Jump to content

both cores (T7200) working!

krille

By krille,
in OSx86 10.4 (Tiger)

 Share
56 posts in this topic

Recommended Posts

krille
krille
Posted
Posted

now xbench: ~160 @ dell inspiron 9400 :o

 

i reinstalled osx 10.4.8 and does not set the kernel flags. System was slow and jumpy. but after wake up from sleep modus everthing worked fine and smooth. anybody else recordnized this?

 

what is changing by waking up from sleep?

 

please help to get us run both cores on every start...

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted

Well I'm using the Mar 24 netkas SSE3 HPET kernel with dsmos.kext at the moment and Sleep mode will not activate.

NvidiaEFI also does not work for me with my GeForce 7800 Go, it used to be fine with 10.4.8 though.

Strange... perhaps there is something different with the JaS combo update, because I actually used the official Apple update and then replaced AppleSMBIOS/IONetworking/mach_kernel myself using Macdrive.

 

I heard the latest JaS 10.4.9 Combo update retains the 10.4.8 Nvidia kexts as default, can someone confirm this for me?

Link to comment
Share on other sites
krille
krille
Posted
  • Author
Posted

same to me - using jas 10.4.8 SSE intel only install with mifki kernel - sleep works right after first boot

after jas 10.4.9 update sleep works too.

 

only using applesmbios from paulicat.

 

but we need to figure out what happend on wake up so we can use both cores permanently.

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted
same to me - using jas 10.4.8 SSE intel only install with mifki kernel - sleep works right after first boot

after jas 10.4.9 update sleep works too.

 

only using applesmbios from paulicat.

 

but we need to figure out what happend on wake up so we can use both cores permanently.

 

I totally agree.

 

If this wake state is doing something to fix dual core then we desperately need to know what it is.

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted

Well, for experimentation I replaced my 10.4.9 Nvidia kexts with the one from 10.4.8 and it 'almost' managed to sleep, but not quite.

It just hangs with a black screen when I close the laptop lid, which is the same thing Titan.kext does when I force sleep via the power button with the 10.4.9 Nvidia kexts.

I'm gathering JaS' 10.4.9 Combo update does not mess with the Nvidia kexts and leaves the 10.4.8 versions there, but clearly there is something else involved which is preventing me from fully sleeping.

Those of whom have JaS' 10.4.9 Combo installed, do you see Rotation enabled in the System Profile? is your GeForce.kext version 1.4.52.6 (8.7.5f3)?

 

Regards,

 

Pawel.

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted
@devilhood

My Geforce.kext is 1.4.40. Date: 4.1.06

Do you mean with rotation the fast change of users in logon options ? (I have a german MacOS)

If so it is disabled.

@fruitylooper

uname -a in Console.

 

Ok, cool.

 

That confirms for me that the JaS 10.4.9 update uses only 10.4.8 Nvidia kexts.

When I said rotation, I meant the rotation of your display. This is a new feature of the 10.4.9 kexts, allowing you to change the rotation in display preferences.

 

I wonder what else it keeps.

Link to comment
Share on other sites
flibblesan
flibblesan
Posted
Posted

Stephen: What exactly did you install? What DVD release, and how did you upgrade to 10.4.9? This information will help us find out how to get sleep working properly, and dual core =)

 

edit: forget me, my brain isn't fully working today.. I saw your post above heh.

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted

Prasys told me that the dual-core stutter bug exists because of the 2nd core going to sleep.

Putting the laptop in sleep mode and waking it properly re-initializes the cores so they work simultaniously.

There could be a solution by using a good ole' soldering iron, but I very much doubt a software solution exists unless one is willing to crack the Dell BIOS.

 

To make this clear, this issue unfortunately cannot be solved in the kernel ;)

Link to comment
Share on other sites
krille
krille
Posted
  • Author
Posted
Prasys told me that the dual-core stutter bug exists because of the 2nd core going to sleep.

Putting the laptop in sleep mode and waking it properly re-initializes the cores so they work simultaniously.

There could be a solution by using a good ole' soldering iron, but I very much doubt a software solution exists unless one is willing to crack the Dell BIOS.

 

To make this clear, this issue unfortunately cannot be solved in the kernel :P

 

But this very unlogical. kernel initializes cores wrong on boot and corectly after wake up. so it's same kernel doing that =)

it's not a real bug but i think it can be solved by the kernel.

