What have you done with a PowerPC today?

[Amethyst1]

Thanks for explaining the graphics drivers and NDRVs, I think I get it now

I mean what is the OpenFirmware name of your ATI Rage 128 pro?

ATY,Rage128Pd

note 10.0 and 10.1 cannot load the GUI without a working NDRV (IE it cant just ride the OpenFirmware frame-buffer) […]

I’ll test 10.2 too. (Not that there’s anything to gain since I’ve run 9, 10.1 and 10.2 at WQUXGA from a Mobility Radeon 7500.)

 

[LightBulbFun]

Thanks for explaining the graphics drivers and NDRVs, I think I get it now


ATY,Rage128Pd


I’ll test 10.2 too. (Not that there’s anything to gain since I’ve run 9, 10.1 and 10.2 at WQUXGA from a Mobility Radeon 7500.)

interesting that one obviously does have an on disk NDRV in 10.0 then , if you nuke it, that should force OS X to fall back to the ROM based NDRV, although

also check the ATIDriverUpdate.plist and see if the Rage128Pd is mentioned in there or not, if it is, remove its entry (or nuke the plist file and NDRV of the same name)

 

[Ataman Honcho Honchev]

Not today, yesterday. Went to rural Bavaria. With a local guy we got the LSI SAS PCI-X controller with 3.3V showing up on G4 MDD. The controller was made for Sun, factory pre-flashed with Open Firmware.

But don't do that at home (yet). The PCI-X slots in G4 are 5V and without some trickery it won't power up.

See here:

PCI-X - Wikipedia

en.wikipedia.org

 

[Ataman Honcho Honchev]

Thanks for explaining the graphics drivers and NDRVs, I think I get it now


ATY,Rage128Pd


I’ll test 10.2 too. (Not that there’s anything to gain since I’ve run 9, 10.1 and 10.2 at WQUXGA from a Mobility Radeon 7500.)

Will show what happened in Bavaria yesterday (strictly 3.3 Volt PCI-X LSI SAS card made for Sun in 5V G4MDD slots). Just need to unpack my stuff.

 

[swamprock]

Stuffed a 2005 Mac Mini (overclocked to 1.42ghz and running unsupported OS 9.2.2 and Tiger) and a iOmega MiniMax external drive/usb/firewire enclosure inside a G4 Cube enclosure, using heavy duty velcro. Completely reversible, in case I find a new logic board for my Cube, as I didn't use the core of the Cube at all. Granted, this really accomplishes nothing but taking up more space on my desk, but hey... I think it looks cool, and the OS 9 install reports the Mini as a Cube anyway, so there you go 🤣

I may reroute the power LED to the power button light diffuser on the top of the Cube, if I gather the motivation.

Pictures later, when I get home from work.

 

[swamprock]

Awaiting a DVI to VGA adapter so I can turn this thing over...

 

[Ataman Honcho Honchev]

Good:

Started the Beta 23 of PowerPC / MacOS 9 SCSI Interface Module ("SIM").
Found a bug, it's not fixed yet but I am working on it. Affects "deep sleep". It's not the infamous 3.3 Volt regulator, just a stupid bug I introduced lately.

Called (Eastern) Germany, verified that they have over 100 proper PCI SATA / eSATA / ATA / USB 2.0 combo cards with VIA chipset. VIA chipset is faster than 3112 and is regulator-neutral. 👍

Bad:

The above bug is not fixed yet. The Open Firmware driver and macOS X driver for VIA needs to be written. Did not call Russia regarding the non-combo VIA card yet. They, too, have hundreds of proper (but non-combo) regulator-neutral cards.
Procrastinating the (local) purchase of SATA M.2 drives for Titanium PowerBooks.
Procrastinating the purchase of 2.5" <--> 3.5" drive adapters to be used in FirmTek external enclosures.

Very Bad: OMG, what is going on with Broadcom Frodo controller? DMA is broken? At least, hot-swap is okay.

