NEW Intel P55 Mobos & Core i7 860 / i5 750 CPU LGA1156

[dtessier]

Can other users of GA-P55-UD2 boards confirm/report their findings with RAM speed?

 

I've got the 12X memory multiplier in the BIOS on my P55-UD2, and that's with BIOS F4.

[aushack]

Rich,

 

I am glad my DSDT provided a good starting point for you. I would suggest you use DSDTSE and go through the code to check that all the patches you want are there.

 

After you boot, open DSDTSE and Extract DSDT, this way you know that you are patching the DSDT file that is actually been used. Go through tonymac's list of patches and make sure they are all there.

 

For audio, I finally managed to get AppleHDA to work following treky's post on page 33 of this thread. The sound is much cleaner than VoodooHDA with no cracking. Try and see if that method works for you, make sure you use AppleHDA version 1.7.4a1 (from 10.6 and 10.6.1).

 

Sleep almost works. It works if you manually put the computer to sleep but not if you let the computer sit and wait until it goes to sleep on its own.

 

hey thanks sorry for the late reply. ended up getting the audio working by putting LegacyHDA in System Extensions and reinstalling your UD4P DSDT from tony's database. works great now! but how is putting LegacyHDA in Extra's Extensions different from in System's? and my Output in Sound preferences is different to the pic posted on tony's blog, Internal Speakers is listed but no Headphone and there is 4x Line outs?

 

also i managed to add Speed Step working and sleep now works with the UD4P posted on tony's blog so big thanks to everyone who contributed! but i can only wake from sleep by pressing the power button and auto sleep doesnt work, is there any chance we can get wake with the keyboard or mouse and auto sleep working in the future? and i can only get sleep working by ticking the last option in Energy Saver, is this a danger after a power failure? or just a luxury?

 

 

EDIT 1/10: lol picimg changed my image! cbf changing it, was just a picture of Energy Saver preferences with the last option ticked.

[Elventanis]

Yep, the kexts in that post both in S/L/E and the DSDT listed for the UD4P on Tony's blog (this is why the link was to the specific post)

 

I tried that, and my sound switched from the back speakers to the center and the subwoofer only. Nothing from the other speakers. There has to be something else going one. I am using Logitech speaker setup with 3 connectors at the back of the machine.

 

I tried this with the DSDT on Tony's site and also my own DSDT that I have been modifying. The same with both.

 

Not sure what else I can check to be honest

[Bansaku]

Well, I tested your plist file and even used your posted dsdt file to be sure I would not miss something, I also used a macpro4,1 smbios.plist file and the results are bad (40% of what I can get without your patch in 64bits and 32 bits modes).

Something might be different in your config., it might not only be a question of dsdt and macpro4,1 config and nullcpm removal ...

Of course I run a native applepowermanagement with turbo detection and doesn't have any disabler ...

 

I am including my current DSDT, as the last one I posted was the compile that was still in progress (not know at the time of posting); I forget what I changed but it resolved my KP after wake. What benchmarks are you basing the 40% dip on?

 

For comparison here's what my system uses that's mod'd/non-vanilla:

 

-Chameleon 2 R3 with the modified boot from Tony

 

In Extras/Extensions folder

-FakeSMC 2.5

-EvOreboot.kext

-JmicronATA.kext

 

In System/Extensions

-my modded AppleHDA.kext

-mod'd IOPCIFamily.kext

-mod'd MacPro4,1.plist in IOPlatformPluginFamily.kext/contents/plugIns/ACPI_SMC_PlatforPlugin.kext/contents/Resources

-RealtekR1000SL.kext

 

-In smbios.plist SMproductname MacPro4,1

 

-My bios version is F3f

dsdt.aml.zip

[TR0LL]

Just wanted to say that my final wave of parts arrived in the mail today, and I now have everything except for the stupid cpu cooler that got held up in shipping due to 'adverse weather conditions'.

Started putting things together tonight but couldn't get too far due to the missing cpu cooler, and I'm excited that its finally happening after so much planning and research .

