]'.
 ok
0 > dev pci1/@2  ok
0 > load hd:5,\ppc\jvt5.rom load-size=ffc0 adler32=91bd8356
 ok
0 > 800000 1 byte-load [ NV43
here:ff9e14a8
 ok

where "@2" is where the card sits and "\ppc\jvt5.rom" is the path to the ROM from top level of the disk and "load-size=ffc0" is the size. Has adler32 to calculated or is it anywhere in the file?

And when all that is done, i can again try to execute the commands from #191?

That's on my G5 so the path to the PCI card is going to very depending on the Mac and the PCI slot, but let's wait and see if Joe can figure why FCODE ROM's newer than the GF4 goto abort before much is really done when loaded from disk. I know Atri told me long ago there is a GPU check in the FCode and it will goto abort if it detects a GPU other than the one hard coded in the FCODE, but I don't think that is the issue here...

 

[flyproductions]

That's on my G5 so the path to the PCI card is going to very depending on the Mac and the PCI slot

Path to PCI would be the easiest to find, if only the card appears at all. But with the PC ROM it should. Not so easy, at least i haven't done it yet would be the access to the HD. Is that number the same that also appears with "diskutil list" from the terminal or can it be found anywhere in the device tree?

 

[DearthnVader]

Path to PCI would be the easiest to find, if only the card appears at all. But with the PC ROM it should. Not so easy, at least i haven't done it yet would be the access to the HD. Is that number the same that also appears with "diskutil list" from the terminal or can it be found anywhere in the device tree?

The partition number will be the same as diskutil list.

 

[joevt]

I wonder because we have a EEPROM size that would accommodate both PC and Mac images if we could stitch a PC ROM and an FCODE ROM into one and it may bypass this check if we put the PC image as the first image, @joevt?

Are we talking about the 6200? The fcode is 64K and the rom is 64K even though it appears as 128K in the expansion rom base register (check the markings on the rom chip). If there was enough space, then I can stitch the Mac and PC roms together. I am thinking about working on a compression method for the fcode rom.

Joe's debug ROM v5 for the PNY card goes to abort very early when I try and load it from file and the 6200 ROM Maker ROM only adds the Open and Close .words then seems to go to abort too.

Here are the PC and Mac ROMS for the EVGA 512MB PCI card:

Do you have both PNY and EGA 6200 cards?

0 > dev pci1/@2  ok
0 > load hd:5,\ppc\jvt5.rom load-size=ffc0 adler32=91bd8356
 ok
0 > .properties
...
subsystem-id            0000b402
...
ok
0 > words

close           open  ok
0 > " pci1/@2" open-dev to my-self r_3 @00061010.l = 92000000
...
ok
0 > 800000 1 byte-load [ NV43
here:ffa04668
get subsystem-id
val93f=b402
r_3 @00061014.l = e0000008
r_3 @00061010.l = 92000000
w_3 @00061010.l = ffffffff
r_3 @00061010.l = ff000000
w_3 @00061010.l = 92000000
r_3 @00061014.l = e0000008
w_3 @00061014.l = ffffffff
r_3 @00061014.l = e0000008
w_3 @00061014.l = e0000008
r_3 @00061018.l = 91000000
w_3 @00061018.l = ffffffff
r_3 @00061018.l = ff000000
w_3 @00061018.l = 91000000
r_3 @00061030.l = 90000000
w_3 @00061030.l = ffffffff
r_3 @00061030.l = fffe0001
w_3 @00061030.l = 90000000
r_3 @00061004.w = 6
w_3 @00061004.w = 6
w_0 @00000004.l = 1000000
r_0 @00000000.l = a100a104
r_0 @00101000.l = 8f9ec523
ABORT:dd7#2
 ok
0 >

You're using a New World Mac so it should work. Does it work with the original fcode? Or is this a PC card that you haven't tried in a Mac before?

ABORT:dd7#2 seems to be related to the subsystem-id/model check.
Note that in #192 that get subsystem-id sets val93f as 4d but you have b402. So before you byte-load the fcode, change the subsystem-id property.
4d encode-int " subsystem-id" property
Also, do a lspci to see what the subsystem-id appears as in the pci config registers.

Yes, i remember.

So the way to go would be reflashing the card again. And assuming these are the matching commands...