Very-Very Bad:

The legacy SATA / ATA chips are history (except these sellers of VIA6421A in Eastern Germany and Russia). No manufacturer has them. Forget these. All 3112 / 3114 / 3512 boards currently on sale anywhere in the world have the bad 3.3 Volt regulator. The costs of replacing the regulator turned out to be astronomic, even in Hungary: you pay $3 for the Micrel regulator, you pay as much as the post asks you to send your board to Hungary. I pay at least HUF 2000 (ca. $6) to the repair shop. Than you pay the postage back.
Unless, of course, you can do that on your own: we are talking about $20 overhead at least, if not $30.

Even worse:

The 3114 even with good Micrel regulator will prevent the PowerMac G4 MMD to start - if the card is in the right-most slot. You can use the other three PCI-X slots tho. I can't do anything about it.

 

[Ataman Honcho Honchev]

Good:

Started the Beta 23 of PowerPC / MacOS 9 SCSI Interface Module ("SIM").
Found a bug, it's not fixed yet but I am working on it. Affects "deep sleep". It's not the infamous 3.3 Volt regulator, just a stupid bug I introduced lately.

Called (Eastern) Germany, verified that they have over 100 proper PCI SATA / eSATA / ATA / USB 2.0 combo cards with VIA chipset. VIA chipset is faster than 3112 and is regulator-neutral. 👍

Bad:

The above bug is not fixed yet. The Open Firmware driver and macOS X driver for VIA needs to be written. Did not call Russia regarding the non-combo VIA card yet. They, too, have hundreds of proper (but non-combo) regulator-neutral cards.
Procrastinating the (local) purchase of SATA M.2 drives for Titanium PowerBooks.
Procrastinating the purchase of 2.5" <--> 3.5" drive adapters to be used in FirmTek external enclosures.

Very Bad: OMG, what is going on with Broadcom Frodo controller? DMA is broken? At least, hot-swap is okay.

Very-Very Bad:

The legacy SATA / ATA chips are history (except these sellers of VIA6421A in Eastern Germany and Russia). No manufacturer has them. Forget these. All 3112 / 3114 / 3512 boards currently on sale anywhere in the world have the bad 3.3 Volt regulator. The costs of replacing the regulator turned out to be astronomic, even in Hungary: you pay $3 for the Micrel regulator, you pay as much as the post asks you to send your board to Hungary. I pay at least HUF 2000 (ca. $6) to the repair shop. Than you pay the postage back.
Unless, of course, you can do that on your own: we are talking about $20 overhead at least, if not $30.

Even worse:

The 3114 even with good Micrel regulator will prevent the PowerMac G4 MMD to start - if the card is in the right-most slot. You can use the other three PCI-X slots tho. I can't do anything about it.

And thinking, that if all the efforts would be focused on the USB 2.0 UIM (USB Interface Module) for "9" supporting the VIA VT6214L chipset - that would be a total hammer. There is no USB 2.0 support for "9" (yet). Than, all USB 2.0 sources are in Darwin. True, for "X". And that chip works under "X" like a charm. All what is needed is determination, time - and some USB know-how. Not even USB 3.0, just the USB 2.0. Regarding USB I am a total idiot, apologies.

 

[mikas]

Did nothing that fancy like all you other fellows out there are doing.

Just gave a 20 yo (early)birthday present to one of my G5 cluster members, and doubled the ram from 1GB to 2GB. While booting the young adult up with her humongous ram, I gave the X800XT/256MB some chewing too, and attached my "new" Eizo S2000 to it. 1792x1344 at 60Hz looks like crisp enough to my old eyes. And it goes up to 75Hz at 1600x1200 which looks just nice.

Need to dig up some old software to delve into that periodic nostalgia trip (like ArchiCAD 5.1).

 

[Amethyst1]

[...] attached my "new" Eizo S2000 to it. 1792x1344 at 60Hz looks like crisp enough to my old eyes.

The S2000's (I just love the design of Eizo's monitors; nice and, uh, professional) native resolution is 1600×1200. It's interesting that it can apparently downscale higher resolutions (no mention of that in the manual), but that comes at the expense of picture quality.

 

[mikas]

native resolution is 1600×1200. It's interesting that it can apparently downscale higher resolutions (no mention of that in the manual), but that comes at the expense of picture quality.