Now I have to figure out what the hell I'm doing as far as setting things up, and hopefully things go relatively smoothly. Feeling like an uber noob now that its real! Deep breaths, it'll all work out...

[aushack]

-mod'd IOPCIFamily.kext

-mod'd MacPro4,1.plist in IOPlatformPluginFamily.kext/contents/plugIns/ACPI_SMC_PlatforPlugin.kext/contents/Resources

 

whats the first one for? is the second to make the processor appear correctly and make your hack appear as a Mac Pro not an iMac, if so is it an appearance mod or the specs will be read correctly thruout the system?

[profdre]

Very odd, according to Gigabyte's own website the you should be good to go up to 2200MHz+ (OC'd I do realize that). I don't see why you do not have the 12x option. Can other users of GA-P55-UD2 boards confirm/report their findings with RAM speed? Perhaps you need to update your BIOS. I had to on my 780 chipset AMD system to get my memory to run at 1066MHz. I feel you pain. I paid good money for my OCZ Reaper DDR2 and I wanted to smash things when I couldn't get the advertised speed. As well, understand why you are overclocking. I read to many OCing forums and people have some crazy OC concepts (some I have tried myself). I see your frustration.

 

I think you'll only have the 12x option when you're using an i7. My highest multiplier is also "only" 10. But given the fact that changing the RAM speed will change the overall speed very slightly (in synthetic benchmarks), forget your RAM speed. Even the iMac's 1066 MHz RAM is fine (and I didn't see a large gap compared to my 1440 MHz RAM. But if you like to play and fine-tune: change the RAM latency...if your RAM is underclocked, you can give it more aggressive timings...I'm using 8-8-8-25 (1333 MHz) instead of 9-9-9-25 (1600 MHz) (giving me great 30 pts more in Geekbench )

[mulan]

Hi all

I have the P55m-UD2 motherboard with BIOS F3 installed.

Is it a good idea to install the newest version of the BIOS (F5) before installing SL?

[YoYelloW]

Hi all

I have the P55m-UD2 motherboard with BIOS F3 installed.

Is it a good idea to install the newest version of the BIOS (F5) before installing SL?

 

 

i updated to the new F5 and it works, i posted the update file in this thread some pages ago, just look it up,

 

it boots alot faster in bios with this new version

[rekursor]

I am including my current DSDT, as the last one I posted was the compile that was still in progress (not know at the time of posting); I forget what I changed but it resolved my KP after wake. What benchmarks are you basing the 40% dip on?

../..

Well except the modded iopcifamily.kext and the platformUUID kext that I prefer to keep for now, we have extremely similar configurations.

I compared the DSDT you gave me with your previous, this is the one that I sent to tony with the USB currents and the IDE0->SATA and IDE1 removed

That's nice to know that it solves the wake KP, pb that I never had btw as this is the DSDT I use for quite a while.

I love what we are doing here, really a team work

 

Concerning the macpro4,1 mod, I don't really need more power though because the few tests I ran show about 9750 in 64bits and 8350 in 32bits average, which is more or less what you get -3% right ? So I think I will stop here this investigation.

[negotiator]

Windows drive boots whether the SL drive is attached or not. It works just fine. I just hit F12 when I boot and select the drive. However, the SL drive only works when the Win 7 drive is unplugged. The SL drive is the first in priority in the BIOS. It will boot fine with the Win 7 drive unplugged. It hangs when I attach the Win 7 drive.

 

In case any one else is having a similar problem, I solved this by just reinstalling Windows 7 (make sure you unplug the Snow drive before installing Win 7). Now, I have Snow working great on the Sata drive and Win 7 on the IDE and I can boot both from PC EFI boot loader.

 

However, I am having a problem with my BIOS clock. Whenever I boot into Snow, Snow changes the clock in the BIOS to about 8 hours ahead of the current time. The time in Snow is correct but when I boot back into Windows, the time gets all messed up. The timezones are the same in both. Nothing else is reset in the BIOS or CMOS other than the clock. All my other settings as to boot order, etc. stay the same.

