Apple Event: May 7th at 7 am PT

Looks like no one’s replied in a while. To start the conversation again, simply ask a new question.

MBA2012 cpu multiplier wrong and no turbo boost in Windows7

MBA2012 Corei 7 2.0GHz/ MEM 8GB / SSD 128G.


core i7 3667U, it support turbo boost, max multiplier 3.2GHz (x32, 2core), or 3.0GHz(x30, 1core).

but check cpu-z 1.60.1.x64, it reports cpu multiplier x8 or x19.

core speed does not rearch 2.0GHz(x20) and can not turbo boost.


User uploaded file

Posted on Jun 17, 2012 7:57 AM

Reply
47 replies

Jun 27, 2012 8:07 AM in response to tori10love

I have noticed the same and reported a bug to Apple. Oddly, this doesn't seem to be affecting the i5 MacBook Airs, as they register higher CPU scores on the Windows Experience Index (6.9) than the i7 (6.3).


I'd suggest sending a bug report, as well. The more this gets reported, the more likely they are to patch the firmware.


http://www.apple.com/feedback/macosx.html

Jun 28, 2012 3:16 PM in response to tori10love

I can confirm this as well on my 11" i7 8GB/256GB. I get the same Windows score of 6.3 whereas my 13" i5 scored 6.9. I hope it dones't take to long to address as it's a fairly simple toggle in the EFI, they probably have to run some regressions first though to make sure nothing else breaks I'm sure. There are definitely applications I use that use the turboboost and one of the reasons I went with the i7 option.

Jun 29, 2012 5:18 PM in response to goinskiing

Me as well. The i7 of the last generation provided a nice boost.


Is there anything I can do to file a report? I don't really like dealing with the 800 apple line as they are not the most knowedgable. I did submit feedback somewhere.


To the guy who said Apple support confirmed the bug, which apple support did you call and is there a case # I can reference?


Thank you

Jul 1, 2012 7:12 AM in response to oxford_guy

Indeed it does suck. I am a developer and Parallels does not cut it for my large Visual Studio project. You can definitely tell. My last generation Mac Air is noticeably faster in Boot Camp.


However, the new Air is noticeably faster in OS X and I assume a fix with be forthcoming for bootcamp.


I am contemplating returning the machine until they issue a fix.

Jul 13, 2012 12:57 AM in response to tori10love

It is assumed in the model CTO in the initialisation of the MSR in EFI, is to set the Target P-State and 19 at the time of turbo boost, and the root cause of the bug (I originally set 32 ​​is correct).


As a temporary workaround, in the application of the Throttle Stop, but it was confirmed that Turobo Boost working effectively, then, I tried to examine the state of the MSR at the before and after application of the Throttle Stop.


As a result, the following was found.

• The CTO model EIST, Turbo Boost is enabled.
• However, in the CTO model, Target P-State is closed when the Turbo Boost is set to 19-fold. Therefore, (Current P-State) can not be set to at least 19-fold magnification of the CPU.
• When you execute the Throttel Stop, so that can be set to 32 times the Target P-State of Turbo Boost, will be able to raise up to 32 times also (Current P-State) magnification of the CPU.


In fact, before and after execution of the Throttle Stop, appears similar to the following and see the state of the MSR. (The value of the MSR, Save Report(.TXT) of the "About" → "Tools" of CPU-Z in the, you can see)


CPU Throttle Stop before) cpu multiplier 19 (max cpu multiplier 19)
MSR 0x000001A0 0x00000000 0x00850089
MSR 0x00000198 0x00001CD2 0x00001300
MSR 0x00000199 0x00000000 0x00001300


⇒ enable EIST, enable TurboBoost, the current P-State 19, the target P-State 19


CPU Throttle Stop after) cpu multiplier 30 (max cpu multiplier 32)
MSR 0x000001A0 0x00000000 0x00850089
MSR 0x00000198 0x000021A0 0x00001E00
MSR 0x00000199 0x00000000 0x00002000


⇒ enable EIST, enable TurboBoost, the current P-State 30 , the target P-State 32


From the above, CTO model, at the time of initialization when the switch to Windows, it would have been set to 0x1E (0x00001E00: DEC

19) is target Performance State Value is IA32_PERF_CTL of MSR in the BootCamp is a cause of trouble, 0x20 (0x00001E00:DEC 32) you can determine it because it does not, and has been a state that does not increase the magnification was also an attempt to invoke the turbo boost when under load.


Core i5 In the standard model, so go up 26 multiplier, probably should be the value of the following.

MSR 0x00000199 0x00000000 0x00001A00


<Reference>
· Intel64 and IA-32 Architectures Software Developer's Manual Combined Volumes1, 2A, 2B, 2C, 3A, 3B and 3C
http://download.intel.com/products/processor/manual/325462.pdf

· CHAPTER 14 POWER AND THERMAL MANAGEMENT
14.1 ENHANCED INTEL SPEEDSTEPR TECHNOLOGY
14.3.2.2 OS Control of Opportunistic Processor Performance Operation
· CHAPTER 34 MODEL-SPEC IFIC REGISTERS (MSRS)
34.7 MSRS IN INTELR PROCESSOR FAMILY (INTELR MICROARCHITECTURE CODE NAME SANDY BRIDGE)


1A0H IA32_MISC_ENABLE
16 Enhanced Intel SpeedStep Technology Enable. (R/W)
38 Turbo Mode Disable. (R/W)
198H IA32_PERF_STATUS
15:0 Current performance State Value.
199H IA32_PERF_CTL
15:0 Target performance State Value.
32 IDA Engage. (R/W).

MBA2012 cpu multiplier wrong and no turbo boost in Windows7

Welcome to Apple Support Community
A forum where Apple customers help each other with their products. Get started with your Apple ID.