Jump to content
1949 posts in this topic

Recommended Posts

OK.

 

mm67 has a q9550 processor (as I do), and a U3D board (I have UD3L).

 

I took mm67's last dsdt.dsl, and editted it so my graphics card was recognized... and changed the PSS method to reflect the values that I saw for my p-states (I'm overclocked to 3.4GHz) that I saw with VoodooMonitor.

 

It compiled fine. It boots fine. Everything works fine, just as it did with the one I rolled.

 

...but CPU-X still doesn't show speedstep working...

 

What the heck?

 

This is just mm67's dsdt.dsl with minimal changes (added a video card definition & only changed the values in Method _PSS).

 

 

mm67's dsdt with videocard added & nudged method _pss

 

==========

afterthought:

hmmmm ... I'm using PC EFI v10.5. Does that disable speedstep for some reason?

This is solved.

 

My speed stepping was working, but was not being detected by CPU-X. I am suspicious that it no longer works in SL as it should. (it only will register the highest/fastest step - 3.4HGz - and doesn't show the changes)

 

But here is how I know that Vanilla SpeedStepping is working for me.

 

1. I installed Voodoomonitor and its kext.

 

2. I installed my basic dsdt.aml that had __no__ speedstep code and booted. I ran Voodoomonitor and top (the CLI load analyzer) and saw that while my system was idle, the CPU speed was stuck @ 3.4 GHz. Here is the screenshot showing that it was idle yet running full bore and at 39 deg celsius.voodoomon-idle-but-no-stepping-and-hot.jpg

 

 

3. I switched to dsdt.aml that does have speedstep code and rebooted. Again I ran Voodoomonitor and top and saw that while my system was idle, the CPU speed was variable, settling at 2.4GHz, and running cooler @ about 32 deg C. voodoomon-showdownstepped-and-cooler-at-idle.jpg

 

4. I removed the VoodooMonitor kext and rebooted. CPU Temps at idle remained about 32 deg C.

 

Big improvement!

 

I'm baffled though that I would be the ___only___ person in the forum that has noticed what seems to be a fault on CPU-X in Snow Leopard.

 

Anyway, now I have vanilla speedstep.

 

Thanks to anyone who helped me out. In the end, the directions on the very front page were __exactly__ on target for me... just minor alterations for my overclocked Quad Q9550 CPU.

OK.

 

mm67 has a q9550 processor (as I do), and a U3D board (I have UD3L).

 

I took mm67's last dsdt.dsl, and editted it so my graphics card was recognized... and changed the PSS method to reflect the values that I saw for my p-states (I'm overclocked to 3.4GHz) that I saw with VoodooMonitor.

 

It compiled fine. It boots fine. Everything works fine, just as it did with the one I rolled.

 

...but CPU-X still doesn't show speedstep working...

 

What the heck?

 

This is just mm67's dsdt.dsl with minimal changes (added a video card definition & only changed the values in Method _PSS).

 

 

mm67's dsdt with videocard added & nudged method _pss

 

==========

afterthought:

hmmmm ... I'm using PC EFI v10.5. Does that disable speedstep for some reason?

 

You can't use CPU-X to monitor speedstepping, use VoodooMonitor or mark-i, or just look if your temperatures are lower with speedstepping code in dsdt.

You can't use CPU-X to monitor speedstepping, use VoodooMonitor or mark-i, or just look if your temperatures are lower with speedstepping code in dsdt.

 

 

It seems true in Snow Leopard.

 

But on my Dell Mini10v, CPU-X shows the stepping happening ... I guess it only works in Leopard. ?

 

mark-i is one I never heard of.

 

thanks mm67

:thumbsup_anim:

kdawg I got it booting again. I forgot to remove openhaltrestart and sleepenabler. After deleting them I could boot into my system with the ICH9 Patch in my DSDT. I added these values because they were the nearest to my System I could find in AppleLPC.kext.

My system is running fine with native Speedstepping but sound is very sloppy as long as I don't move my mouse around.

I changed the PX40 Offset to 3a18 but sloppy sound is still there :/

Moreon the system feels slow when using the native speedstepping compared to using voodoopowermini. Maybe I have a big glitch in my DSDT?

Hi waldo_pepperjack

I'm baffled though that I would be the ___only___ person in the forum that has noticed what seems to be a fault on CPU-X in Snow Leopard.

Lol.. don't take this the wrong way but ApexDE also noticed CPU-X doesn't work correctly in Snow Leopard, that was in post #16 on page 1. I know you're keen to find answers, but a lot has already been covered in this thread.

