Hi,

is there any way to use the Sonnet Duet G4 CPU-upgrade with MorphOS?
At the Moment my G4 is shown with 600MHz on the boot pic.
It's no big issue as MorphOS is running very well with 600MHz, but would be cool to use the CPU-upgrade as well.

Regards,
Nimrod

Believe Redrumloa had the Sonnet Encore g4 card on his Quicksilver Mac a few years ago. Mind you that MorphOS will only use one of these CPUs and not both. :)

But if you have the that card, throw it in! :)

****Just be sure you've done any firmware upgrading to the quicksilver if you need. You may have to google it. I know I had to flash a MDD mac before I installed a Sonnet card into the MDD. May be the same story for the Quicksilver.

Quote:

Yes it works. As Doffo mentioned I had one in a Quicksilver for a while. As a matter of fact I had the 1.8 and it overclocked well to 2.0.

Also Doffo mentions Google to get the manual for the Sonnet card and check your firmware. Chances are you are fine as I was fine, just dropped in and went. You may or may not have to update the firmware first. My memory is a little vague, but it isn't a huge problem if you have to update your firmware.

This thread talks about me benchmarking Quake 3 with that system. My reflashed Radeon 8500LE seemed to be the limiting factor.

Quote:

thanks for the answers.
After I have installed the Sonnet expansion I have upgraded the firmware using OSX. Both cores were running fine under OSX with full speed, therefore I think that everything is fine on Sonnet side.
At the moment MorphOS is running exclusivly on the system, no OSX any longer, but with just 600MHz.
I understand that the second processor is not usable with MorphOS.
But, is there any way to increase the processor clock from 600MHz to 1.8GHz as it should be?

Thanks,
Nimrod

Question is wether the system is really running at 600MHz or just claiming so in the splash screen.

Does typing "cpu" in the shell also report 600MHz ?

AFAIK there is no way to underclock these cards in SW and both my 2x1800 and 1x1800 (both Sonnet) were reported as 1800MHz way back when I had them in regular use.

[ Edited by Kronos 07.05.2016 - 20:53 ]

really strange...the shell command cpu tells me that there are 2 CPUs:
System: PowerMac3,5
Machine: 0x1
PageSize: 4096 Bytes
CPUCount: 2
CPU0: 7445/7455 (G4) Version 32769 Revision 771
CPU0: CPUClock 599999998 BusClock 133160004
CPU0: FPU Unit exists
CPU0: Altivec Unit exists
CPU0: Performance Monitor Unit exists
CPU0: DataStream Unit exists
CPU0: ReservationSize 32
CPU0: Ticks 4
CPU0: CacheL1Type 0x0
CPU0: CacheL1Flags 0x1000F
CPU0: ICacheL1Size 32768
CPU0: ICacheL1Lines 1024
CPU0: ICacheL1LineSize 32
CPU0: DCacheL1Size 32768
CPU0: DCacheL1Lines 1024
CPU0: DCacheL1LineSize 32
CPU0: CacheL2Type 0x0
CPU0: CacheL2Flags 0x1001D
CPU0: ICacheL2Size 262144
CPU0: ICacheL2Lines 0
CPU0: ICacheL2LineSize 0
CPU0: DCacheL2Size 262144
CPU0: DCacheL2Lines 0
CPU0: DCacheL2LineSize 0
CPU0: CacheL3Type 0x0
CPU0: CacheL3Flags 0x1001D
CPU0: ICacheL3Size 2097152
CPU0: ICacheL3Lines 0
CPU0: ICacheL3LineSize 0
CPU0: DCacheL3Size 2097152
CPU0: DCacheL3Lines 0
CPU0: DCacheL3LineSize 0
CPU1: 7445/7455 (G4) Version 32769 Revision 771
CPU1: CPUClock 599999998 BusClock 133160004
CPU1: FPU Unit exists
CPU1: Altivec Unit exists
CPU1: Performance Monitor Unit exists
CPU1: DataStream Unit exists
CPU1: ReservationSize 32
CPU1: Ticks 4
CPU1: CacheL1Type 0x0
CPU1: CacheL1Flags 0x1000F
CPU1: ICacheL1Size 32768
CPU1: ICacheL1Lines 1024
CPU1: ICacheL1LineSize 32
CPU1: DCacheL1Size 32768
CPU1: DCacheL1Lines 1024
CPU1: DCacheL1LineSize 32
CPU1: CacheL2Type 0x0
CPU1: CacheL2Flags 0x1001D
CPU1: ICacheL2Size 262144
CPU1: ICacheL2Lines 0
CPU1: ICacheL2LineSize 0
CPU1: DCacheL2Size 262144
CPU1: DCacheL2Lines 0
CPU1: DCacheL2LineSize 0
CPU1: CacheL3Type 0x0
CPU1: CacheL3Flags 0x1001D
CPU1: ICacheL3Size 2097152
CPU1: ICacheL3Lines 0
CPU1: ICacheL3LineSize 0
CPU1: DCacheL3Size 2097152
CPU1: DCacheL3Lines 0
CPU1: DCacheL3LineSize 0
System:>

But still it reports 600Mhz

Quote:

Nothing strange, MorphOS can easily detect a 2nd CPU, just not use it.