That 1792x1344 was an analog signal with an ADC to VGA adapter. Digital signal with DVI cable gives 1600x1200 resolution as the highest option.

ps. the best part is the monitor was donated free of charge. I had to pick it up though, no free delivery.

 

[Doq]

one of my G5 cluster members

Hol' up you have a G5 cluster? And you're saying we're fancy

 

[mikas]

😊
well I guess I just worded it to sound more worth it.
One is a computer, two is dual workstations and three is a cluster.

 

[mmphosis]

Monday to Friday, I use iCal. Today I added a URL in iCal which opened Safari which is safe since the web server is on the local LAN. I pretty much always use TenFourFox but I guess Safari is the default.

 

[Tratkazir_the_1st]

three PCI-X slots tho

AFAIK G4 MDD has 64bit PCI slots, not PCI-X. And they are more problematic in terms of adapters compatibility.

 

[Ataman Honcho Honchev]

AFAIK G4 MDD has 64bit PCI slots, not PCI-X. And they are more problematic in terms of adapters compatibility.

You are right: these slots are 64-bit wide but only 33MHZ.

 

[Ataman Honcho Honchev]

After Victory Day results, now come the Memorial Day results.
During these three weeks the new SATA / ATA driver advanced to the "Beta 24".

Three cards were tested today:

- The engineering reference 4-port Silicon Image 3114 ("Simagic4")
- The "German" USB + SATA "combo" card with VIA 6421A chipset (a flavor of "Viaduct3")
- The "Hungarian" VIA6421A card (no secondary chip, just SATA).

The results as shown.

Here the unloved Silicon Image 3114 was a small surprise: the writes were always slow (and still are).
On reads that card was able to roughly reach the rest (but still slower of course).
All four channels are working here. But the number of transactions is not bad at all.

As expected, the "pure" Viaduct3 beats everything of it's class (middle screenshot).

This driver is better than what was made 20+ years ago but there are still bugs.

Important: this driver is (in theory) fully 48-bit aware and should break the 32-bit 2TB barrier.
Unfortunately there is no SCSI storage driver (FWB, Silverlining, Apple, etc-etc) which issues 16-byte SCSI commands.

Years ago I wrote a SCSI storage driver - but only a modification of it made for early Apple ATA bus and sold as part of FDW HDT was sold. If nothing else, I may go back to that SCSI driver and fix the 16-byte command problem.

Than the 2TB limit for "classic" MacOS is gone. Unfortunately my driver installation tool is just for home use, it's no match to anything nice being sold.


This time the machine is a modified G4. Originally it was bought new at Fry's in Campbell, so it's maybe the only one-owner Mac in our properties. It was a dual 1GHZ MDD model with the bad aluminum heatsink and recalled power supply. Apple sent me a new power supply and after some years I found an 1.42 GHz dual processor and the copper heatsink on fee-Bay. Currently this is the one of the two fastest G4 I have, the second one with a similar processor is in Alsace.

 

[DouglasCarroll]

Today I listened to Pandora on my PowerBook G4 using the terminal program "PianobarPPC" from here...

http://desktopecho.com/pianoppc/

...works great still in 10.4 Tiger.

I also checked out a book on tape from the local library using "InterWebPPC" and was able to LISTEN TO IT ONLINE!! (Yay!) with the same browser!!

Good times indeed!!

 

[pipetogrep]

Had a drink.

 

[Amethyst1]

Had a drink.

For a second, I thought you were running an OS X preview.

 

[ojfd]

Years ago I wrote a SCSI storage driver - but only a modification of it made for early Apple ATA bus and sold as part of FDW HDT was sold. If nothing else, I may go back to that SCSI driver and fix the 16-byte command problem.

Hang on, are you saying that FWB HDT's ATA driver is your creation?
I was experimenting with various HD formatters for internal drive in OS9 (PowerBook G4 17" 1GHz, mSATA drive on ATA-mSATA adapter) and using FWB's drivers resulted in highest sustained transfer speeds, hitting around 90MB/sec.