:blink:

Hi waldo_pepperjack

 

Lol.. don't take this the wrong way but ApexDE also noticed CPU-X doesn't work correctly in Snow Leopard, that was in post #16 on page 1. I know you're keen to find answers, but a lot has already been covered in this thread.

:)

keen

;) or just a bit impatient :D

 

 

Honestly, thanks for pointing that out. I'd thought I'd read things thoroughly, but I guess I overlooked this.

 

'tis true, it is a very "rich" or "dense" thread ... but this has been a good learning experience ...

 

The result is that everything works perfectly, and I can pare back a few more kexts... guess I gotta update my sig now.

Hi! My system works stable but CPU temps are a bit high (48 Celsius) with normal use. Speedstep (AppleIntelCPUpowermanagement) should help lower temps. I'm looking for a safe way to test the system with AppleIntelCPUpowermanagement.kext so in case of KP i can still load NullCPU kext to boot.

 

Running Snow Leopard

My specs: Asus P5Q (ICH10) , E6600, GT7300

 

Here is my dsdt:

 

DefinitionBlock ("./dsdt.aml", "DSDT", 1, "A0993", "A0993001", 0x00000001)
{
   Scope (_PR)
   {
       Processor (CPU0, 0x01, 0x00000810, 0x06) {}
       Processor (CPU1, 0x02, 0x00000810, 0x06) {}
       Processor (CPU2, 0x03, 0x00000000, 0x00) {}
       Processor (CPU3, 0x04, 0x00000000, 0x00) {}
   }

   Scope (_PR.CPU0)
   {
       Method (_PSS, 0, NotSerialized)
       {
           Return (Package (0x02)
           {
               Package (0x06)
               {
                   Zero, 
                   Zero, 
                   0x0A, 
                   0x0A, 
                   0x091A, 
                   Zero
               }, 

               Package (0x06)
               {
                   Zero, 
                   Zero, 
                   0x0A, 
                   0x0A, 
                   0x0714, 
                   One
               }
           })
       }

       Method (_PSD, 0, NotSerialized)
       {
           Return (Package (0x05)
           {
               0x05, 
               Zero, 
               Zero, 
               0xFC, 
               0x04
           })
       }

       Method (_CST, 0, NotSerialized)
       {
           Return (Package (0x02)
           {
               One, 
               Package (0x04)
               {
                   ResourceTemplate ()
                   {
                       Register (FFixedHW, 
                           0x01,               // Bit Width
                           0x02,               // Bit Offset
                           0x0000000000000000, // Address
                           0x01,               // Access Size
                           )
                   }, 

                   One, 
                   0x9D, 
                   0x03E8
               }
           })
       }
   }

   Scope (_PR.CPU1)
   {
       Method (_PSS, 0, NotSerialized)
       {
           Return (^^CPU0._PSS ())
       }

       Method (_PSD, 0, NotSerialized)
       {
           Return (^^CPU0._PSD ())
       }

       Method (_CST, 0, NotSerialized)
       {
           Return (Package (0x04)
           {
               0x03, 
               Package (0x04)
               {
                   ResourceTemplate ()
                   {
                       Register (FFixedHW, 
                           0x01,               // Bit Width
                           0x02,               // Bit Offset
                           0x0000000000000000, // Address
                           ,)
                   }, 

                   One, 
                   Zero, 
                   0x03E8
               }, 

               Package (0x04)
               {
                   ResourceTemplate ()
                   {
                       Register (FFixedHW, 
                           0x08,               // Bit Width
                           0x00,               // Bit Offset
                           0x0000000000000414, // Address
                           ,)
                   }, 

                   0x02, 
                   One, 
                   0x01F4
               }, 

               Package (0x04)
               {
                   ResourceTemplate ()
                   {
                       Register (FFixedHW, 
                           0x08,               // Bit Width
                           0x00,               // Bit Offset
                           0x0000000000000415, // Address
                           ,)
                   }, 

                   0x03, 
                   0x55, 
                   0xFA
               }
           })
       }
   }

   Scope (_PR.CPU2)
   {
       Method (_PSS, 0, NotSerialized)
       {
           Return (^^CPU0._PSS ())
       }

       Method (_PSD, 0, NotSerialized)
       {
           Return (^^CPU0._PSD ())
       }

       Method (_CST, 0, NotSerialized)
       {
           Return (^^CPU1._CST ())
       }
   }

I just followed the process on a GA-EP45-DS3R and tough I got speedstepping working, the windows started to not to move smoothly when I drag them. Also boot time is longer (just after the video driver is initialized, the screen remains black for a while, and then blue, while with no speedstep you go directly from the gray screen to the desktop).