Quote:

Most likely a bug in the detection routine, you migth want to run some CPU benchmark to confirm your running at 1.8GHz.

And file a bugreport.

These are the results of AmigaMark:

==========================================================================·
¡ ¡
| C P U R E S U L T S |
| |
·--------------------------------------------------------------------------·
| BogoMIPS ppc-assembler inline |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| 100 Percent at : 5000.0 BogoMIPS |
| Score weight : 80.0% |
| CPU Cache : YES |
| Multitasking : NO |
| |
| Result Result |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| [+] BogoMIPS 2526.45 |
| |
| AmigaMARKS : 40.42 Percent |
! !
·--------------------------------------------------------------------------·
| Dhrystones |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| 100 Percent at : 10000000.0 Dhrystones/sec |
| Score weight : 100.0% |
| Rounds : 5000000 |
| CPU Cache : YES |
| Multitasking : NO |
| |
| Method Result |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| [+] Dhrystones/sec 2287453.27 |
| |
| AmigaMARKS : 22.87 Percent |
! !
·--------------------------------------------------------------------------·
| FPU query [Double Precision] - Al Aburto |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| 100 Percent at : 1000.0 MFlops |
| Score weight : 120.0% |
| CPU Cache : YES |
| Multitasking : NO |
| |
| Method MFlops |
·- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ·
| [+] Integral of df(x)/f(x) 299.41 |
| [+] PI from Taylor Series expansion of atan(1.0) 200.67 |
| [+] Integral of sin(x) from 0.0 to PI/3.0 (Trapazoidal) 487.34 |
| [+] Integral of cos(x) from 0.0 to PI/3.0 (Trapazoidal) 291.12 |
| [+] Integral of tan(x) from 0.0 to PI/3.0 (Trapazoidal) 348.38 |
| [+] Integral of sos(x)*cos(x) from 0.0 to PI/4 (Trap.) 522.81 |
| [+] Int. 0 to sa of 1/(x+1), x/(x*x+1) & x*x/(x*x*x+1 100.26 |
| [+] Int. sin(x)*cos(x)*cos(x) from 0 to PI/3 (Trap.) 504.93 |
| |
| [+] Iteration 512000000 |
| [+] NullTime (usec) 0.0 |
| [+] MFLOPS(1) 248.45 |
| [+] MFLOPS(2) 202.49 |
| [+] MFLOPS(3) 323.74 |
| [+] MFLOPS(4) 452.8 |
| |
| AmigaMARKS : 36.80 Percent |
! !

Sorry, no idea how I can tell now if the system runs with 600MHz or more.
Maybe you can help Kronos?

Tried AmigaMark on my PowerBook (1,67GHz) but results were all over the place compared to yours, even the 2.3GHz G5 go lower results for BogoMIPS while both had much higher Dhrystones/sec compared to your results.

No idea wether this is to different cache sizes or just a sign that the SW is useless....

thanks for testing.
Would it help to repeat the test with cache turned OFF?

Not really into benchmarking, but I just did:

Download lame

Downlaod AKsackl.waw

Placed both in ram:

":lame-3.99.5-morphos/lame :AKsack.wav x.mp3"

24s for the G5
31s for the PowerBook

If you were running at 600MHz you should get about 1 minute, 30s for 1800 (RAM speed is slow on those old G4s so you won't see anything near x3 by trippling clock).

Thanks for the advice.
Did so, result is 39sec. (system wasn't freshly booted)
It seems that the machine isn't really running at 600MHz.

39s seem pretty realistic for 1800MHz compared to the 31 I get at 1670MHz considering that the PB has faster RAM.

Differnt revisions of the "G4" may also play a role.

ok, so there is just the minor bug with the cpu command left.
Which is the best way to report the bug?

Quote:

just replay a 720p video with Mplayer and look whether it stutters and frameskips like mad (600MHz) or gets replayd fluently (above 1.5 GHz).

Quote:

Right click in Ambient and go to Ambient/about Ambient/ chose the "Support" tab there and file the bug report.

Quote:

Not in Ambient's About, but in "About MorphOS...".

Quote:
Uhh, I got 19s here on my PowerBook 1,67!
Code:

thanks for the help Kronos and Zylesea :)

*** closed ***

Das ist die unoptimierte Lame-Version.
Nutze lame_vmx für die Altivec optimierte Version.

Hi Tom,

danke für den Hinweis, es ging lediglich um einen Geschwindigkeitsvergleich unserer beiden Rechner, von daher war es sinnvoll die Selbe Version zu verwenden um keinen Vorteil zu haben.

Additionally, I want to point out that other 3rd-Party-CPU upgrades for the G4 Powermacs will work, too. I had a Newertech 1,6 GHz 7447 upgrade for some years in my Powermac 3,3, and it did good work.

I have a single core sonnet 1.8 ... Shows up fine in Morphos.