Escape character is '^]'.
 ok
0 > dev pci1/@2  ok
0 > load hd:5,\ppc\jvt5.rom load-size=ffc0 adler32=91bd8356
 ok
0 > 800000 1 byte-load [ NV43
here:ff9e14a8
 ok

where "@2" is where the card sits and "\ppc\jvt5.rom" is the path to the ROM from top level of the disk and "load-size=ffc0" is the size. Has adler32 to calculated or is it anywhere in the file?

And when all that is done, i can again try to execute the commands from #191?

Beige G3 can't load files from HFS+ disks, can it? Unless from floppy disk?
The load command is load hd:5,\ppc\jvt5.rom. Remember that Open Firmware often replies on the same line you made the command from which means the rest of that line is the response. It may be useful to edit these logs so that responses from Open Firmware commands are on a separate line.

Path to PCI would be the easiest to find, if only the card appears at all. But with the PC ROM it should. Not so easy, at least i haven't done it yet would be the access to the HD. Is that number the same that also appears with "diskutil list" from the terminal or can it be found anywhere in the device tree?

I made a command to list partitions on my Power Mac 8600

First you need to define some things (unless you have Open Firmware 2.4 or Open Firmware 4)

unselect-dev

struct
/w field >pmsig
/w field >pmsigpad
/l field >pmmapblkcnt
/l field >pmpypartstart
/l field >pmpartblkcnt
20 field >pmpartname
20 field >pmparttype
/l field >pmlgdatastart
/l field >pmdatacnt
/l field >pmpartstatus
/l field >pmlgbootstart
/l field >pmbootsize
/l field >pmbootload
/l field >pmbootload2
/l field >pmbootentry
/l field >pmbootentry2
/l field >pmbootcksum
10 field >pmprocessor
drop

504d constant pmsig

Then you need to open the disk:
" pci1/@F/@2/sd@0" 2dup find-device open-dev to my-self
In this case, @0 is the primary drive of an ATA connected disk.

Then you add the commands (see attached file for full listing):

: mytype
...
: list-partitions
...

Then execute list-partitions

 

[flyproductions]

First you need to define some things (unless you have Open Firmware 2.4 or Open Firmware 4)

I have 2.4. But if after all, i just cannot load files from the disk, as it is HFS+, this might be pointless anyway, right?

And not beeing able to load ROMs from the disk while the card is completely invisible to OF to test things, if an fCode ROM is on the card, sounds a bit like a dead end to me. Correct me, if i'm wrong.

Also i did some benchmarks in the Sawtooth, where all the cards run fine. Used OpenGL Extensions Viewer 3.0. And here is the outcome:

6200 PCI (sure reflashed to the "normal use" ROM)

6200 AGP

Radeon 9000pro PCI

Radeon 9000pro AGP

...and, for reference to the high side, the 6800xt.

Sure, no shader 3 support for the Radeons and so no OpenGL 2.0 and 2.1. But for the rest they are significantly faster than the 6200s. So i begin to ask myself, if it is worth all the effort.

 

[flyproductions]

Then you add the commands:

: mytype ( straddr strlen -- )
    bounds do i c@ dup 20 7e between if emit else drop space then loop space ;

: list-partitions { ; partstart blockbuffer }
    cr
    600 alloc-mem -> blockbuffer
    64 1 do
        blockbuffer i 1 read-blocks 1 = if
            blockbuffer >pmsig w@ pmsig = if
                i 2 u.r space
                blockbuffer >pmpypartstart l@ dup -> partstart 8 u.r space
                blockbuffer >pmpartblkcnt l@ 8 u.r space
                blockbuffer >pmpartname    20 mytype
                blockbuffer >pmparttype    20 mytype
                blockbuffer >pmprocessor 10 mytype
                blockbuffer partstart 3 read-blocks 3 = if
                    blockbuffer 2 mytype
                    blockbuffer 200 + 2 mytype
                    blockbuffer 400 + 2 mytype
                else
                    ." ..."
                then
                cr
            else
            then
        else
            ." didn't read first partition block" cr leave
        then
    loop
    blockbuffer free-mem
;

Then execute list-partitions

Is all this necessary, if there is just one disk with one partition on it?

 

[joevt]

Is all this necessary, if there is just one disk with one partition on it?

Apple Partition Map formatted disks contain a bunch of driver partitions so it might be useful to know exactly which partition contains the file system.

Sure, no shader 3 support for the Radeons and so no OpenGL 2.0 and 2.1. But for the rest they are significantly faster than the 6200s. So i begin to ask myself, if it is worth all the effort.

Can the Radeon's work in a Open Firmware 1.0.5 Mac?
If we get the 6200 working then we can do the same for other Nvidia GPUs.