 

Any idea?

 

Also I had to install netkas hpet kext, as I don't know why (my dsdt already has an hpet device with proper irq's) appleintelcpumanager kext complained there was no hpet device.

I just followed the process on a GA-EP45-DS3R and tough I got speedstepping working, the windows started to not to move smoothly when I drag them. Also boot time is longer (just after the video driver is initialized, the screen remains black for a while, and then blue, while with no speedstep you go directly from the gray screen to the desktop).

 

Any idea?

 

Also I had to install netkas hpet kext, as I don't know why (my dsdt already has an hpet device with proper irq's) appleintelcpumanager kext complained there was no hpet device.

 

My board is a close cousin of yours - I have the GA-EP45-UD3L. Which CPU do you have? Me: Q9550@3.4GHz.

 

What sort of video setup do you have? EFI strings, DSDT/gfx or something else?

 

Is Quartz Extreme/CoreImage enabled on your box?

 

I haven't updated my sig, but now literally everything is via DSDT:

  • MSI GeForce 8400 GS 256MB
  • ALC888 HDEF audio
  • gigabit ethernet (probably same chip as yours)
  • USB fix so that sleep works correctly
  • CMOS patch so that CMOS/BIOS isn't reset (though this can be fixed w/ a BIOS update, perhaps)
  • Vanilla SpeedStep, via the directions on the very first post of this thread

My board is a close cousin of yours - I have the GA-EP45-UD3L. Which CPU do you have? Me: Q9550@3.4GHz.

 

What sort of video setup do you have? EFI strings, DSDT/gfx or something else?

 

Is Quartz Extreme/CoreImage enabled on your box?

 

I haven't updated my sig, but now literally everything is via DSDT:

  • MSI GeForce 8400 GS 256MB
  • ALC888 HDEF audio
  • gigabit ethernet (probably same chip as yours)
  • USB fix so that sleep works correctly
  • CMOS patch so that CMOS/BIOS isn't reset (though this can be fixed w/ a BIOS update, perhaps)
  • Vanilla SpeedStep, via the directions on the very first post of this thread

 

Hey Waldo,

 

Do you mind posting your dsdt? I'm interested in your OC settings for the Q9550

Thanks for the answer I will give it a try :)

 

:( Didn't work. RTC already had no IRQ's, removed the one from TMR, but the 1 minute delay between the gray screen and the desktop was still present in addition to skippy window dragging.

 

I disabled back appleintelcpumanager, and everything worked fine again...but no speedstep obviously.

 

I'm using a vanilla 10.6.2 AHCI install in my GA-EP45-DS3R with a Q6600 cpu (not overclocked), the VGA (8600GT QE&CI working) and Ethernet as internal is set up via EFI string.

:( Didn't work. RTC already had no IRQ's, removed the one from TMR, but the 1 minute delay between the gray screen and the desktop was still present in addition to skippy window dragging.

 

I disabled back appleintelcpumanager, and everything worked fine again...but no speedstep obviously.

 

I'm using a vanilla 10.6.2 AHCI install in my GA-EP45-DS3R with a Q6600 cpu (not overclocked), the VGA (8600GT QE&CI working) and Ethernet as internal is set up via EFI string.

 

Do you have the Disabler.kext installed with the AppleUpstreamUserClient.kext in its blacklist? This can cause the symptoms you describe on the 8??? series of cards.

Do you have the Disabler.kext installed with the AppleUpstreamUserClient.kext in its blacklist? This can cause the symptoms you describe on the 8??? series of cards.

 

Yep I let AppleUpstreamUserClient.kext disabled since I heard it casued skippy video with latests QT versions.

Hey Waldo,

 

Do you mind posting your dsdt? I'm interested in your OC settings for the Q9550

yep, here it is ...

dsdt-waldo-kitchensink-and-speedstep.dsl

 

When I said I had _everything_ in it, well, that was an overstatement, but the list I have up is accurate.

 

I haven't done anything beyond:

- initial creation w/ the handy Using DSDT with the Gigabyte ... howto by blackosx (did the ethernet/gigabit LAN, GFX, CMOS fix, USB-Sleep-Unmount fix, ALC888 Legacy HDEF)

 

- some more USB fixes, including MasterChief's/UnsubscribeMe's USB0->UHC1 suff

 

- the Vanilla SpeedStep, pretty much just as is shown in the very 1st post of this thread.

 

BTW, my bus is @ 400MHz, using the 8.5multiplier. I have the ArticPro 7 Coolder... I don't see nay temperatures over 41C.