P.S. ATTO Express ProTools is another nice tool when testing transfer speeds.

 

[Ataman Honcho Honchev]

And today I learned that after Intel got the Vitesse 7174 controller (which was originally designed by APT), it got the code name "Artisea".

Wow, learned that after 20 years. All because after yesterday's tests with "beta24" I found that I introduced a bug and the 7174 is not working properly. Where is the manual? Was unable to find it, found on-line... and the manual from Intel had that code name.

I think, "Vitamin4" as well as it's grand-grand-daddy ("SeriTek/1V4") are gone, "Artisea" sounds so much better.

To keep up with the tradition: "Artisea4", because it has four ports.

(Honestly: I did not like that "Vitamin4" name very much). What do you think?

Hang on, are you saying that FWB HDT's ATA driver is your creation?
I was experimenting with various HD formatters for internal drive in OS9 (PowerBook G4 17" 1GHz, mSATA drive on ATA-mSATA adapter) and using FWB's drivers resulted in highest sustained transfer speeds, hitting around 90MB/sec.

P.S. ATTO Express ProTools is another nice tool when testing transfer speeds.

It started there, with HDT 2.0. The latest HDT ATA drivers (for Apple ATA) were made probably by others.
I am not with FWB since 1996.
Originally FWB hired me after I demoed them a SCSI driver written entirely in Pascal.
While I was moving, they hired someone else for than-new ATA driver.

The second hire was delayed by few months, so I was assigned to the ATA work.

These ATA or SCSI drivers are mass storage drivers interacting with the OS and passing the commands to the controller driver. They have nothing to do with a particular ATA or SCSI hardware.

 

[Ataman Honcho Honchev]

Hang on, are you saying that FWB HDT's ATA driver is your creation?
I was experimenting with various HD formatters for internal drive in OS9 (PowerBook G4 17" 1GHz, mSATA drive on ATA-mSATA adapter) and using FWB's drivers resulted in highest sustained transfer speeds, hitting around 90MB/sec.

P.S. ATTO Express ProTools is another nice tool when testing transfer speeds.

Fun fact: at that time (around 1995) IDE / ATA was horrible.
Many manufacturers broke the barely existing standards.

The first goal with that ATA driver was to support the hot-swap magneto-optical drive from Fujitsu, made for Apple notebooks. The code name of that drive was "Pismo", i.e. "PisMO". And it was really non-compliant.

That drive was given to me and a second copy to my colleague who came from Kiev.

Probably Tratkazir_the_1st will enjoy the most, what happened after that.

Out of frustration the poor "PisMO" drive earned a very special nick: we changed "s" for "z" and added "da" before "MO". Everyone at FWB knew, what it was. They felt sorry for our ordeal and everyone agreed: the poor PisMO drive deserved that nick.

Once during a phone call with Fujitsu I had a tongue sleep. Luckily, the conversation partner wasn't from (ex) Soviet Union - otherwise the ATA project at FWB would be over.😂

 

[ojfd]

we changed "s" for "z" and added "da" before "MO"

Ha ha. You should have kept that name and tried to convince you bosses that with "z" it "sounds better".
(For example, someone at Mackie gave their control surface the name "HUI").

Back to drivers. Who did FireWire drivers at or for Apple? Also you?

 

[Ataman Honcho Honchev]

Ha ha. You should have kept that name and tried to convince you bosses that with "z" it "sounds better".
(For example, someone at Mackie gave their control surface the name "HUI").

Back to drivers. Who did FireWire drivers at or for Apple? Also you?

1) No FireWire unfortunately. Otherwise the FW800 G4 problem with MacOs "9" would be solved.
I have no idea how FWIM (FireWire Interface Module) looks and now it's not as important as UIM (USB Interface Module) for "9". Anyone with some info please come forward: would love to utilize the USB 2.0 ports of that "German" card. Eventually will try to study by myself, but it's an uphill battle.

2) That "PisMO" is the opposite thing (see: Ying vs. Yang) to that printshop.

But the worst English, German or Slavonic vocabulary is very timid and innocent compared to what you may hear on Hungarian or Romanian streets.