42 replies to this topic

[mitch_de]

As others, check geekbench points to be sure you dont get stuck with slowest Pstate (speedstep stopped working) !

[^Andy^]

mitch_de, on Nov 20 2011, 08:25 AM, said:

As others, check geekbench points to be sure you dont get stuck with slowest Pstate (speedstep stopped working) !

Just for info - Speedstep is still working on P8P67 Pro using the modified bios.

[Majonéz]

Maybe another bug in 11D16... When I want to start/stop the samba sharing in the system preferencies, the system stop responding... :S

[juanerson]

^Andy^, on Nov 21 2011, 02:49 AM, said:

Just for info - Speedstep is still working on P8P67 Pro using the modified bios.

Hi. Please tell us, How is possible? If NOT working in genuine Macs (I have 4 reports)... AND of course 4 Hackintosh more.

[Kynyo]

Don't FORGET this relase is 1st BETA!!! It's normally to have a lot of problems...

[fantomas1]

Kynyo, on Nov 21 2011, 07:21 PM, said:

Don't FORGET this relase is 1st BETA!!! It's normally to have a lot of problems...

not necessarily !    

[^Andy^]

juanerson, on Nov 21 2011, 05:55 PM, said:

Hi. Please tell us, How is possible? If NOT working in genuine Macs (I have 4 reports)... AND of course 4 Hackintosh more.

I don't know HOW it's possible - I'm just telling you what I am seeing. My Geekbench scores are still over 12,000 which to me suggests that speedstep is still working.

Without speedstep my Geekbench scores are usually around 8,000.

This is on a P8P67 Pro with a 2500k and 10.7.3 11D16

 Screen_Shot_2011_11_22_at_06.45.37.png   132.98K   69 downloads

[mitch_de]

If the bug has no effect it may depends from CPU type and/or smbios MacModel.

[^Andy^]

mitch_de, on Nov 22 2011, 06:50 AM, said:

If the bug has no effect it may depends from CPU type and/or smbios MacModel.

I will try changing the mac model in smbios tonight after work and see if that changes things (unless anybody with the problem would like to beat me to it lol).

[J_Freman]

Modifed BIOS for ASUS p8P67 Pro? Where??

[^Andy^]

J_Freman, on Nov 22 2011, 09:33 AM, said:

Modifed BIOS for ASUS p8P67 Pro? Where??

It's discussed from page 16 of this thread onwards http://www.insanelym...p...1312&st=300 but be warned if done wrong you could brick your mobo.

[kirasir]

^Andy^, on Nov 22 2011, 02:26 PM, said:

Geekbench 2.1.13  

[^Andy^]

kirasir, on Nov 22 2011, 02:44 PM, said:

Geekbench 2.1.13  

Why is that a problem? Without Speedstep using the same version my results fall to the 8000 range - the version of geekbench should have nothing to do with it.

[mitch_de]

Yep, geekbench version DOESNT matter to check if xpu stays on slowest PSTATE. There are only very minor results diff between gb versions. And the diff by non working speedstep is so huge (diff >= 30%)  that minor gb diffs between versions (much less than 5%)  doesnt matter.

[juanerson]

And what about MSR Dumper check? "Maybe" his CPU is stuck in the HIGHEST P-State.

[^Andy^]

mitch_de, on Nov 22 2011, 07:13 PM, said:

Yep, geekbench version DOESNT matter to check if xpu stays on slowest PSTATE. There are only very minor results diff between gb versions. And the diff by non working speedstep is so huge (diff >= 30%)  that minor gb diffs between versions (much less than 5%)  doesnt matter.

Just tried smbios using MacBookPro8,3 definition and still get 12,000 (12,264).

If I can find the post for msrdumper I will try that and see if it is stuck at highest speed..

[^Andy^]

juanerson, on Nov 22 2011, 07:27 PM, said:

And what about MSR Dumper check? "Maybe" his CPU is stuck in the HIGHEST P-State.

Nope I have the following in the kernel log..


Nov 22 20:06:42 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper CoreMulti(16)
Nov 22 20:06:42 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper PStatesReached: 16 48
Nov 22 20:06:43 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper CoreMulti(48)
Nov 22 20:06:43 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper PStatesReached: 16 48

So it is stepping up and down correctly (I've only ever had 2 steps, never looked into why).

[^Andy^]

^Andy^, on Nov 22 2011, 08:08 PM, said:

Nope I have the following in the kernel log..


Nov 22 20:06:42 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper CoreMulti(16)
Nov 22 20:06:42 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper PStatesReached: 16 48
Nov 22 20:06:43 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper CoreMulti(48)
Nov 22 20:06:43 Andy-Stubbss-Mac-Pro-3 kernel[0]: MSRDumper PStatesReached: 16 48

So it is stepping up and down correctly (I've only ever had 2 steps, never looked into why).

Just to clarify this is with a P8P67 Pro with an i5 2500K (clocked to 4.8).

[^Andy^]

^Andy^, on Nov 22 2011, 08:45 PM, said:

Just to clarify this is with a P8P67 Pro with an i5 2500K (clocked to 4.8).

Ok I have done some more messing around with this and there definitely is an issue the problem is I can't work out where lol.

I was messing around creating a new dsdt earlier today so thought I'd try it on my 10.7.3 installation - the result was a geekbench score of 4448 which is roughly half of what I get at stock speeds on an i5 2500K which implies its running at the lowest clock speed - i.e. a 16x multiplier. I then put the original dsdt back and nothing changed 10.7.3 still runs at the lowest multiplier. To test the dsdt wasn't hosed I tried it on 10.7.2 - and it was fine, runs with speedstepping working as you would expect. Now for the life of me all I changed between 10.7.3 with working speedstepping and a broken 10.7.3 was the dsdt and even putting the original back now wont fix it.

Something in fruitland aint right lol

[kirasir]

With this "Just to clarify this is with a P8P67 Pro with an i5 2500K (clocked to 4.8)" intel technology not working.