Hello guys. Found dsdt for my laptop, but with different CPU, now speedstep working wrong :)

My CPU is core2duo T5450, tried to modify dsdt with this guide, no success :huh:

Screenshot of PStateChanger and dsdt attached, can anyone help to modify PR section for my CPU? Or may be has dsdt for this CPU?

post-204570-1263017179_thumb.png

dsdt.zip

Here are the details of your CPU.

http://ark.intel.com/Product.aspx?id=30787...pec-codes=SLA4F

 

What excat do you mean with "works wrong "

The MHZ Pstates (with 166 MHZ FSB *6...*10(=0Ahex)) seem to be OK for your CPU !

 

You may only change the VID (mVolts).

Intel Informastion for your CPU VID Voltage Range 1.075V-1.250V (depends on CPU , 1075 maybe to less for some CPUs).

What excat do you mean with "works wrong "

The MHZ Pstates (with 166 MHZ FBS *6...*10(=0Ahex) seem to be OK for your CPU !

 

You may only change the VID (mVolts).

 

thx, I'll try. "works wrong" means CPU cooler with VoodoPState than with native speedstep.

yep, here it is ...

dsdt-waldo-kitchensink-and-speedstep.dsl

 

When I said I had _everything_ in it, well, that was an overstatement, but the list I have up is accurate.

 

I haven't done anything beyond:

- initial creation w/ the handy Using DSDT with the Gigabyte ... howto by blackosx (did the ethernet/gigabit LAN, GFX, CMOS fix, USB-Sleep-Unmount fix, ALC888 Legacy HDEF)

 

- some more USB fixes, including MasterChief's/UnsubscribeMe's USB0->UHC1 suff

 

- the Vanilla SpeedStep, pretty much just as is shown in the very 1st post of this thread.

 

BTW, my bus is @ 400MHz, using the 8.5multiplier. I have the ArticPro 7 Coolder... I don't see nay temperatures over 41C.

 

You might want to check again your UHCI and EHCI devices. UHCI devices work like that but these parts in in definitions are useless for ICH10:

                Method (_DSM, 4, NotSerialized)
               {
                   Store (Package (0x02)
                       {
                           "device-id", 
                           Buffer (0x04)
                           {
                               0x34, 0x3A, 0x00, 0x00
                           }
                       }, Local0)
                   DTGP (Arg0, Arg1, Arg2, Arg3, RefOf (Local0))
                   Return (Local0)
               }

 

Those are only useful if you want to make for example ICH9 UHCI devices look like ICH10, but you already have an ICH10 board. And your EHCI devices now have no patches at all for usb removal after sleep problem. Plug in a usb stick, put system to sleep, wake up and see what happens.

You might want to check again your UHCI and EHCI devices. UHCI devices work like that but these parts in in definitions are useless for ICH10:

...

Those are only useful if you want to make for example ICH9 UHCI devices look like ICH10, but you already have an ICH10 board. And your EHCI devices now have no patches at all for usb removal after sleep problem. Plug in a usb stick, put system to sleep, wake up and see what happens.

mm67, you are right. :P Thanks for that spot ...

OK, so much for everything being done.

 

I guess I've got (yet some more) reading to do.

 

Tips/hints? A pointer to the/a relevant group of posts would be great.

mm67, you are right. :P Thanks for that spot ...

OK, so much for everything being done.

 

I guess I've got (yet some more) reading to do.

 

Tips/hints? A pointer to the/a relevant group of posts would be great.

 

Just read the Gigabyte fixes thread some more, you should find several examples there.

Just read the Gigabyte fixes thread some more, you should find several examples there.

yup, will do ... I also see this now trick retail drivers by changing device ide (e.g. USB) - post RE patches to p35 vs. p45 boards ...

Also I should look at how both you and kdawg manage it in your dsdt setups.

yup, will do ... I also see this now trick retail drivers by changing device ide (e.g. USB) - post RE patches to p35 vs. p45 boards ...

Also I should look at how both you and kdawg manage it in your dsdt setups.

 

Thanks for that.

 

There are also a couple of extra simple additions you may want to consider as well.

The latest boot files have the restart fix incorporated into them here. You should be able to remove OHR kext. With my board I only use FakeSMC, LegacyHDA and IOBlockStorage kexts (bonjour is via ifconfig).

 

Check out my dsdt, its pretty well annotated with the fixes etc, compare it to yours, wake via apple usb keyboard has also been fixed.

 

I'm still tinkering with my pstates and they are for a Q9550.

 

Credit goes to mm67 for his excellent work.

 

 

EDIT:

Your link to your dsdt appears to be broken :(

Current___complete.zip

×
×
  • Create New...