Not changing the version number won't help with validation. The bugged version can be used and it will be harder to spot

Trying CPU-Z + 990FX Sabertooth at the moment, 1.60.1/1.60 is the only one working atm, neither 1.61old, 1.61new or 1.61installer are working

Yeah: 8.7 GHz 1.2v validated on air with stock cooler:

 

Edited July 6, 201212 yr by Christian Ney

Seems legit!

So 1.60 is also borked?

Yeah,

 

I did 11 GHz but says file corrupted :cry:

2 GHz Memory Clock validated (DDR3-4xxx)

 

 

Any difference between x86 and x64 versions? I've been using the 32bit version for quite some time and haven't seen such strange behavior, but maybe I didn't made the "correct" steps.

used both, same, publish results soon hope it can help guys from CPUz to fix

1.60 sucks on my qx6850 as well, check my ucbench score:p Funny thing is it works at stock from what i remember...

Can any MHz submissions with these versions be trusted?

I'm wondering if it happens when using TurboEvo + opened cpuz at the same time only or it happens with other tools like AMD Overdrive + opened cpuz, where you can change multi/htref as well. Seen this only with asus 990FX boards until now. All rejected "records" were done with Asus boards (Sabertooth and Crosshair V Formula).

Haven't noticed that with my Giga 970A-UD3 where I'm using AMD Overdrive.

 

PS: Hmm, ok...Mr.Scott is using Giga.

Edited July 6, 201212 yr by I.nfraR.ed

Hi guys,

 

Yes Im using the latest CPU-Z 1.61 zipped package, the one that is supposed to be fixed.

 

I carried on testing last night and the results were repeatable although sometimes the clock speeds would go down aswell as up

 

I never tried with any other software than Turbo V Evo though.

Test Setup:

Microsoft Windows 7 x64

Asus Sabertooth 990FX (BIOS 1208)

Asus TurboV Evo and AMD OverDrive

AMD FX-8150

blablablabla.

 

CPU-Z in screenshots:

 

TurboV Evo ------- 1.61 (installer version) ------- 1.61new ------

TurboV Evo ------- 1.61(old) -------------------- 1.60.1 --------

 

So there is no tricky stuff to do in fact. just overclock like usual, open cpu-z and ajust bus speed and multiplier on the fly within Windows.

 

 

Changed 200 Mhz to 203 Mhz, only 1.60.1 see the change:

 

Same for 211 Mhz and so, only 1.60 sees it (cpuz 1.61new shows 117 MHz wtf):

 

Booting with 211 Mhz from BIOS (all read fine):

 

Change bus from 211 to 215, only 1.60 sees the change:

 

Ajust CPU Ratio from 19 to 14: CPU-Z goes crazy (all version):

Real bus speed: 216 Mhz

CPUZ: 150 MHz or 300 MHz

 

Boot with 150 Mhz x10:

CPUZ 1.61installer, 1.61old and 1.60 shows right frequency

1.61new shows 91.2 Mhz

 

Change CPU multiplier from 10x to 14x:

CPUZ goes crazy again:

CPUZ 1.61installer, 1.61old and 1.60 shows 211 MHz

1.61new shows 127.7 Mhz

127/91 = 1.4

210/150 = 1.4

 

From 14x to 20x:

CPUZ 1.61installer, 1.61old and 1.60 shows 300 MHz

1.61new shows 182.4 Mhz

300/150 = 2

182.4/91 = 2

 

Change from 20 to 27:

CPUZ 1.61installer, 1.61old and 1.60 shows 406 MHz

1.61new shows 246.3 Mhz

You got it no ? guess what is 406/150 and 246/91 ! 2.7

 

Now cpuz 1.61old gets even more retarded:

CPUZ 1.61installer, 1.61old and 1.60 shows right frequency

1.61new shows 91.2 Mhz

CPUZ 1.61installer, 1.61old and 1.60 shows right frequency

1.61new shows 91.2 Mhz

 

BUT, IF YOU ACTUALY OPEN CPU-Z AFTER THE CHANGES, I SHOWS THE RIGHT CLOCKS:

 

Same goes for memory clocks:

 

 

Some valids I made:

 

Stock valid (rejected wtf)

 

One of the 300 Mhz bugged (211 mhz real) (validated)

 

6 GHz on air with 1.2v (aka bugged) (4GHz real) (validated)

 

8.7 Ghz (4 GHz real) (validated)

 

2'076.9 Mhz Memory Clock (validated)

 

But still can't validate 11 GHz, all files are detected as corrumpted by valid.canardpc :cry:

 

 

EDIT: Just tried CPU-Z 1.59 and it is working perfect, none of the bugs I pointed here are present with CPU-Z 1.59

 

EDIT2: Installed windows XP and tried, 1.59 is working perfect, 1.60 is fucked, 1.61old is fucked and 1.61new is even more

Edited July 6, 201212 yr by Christian Ney

Yeah: 8.7 GHz 1.2v validated on air with stock cooler:

 

 

I was also getting some very bugged results with 1.60.1

 

 

My problem was that PSCheck was changing the multi properly, but cpuz thought that the fsb was increasing also (which it wasn't).

 

Hopefully the new one will fix this.

Yes it's the same effect I reported here.

 

Either AOD, TurboV Evo and PSCheck

 

 

EDIT: guys from CPUID sent me a cpuz 1.61.2 to test I will do it aas soon as I come back to the office (no hardware home)

Edited July 6, 201212 yr by Christian Ney

My qx6850 ucbench submission is also bugged. Works at stock, but not OCed. MP is fine, but FSB reading is way off:p

 

I hope Franck will go back to the old algorithm, that worked very well.

Which system knut ?

 

Btw anyone having problems with Intels ?

Which system knut ?

 

Btw anyone having problems with Intels ?

 

REX + QX6850 2k3, but same issue with Maximus Extreme.

REX + QX6850 2k3, but same issue with Maximus Extreme.

 

hum, talk to you on msn

Went to office quickly this morning to test the 1.61.2, it's even worse than others, it shows 5 MHz or 9.6 MHz bus frequency

 

Sent 16 cpuz report.txt files to CPUID

Edited July 7, 201212 yr by Christian Ney

Went to office quickly this morning to test the 1.61.2, it's even worse than others, it shows 5 MHz or 9.6 MHz bus frequency

 

Sent 16 cpuz report.txt files to CPUID

 

Just like on Win8.

Tested the 1.61.3, not getting better, the effect is now reversed and there are other side effects.

 

Tomorrow Franck D. (guy from CPUID/Z) will remote control my Bulldozer system and do some testing on it, hope we can fix this issue this week-end.

Hm...Im stay at 1.60x64....

fyi, to verify your clocks you can use CPUZ 1.59 which works just perfect.

 

1.60.x, 1.61.x are all bugged.

 

Ho FYI, you can't validate anymore online with CPUZ 1.60.x

Ho FYI, you can't validate anymore online with CPUZ 1.60.x

 

Wow, put out 3 bugged versions and block those that still work.