 

I have Tony's most recent DSDT installed for my GA-P55M-UD2 board (Core i5 750 and Nvidia 9500GT). Anyone else having a similar problem or a solution.

[tonymacx86]

However, I am having a problem with my BIOS clock. Whenever I boot into Snow, Snow changes the clock in the BIOS to about 8 hours ahead of the current time. The time in Snow is correct but when I boot back into Windows, the time gets all messed up. The timezones are the same in both. Nothing else is reset in the BIOS or CMOS other than the clock. All my other settings as to boot order, etc. stay the same.

 

A known issue- I've had the same problem- you have to set up a script in Windows that changes the clock to time.gov or something at login. I did it years ago, but I don't remember exactly how, and haven't fixed it recently. I'm pretty much 99.9% in OS X these days

[profdre]

Hey, I had 3 kernel panics recently and I can say for sure that it happened since I changed the DSDT.aml (got it from tony's blog)...or could it be that using the sleep mode is the reason? One kernel panic during reading a web page (BOINC was running also), one in the Bioshock Demo and one as I was on the Adobe Flash page...

CPU is not overclocked and Windows 7 seems to run more stable (okay I use it less also)...In my Extensions folder there are only EvOreboot.kext, fakesmc.kext and PlatformUUID.kext...

Or could the reason be that I changed the RAM settings from Turbo to Extreme and lowered the CL from 9 to 8 ?

I'm just running a prime stress test in OS X to check that...

 

Should I remove Platform.UUID.kext?

 

Thanks!

[Jared Kipe]

I've got the 12X memory multiplier in the BIOS on my P55-UD2, and that's with BIOS F4.

 

 

Odd, one of the first things I did was upgrade to F5 bios (from the file posed by Yellow) and 10X is the highest mine goes. MAYBE it is because I'm on an i5 instead of i7. But thats a silly limitation if its true.

 

 

Regardless, I get bad performance when using everything set down to stock. without NullCPU..

 

 

Oh and for anyone who's still having high pitched whining when moving windows and what not. Apparently this can be caused by many things, and the CPU going into C states is one of them. I had my system set to stock 133MHz fsb and multiplier 20 with NullCPU and had the noise really badly. Messed around with enabling/disabling the various C states and couldn't get rid of the noise. Went to XMP profile so 160MHz FSB and Multiplier 17, and the noise went away.

 

So if you have the noise, just mess with your various CPU/RAM related parameters.

[HerbM]

The rest of the components for my build should be here tomorrow.

 

I went with the P55A-UD3. Hopefully you all can help me get it working

 

My 30GB Vertex SSD won't be hear until Monday. Can I go ahead and install Snow Leopard on my big HD and then clone it to the SSD when it arrives?

 

i7-860

P55A-UD3

1.5 TB Samsung

H50 Water Cooler

4GB G.SKILL DDR3 1600 (PC3 12800) Model F3-12800CL9D-4GBNQ

MSI N250GTS TwinFrozr 1G OC GeForce GTS 250 1GB 256-bit GDDR3

Cooler Master RC 690 case

OCZ ModXstream 600W PS

Sony Optiarc 24X DVD/CD Rewritable Drive Black SATA Model AD-7240S-0B

[profdre]

Odd, one of the first things I did was upgrade to F5 bios (from the file posed by Yellow) and 10X is the highest mine goes. MAYBE it is because I'm on an i5 instead of i7. But thats a silly limitation if its true.

 

 

Regardless, I get bad performance when using everything set down to stock. without NullCPU..

 

 

Oh and for anyone who's still having high pitched whining when moving windows and what not. Apparently this can be caused by many things, and the CPU going into C states is one of them. I had my system set to stock 133MHz fsb and multiplier 20 with NullCPU and had the noise really badly. Messed around with enabling/disabling the various C states and couldn't get rid of the noise. Went to XMP profile so 160MHz FSB and Multiplier 17, and the noise went away.