Link to comment
Share on other sites
quasar_x
quasar_x
Posted
Posted
But this very unlogical. kernel initializes cores wrong on boot and corectly after wake up. so it's same kernel doing that =)

it's not a real bug but i think it can be solved by the kernel.

 

I agree. If others can reproduce the above (getting dual-core working after waking up), there must be ways to "simulate" such an event by means of software and without touching the BIOS.

 

I am also on an Inspiron E1505/6400 with 10.4.9 (latest netkas), but haven't yet managed to wake it up after activating standby.

Link to comment
Share on other sites
Stephan
Stephan
Posted
Posted

After boot the cursor is in the upper left corner of the screen (just over the apple), so that it is no problem to activate sleepmode with the arrow keys (mouse is not working).

If you don't go to sleepmode, a black screen is appearing after about 30sec.

 

A automatic sleep after boot may be possible (perhaps there is something like /proc/acpi/sleep in linux)

A little startup skript may do the job (sorry I know only some linux, I'm not a Mac crack)

But you still will have to push the button for wakeup.

In linux a solution over RTC for auto wakeup is possible , but in MacOS ?

 

I hope we will find a more simple solution.

 

I think Titan is loaded twice at startup (only dual core boot), perhaps this is causing some problems too.

Link to comment
Share on other sites
devilhood
devilhood
Posted
Posted
I agree. If others can reproduce the above (getting dual-core working after waking up), there must be ways to "simulate" such an event by means of software and without touching the BIOS.

 

I am also on an Inspiron E1505/6400 with 10.4.9 (latest netkas), but haven't yet managed to wake it up after activating standby.

 

The kernel isn't actually doing anything wrong. It is purely BIOS related, with the way in which Power Management (S3) has been set.

Waking from sleep re-initializes both cores simultaniously, stopping this whole 2nd core going to sleep problem.

Dell are pretty strict with their settings and lock all functionality of their BIOS the moment a Service Tag number is entered in.

 

You cannot fix something like this with software, unless you can script a sleep/wake state during the boot procedure, though I'm unsure how viable that would be.

The best possible route is to try and reverse engineer the BIOS itself, because then we would also be open to the possibility of enabling AHCI too.

We'd essentially have a Macbook Pro clone if that was possible ;)

Link to comment
Share on other sites
quasar_x
quasar_x
Posted
Posted
You cannot fix something like this with software, unless you can script a sleep/wake state during the boot procedure, though I'm unsure how viable that would be.

 

That's exactly what I meant. It might not be as elegant as hacking the BIOS, but IMHO more realistic to achieve.

 

The best possible route is to try and reverse engineer the BIOS itself, because then we would also be open to the possibility of enabling AHCI too. We'd essentially have a Macbook Pro clone if that was possible :)

 

Ok, anybody volunteering to flash such a modified BIOS? ;)

Link to comment
Share on other sites
sj1069
sj1069
Posted
Posted

thanks everyone for your efforts.

after reading this thread, i reinstalled my system again as per the following order:

1. JAS 10.4.8 DVD

2. JAS 10.4.9 combo

3. Titan.kext

4. IOnetwork from 10.4.6

5. Speedstep.mpkg downloaded from this forum

6. AppleHDAPatcher_v1.15

Then i tried the sleep and wake up, which both succeeded.

Then i enabled both cores in the bios, and proved that after sleep and wakeup, the system went smoothly with both cores enabled.

However, when the second time i rebooted the system with both cores enabled, the system directly went into "four language page" requiring for restart. This repeated for 3 times, after which i gave up the duo core attempt and disabled one core in the bios.

Unfortunately, i was no longer able to mount dmg files again. (i believe it has something to do with the speedstep.mpkg, because when i reinstalled the sytem again without it, everying was ok)

in addition, there is another thing. about the applehdapatcher. I don't know whether you have a stuttering of sound at the very beginning. but whenever i play music, the first 2-3 seconds suffer a stuter. but it will disappear later on. Interestingly, during the experience with the speedstep.mkpg installed (which casued unable to mount dmg files) the sound was perfect.

 

above is my experience this afternoon.

 

meanwhile, i am wondering which is more difficult, studying the possibility of rewriting a bios or trying to hack the kernal? lol

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...