 

So if you have the noise, just mess with your various CPU/RAM related parameters.

 

Your PSU could also be the source of that whining. It's a well known issue at Power Mac G5's. Mine is emitting these sounds when the machine is idle (the changing amount of currents is causing the toroids to oscillate?!) When the CPU is maxed out, these sounds disappear.

[Bansaku]

Hey, I had 3 kernel panics recently and I can say for sure that it happened since I changed the DSDT.aml (got it from tony's blog)...or could it be that using the sleep mode is the reason? One kernel panic during reading a web page (BOINC was running also), one in the Bioshock Demo and one as I was on the Adobe Flash page...

CPU is not overclocked and Windows 7 seems to run more stable (okay I use it less also)...In my Extensions folder there are only EvOreboot.kext, fakesmc.kext and PlatformUUID.kext...

Or could the reason be that I changed the RAM settings from Turbo to Extreme and lowered the CL from 9 to 8 ?

I'm just running a prime stress test in OS X to check that...

 

Should I remove Platform.UUID.kext?

 

Thanks!

 

Does this happen only after sleep?

 

 

I compared the DSDT you gave me with your previous, this is the one that I sent to tony with the USB currents and the IDE0->SATA and IDE1 removed

That's nice to know that it solves the wake KP, pb that I never had btw as this is the DSDT I use for quite a while.

I love what we are doing here, really a team work

 

Yessir! It's like being pioneers exploring the new frontier.

 

I had to laugh though. The DSDT I uploaded was from my Test SL install. It was labeled in Red (what I do to all my current working DSDT for quick reference) so I assumed it was mine. ;P

[Jared Kipe]

Your PSU could also be the source of that whining. It's a well known issue at Power Mac G5's. Mine is emitting these sounds when the machine is idle (the changing amount of currents is causing the toroids to oscillate?!) When the CPU is maxed out, these sounds disappear.

 

 

Or... it could be because it is when idle or not working hard that the CPU enters C states, when CPUs are at full load not entering C states at all.

 

I have a hard time believing that just because some PMG5 PSUs were defective that that is what a lot of us are experiencing now... on a wide array of PSUs. Probably better to look at the components that we have that are identical.

[Jared Kipe]

Ah yes, I remember now. Let me ask you this, do you care about speed (and the loss of sleep) or do you want to get your system as vanilla as possible? Overclocking is great and all (having my RAM at 2000MHz vs 1600MHz was nice) but using the DSDT from Tony's blog will get you {censored} Geekbench (and overall performance of the OS) if you are overclocking. This is due to the HPET fix. If you use only the CMOS reset and USB/SATA fix in your DSDT, and obviously using NullCPUPM.kext, you should get around 10 000 easy. However, from my experience using Handbrake to stress CPU (it will use ALL available CPU power to compress the video) all I see is temps reaching 80'c and a pitiful 2 minutes off the total compression time. Sure, gaming is faster as well, but PS3s are only $299 now.

 

 

Edit: 2 minutes equals about 14%

Edit 2: You can probably get rid of PlatformUUID.kext now (please correct me if I am wrong people), and I recommend EvOreboot.kext to OpenHaltRestart.kext, just because it's community based rather than Psystar {censored}.

Edit 3: What's the native speed of your RAM?

 

I'd take sleep and lower temps over speed IF the speed was acceptable.

 

I have spent the day running a real test. I ripped Quantum of Solace with my bluray burner and makemkv. Then I use the newest handbrake and the "High Profile" to compress said 27GB movie. When it reaches 10% I mark down the time it took (using a real life stopwatch) and the average FPS that handbrake tells me. Then I stop it, try to watch the file it created, delete the created file, reboot and start over.

 

Run 1

30.5 minutes, average 8.31 fps

CPU temps 59-62ºC.

w NullCPUPowerManagement

XMS profile (ram@1600MHz, CPU@2.72GHz) 9-9-9-24 memory timing

 

Run 2

55.8 minutes, average 4.54 fps

CPU temps 51-54ºC

w/o NullCPUPowerManagement.kext

XMS profile (ram@1600MHz, CPU@2.72GHz) 9-9-9-24 memory timing

 

Run 3

68 minutes, average 3.68 fps

CPU temps 35-37ºC

w/o NullCPUPowermanagement.kext

everything to stock so it is (ram@1333MHz, CPU@2.66GHz (hopefully 2.8GHz if turbo boost is working)) 8-8-8-21 timing

 

 

 

What could be worse?? Not being able to watch my newly compressed video file after run 2 or 3, it just stutters ever 5 seconds of smooth playback. After run 1 (and after run 3 and resetting my settings) I can play the video just fine. Smooth as could be. Is having 10-20ºC off my CPU clock worth a STAGGERING 50% raw performance reduction?!? Not to me.

[rekursor]

../..

Run 1

30.5 minutes, average 8.31 fps

CPU temps 59-62ºC.

w NullCPUPowerManagement

XMS profile (ram@1600MHz, CPU@2.72GHz) 9-9-9-24 memory timing

 

Run 2

55.8 minutes, average 4.54 fps

CPU temps 51-54ºC

w/o NullCPUPowerManagement.kext

XMS profile (ram@1600MHz, CPU@2.72GHz) 9-9-9-24 memory timing

../..

Jared, just two remarks:

1. You might be interested to test also the Run 1 with no overclocking and see if it's really worth it (with no overclocking u can also set your ram in turbo mode, if you have fast cas memory(7-7-7-21), use the Profile1 option as well)

2. The problem with your test IMHO is that you did not troubleshoot yet why u get {censored} perfs with vanilla power management from the results I read in your previous posts. You should at least get the average imac i5 geekbench results. For instance, my i7 config gives me more or less the average results imac 11,1 core i7, so it's consistent with my hardware system.

[Elventanis]

I also had a kernel panic because of CPUPowerManagement with my vanilla kernel in 10.6.2, so:

I installed the (10.6 version of) Disabler.kext from my leopard partition,

checked the permissions,

moved all extra kexts to /System/Libraries/Extensions (to avoid conflicts with leo) and changed their color category so that I can handle them easily,

changed the com.apple,Boot.plist as follows:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>Kernel</key>
<string>mach_kernel</string>
<key>Kernel Flags</key>
<string>-v arch=i386</string>
<key>ForceHPET</key>
<string>y</string>
<key>ethernetbuiltin</key>
<string>y</string>
<key>GraphicsEnabler</key>
<string>y</string>
<key>USBBusFix</key>
<string>y</string>
<key>Timeout</key>
<string>1</string>
</dict>
</plist>

 

I also had to adapt my DSDT.aml file HDEF to match ALC888 as opposed to tony's ALC888B(aka ALC887)

and I have choosen a different (ALC888 3out/2in HDA) LegacyHDA.kext file you can find in this forum from the 'series of LegacyHDA'

 

The changed HDEF definition in my DSDT.aml is as follows:

            Device (HDEF)
           {
               Name (_ADR, 0x001B0000)
               Method (_PRW, 0, NotSerialized)
               {
                   Return (Package (0x02)
                   {
                       0x0D, 
                       0x05
                   })
               }

               Method (_DSM, 4, NotSerialized)
               {
                   Store (Package (0x08)
                       {
                           "codec-id", 
                           Buffer (0x04)
                           {
                               0x88, 0x08, 0xEC, 0x10
                           }, 

                           "layout-id", 
                           Buffer (0x04)
                           {
                               0x78, 0x03, 0x00, 0x00
                           }, 

                           "device-type", 
                           Buffer (0x0F)
                           {
                               "Realtek ALC888"
                           }, 

                           "PinConfigurations", 
                           Buffer (0x28)
                           {
                               /* 0000 */    0x50, 0x70, 0x21, 0x01, 0x40, 0x01, 0x13, 0x90, 
                               /* 0008 */    0x20, 0x30, 0x81, 0x01, 0x41, 0x40, 0x01, 0x01, 
                               /* 0010 */    0x30, 0xE0, 0xC5, 0x01, 0x60, 0xE0, 0x45, 0x01, 
                               /* 0018 */    0x0F, 0x00, 0x00, 0x40, 0x0F, 0x00, 0x00, 0x40, 
                               /* 0020 */    0x10, 0x01, 0xA0, 0x90, 0x11, 0x01, 0xA0, 0x90
                           }

I used DSDTSE to edit the tony's DSDT (which is very close to our need) and changed HDEF and also fixed the _WAK return warning (see DSDT Hacks in DSDTSE tool)

 

Hope this helps.

 

 

@Rekursor

 

I am still trying to get my sound to work properly on an P55 UD4P. I came across this post and was wondering if I need to change my Pinconfigurations section in the DSDT to get them to work. If I use the IOreg utility and look at the HDEF section the Pinconfiguration values are different compared to what I have in my DSDT. Is that normal or do I need to sync them up?

 

Also, how do I determine what value goes where when I am reading this IOREG information. Thx

[Pieroman]

Hi everyone !

 

I just saw that Chameleon 2.0 RC4 is out, in the release notes they say something about i5 detection, but not much more. Have you tried it ? Does it make life easier with P55 boards ?

[profdre]

Does this happen only after sleep?

 

Yes I'm not absolutely sure, but at least in 2 cases. Before that I never had a kernel panic in SL. Now I try to shut down and reboot and I'll see if I won't have kernel panics. The problem for me is that the kernel panics are not really traceable, the hackintosh does not crash right after terminating sleep, it can last some hours... At the moment it's just running fine... (without using sleep)

If anybody can check my DSDT if it's appropriate for my hardware, that would be great!

 

Thanks!

 

 

Or... it could be because it is when idle or not working hard that the CPU enters C states, when CPUs are at full load not entering C states at all.

 

I have a hard time believing that just because some PMG5 PSUs were defective that that is what a lot of us are experiencing now... on a wide array of PSUs. Probably better to look at the components that we have that are identical.

 

I just thought so, because I have none of these sounds at my machine...Everything is silent, okay, the front panel jack doesn't work, but I hear no hissing sounds through the backside ports and no strange sounds coming out of the computer...

DSDT.aml.zip

[Alex HQuest]

A known issue- I've had the same problem- you have to set up a script in Windows that changes the clock to time.gov or something at login. I did it years ago, but I don't remember exactly how, and haven't fixed it recently. I'm pretty much 99.9% in OS X these days

 

When ntpdate command runs on OSX boot process, it stores time settings in UTC but system displays local time. Windows do uses local time, hence the difference.

[tonymacx86]

Hi everyone !

 

I just saw that Chameleon 2.0 RC4 is out, in the release notes they say something about i5 detection, but not much more. Have you tried it ? Does it make life easier with P55 boards ?

 

Just tried Chameleon RC4- here are my initial findings:

 

1. Has the Nehalem patch as mentioned in my modified PC-EFI 10.5 post. This enables P55 support without the 'system too fast' issue.

 

2. Breaks GraphicsEnabler-Yes for my GeForce 9400 GT. This works fine with the PC EFI 10.5. I'm pretty sure that PC-EFI recognizes cards better- don't know why yet.

 

3. Still no love for the GA-P55M-UD2 with EthernetBuiltIn=Yes. I still have to use RealtekR1000SL.kext in S/L/E.

 

4. Works without PlatformUUID.kext- no errors with Time Machine or SuperDuper!. I still have to use this kext with PC-EFI 10.5.

 

5. In About this Mac, CPU still Quad-Core Intel Xeon- but with the RC4 the i5 shows up as 2.67 instead of 2.7Ghz.

 

My guess is that netkas will update PC-EFI to 10.6 in the next few days incorporating the changes. What are everyone else's findings?

 

http://tonymacx86.blogspot.com/2009/12/cha...-supported.html