Jump to content
HWBOT Community Forums

Recommended Posts

Posted (edited)

I just discovered a strange issue when raising FSB using SetFSB in Win XP, when I add more than 5MHz of FSB the multiplier starts to drop.

This only seems to appear in XP, I used the same version of SetFSB in Win 7 x64 and it worked just fine (both CPU-Z 1.85)...

Setup used was a ASUS P5E64-WS EVO and a Pentium E5800 but I also had the same issue on a EP45-UD3R but with 10MHz above boot-FSB...

EDIT: Seems like a CPU-Z bug, CPU-Z 1.63 displays the right multi, but I can't use that to validate for obvious reasons :(

EDIT2: The CPU itself isn't lowering clocks at all, spi runs are nearly the same within the 1MHz where the bug accurs...

Anyone ever had a simmilar problem and might know a fix?

Any help would be greatly appreciated :)

Edited by TAGG
  • Thanks 1
Posted

Huh, nice testing. I noticed it on a Celeron 430 IIRC, but it was while I was binning so didn't test any benches. I found booting higher helped prevent it from occuring as soon; not a solution but the best I could find at the time.

Posted

Yep booting higher helps but only linear i.e. booting 400MHz instead of 395 lifts the bug to 405 instead of 400...

I will try a brandnew XP install tomorrow...

At least it's  only a display bug, so benches are not affected, you just need to use antique CPU-Z to have the right Frequency displayed :)

 

Posted
44 minutes ago, TAGG said:

Yep booting higher helps but only linear i.e. booting 400MHz instead of 395 lifts the bug to 405 instead of 400...

I will try a brandnew XP install tomorrow...

At least it's  only a display bug, so benches are not affected, you just need to use antique CPU-Z to have the right Frequency displayed :)

 

What happens if you open 1 window of old correctly working cpuz (leave open), then open the new required version. Does it show correctly? If yes then you can open rest of cpuz windows you need, then close the old version and take your screen shot. This is a trick that fixed weird issue with AFD on mocf for kabylake. No idea if it will work for you but worth a shot.

Posted

Doesn't work, neither does the win98 version or anycombination of settings in the ini. file.

As soon as you hit the threshold of 5MHz new CPU-Z displays one multi lower no matter if you have it open during changing FSB or if you open it afterwards if you had old CPU-Z open...

Posted
21 hours ago, TAGG said:

Yep booting higher helps but only linear i.e. booting 400MHz instead of 395 lifts the bug to 405 instead of 400...

I will try a brandnew XP install tomorrow...

At least it's  only a display bug, so benches are not affected, you just need to use antique CPU-Z to have the right Frequency displayed :)

 

Technically don't you need to show the freq that you benched at?

  • 6 months later...
Posted
9 hours ago, unityofsaints said:

I am seeing the same on the Rampage III Extreme with locked CPUs.

On 1366 you can get around it by upping the multi using Throttlestop. I've seen this issue on 775 (P5Q Deluxe and EP45-DS3) as well as on various 1366 boards (at least Rampage II Extreme, x58a-oc and x58a-UD7) and also on 1156 (H55m-UD2h and P55-USB3)

Posted

You can bug it in the other direction as well if that helps as motivation to fix it, lowering bclk raises multi to impossible levels (see attached pic)

IMG_20190203_155733.jpg

  • Haha 1
  • 1 year later...
  • 5 months later...
Posted

Hi Guys :)

A few Days ago i got the same Issue with an E5800 on an Asus P5E3 Deluxe/WIFI and Windows XP SP3.
Like 'TAGG' said, even if you boot with higher FSB, the Problem is still there, it just moves.

I tested all CPU-Z Versions down to v1.58. The Issue starts with CPU-Z v1.78.
With CPU-Z v1.77 i got some fast "hopping" of Numbers when i start to rise FSB.
But a few Seconds later it stabilizes. I have about 2 of 5 corrupted CPU-Z Validation Files with v1.77.
With v1.69 no Issues are found.

The Point is, like 'Luumi' said, that it's only a Visual Bug.
But if i try to get some Validation File for high FSB or CPU-Freq. this Bug doesn't let you get the right Speeds.
The Validation-File saves the "false-Multi", so if you Upload this File, you will have a droped Multi on you CPU-Z Validation Page.
And using older CPU-Z is senseless, because you can't Upload these and on the other Hand you have to use 1.94+ for HWBot-Submissions:

So...
Has someone of you Guys found a Solution or a Workaround for this strange Behavoir?

Thank you in Advance and best Greetings,

mrmouse
 

 

20200101-005252.png

20200101-005303.png

Posted

Thanks a lot for the FAQ :)
Haven't a look at the very Bottom.

Just to be on the safe side; if i use old enough Hardware, there is no need to get a CPU-Z Validation on there Site (https://valid.x86.fr/)?
So it would be enough, when i do a normal Screenshot instead of Validation-Link for CPU, Mem and FSB-Frequency?


For me, the Validation-Page for CPU-Z doesn't take/upload any Files under v1.94; so i even can't upload them.

You said "Unchecked" on the Validation-Website is not allowed for Submissions like CPU, Mem and FSB-Frequency?
A lot of the Top-Overclockers have unchecked Submissions; you get them if you set CPU-Z to XOC=1.


Was this a Time ago allowed or how is this Checked/Unchecked behavior in 2021?

Best Greetings,
mrmouse

Posted

Yes, i tried the Vintage-CPUz-Edition, without success.
Also i tried to move the CPU-Z.exe into different Folders with different .ini-Files.
Tried a lot, but nothing seemed to get the Success.

The only thing, to get rid of these Error on CPU-Z 1.78 to 1.95, is to set the Multiplier to 15,0x.

Mulitplier 16,0 and 15,5 -> strange Behavior.
Multiplier 15,0 and lower -> no Problems

But this isn't for me a "real Solution"...

Posted

I'm thinking is an issue with newer versions of CPU-Z vs SetFSB itself, like it is with older Socket A systems locking up when you open CPU-Z with any CPU-Z version 1.60 and later. Although it doesn't happen with all Socket A's it occurs on alot of them, I've ran into that too many times myself.

Also; Do bear in mind SetFSB is an older tweaking util and I seriously doubt it's been updated for a VERY long time now.
If the guys working with CPU-Z can solve all that it would be great.

  • 3 years later...
Posted (edited)

Hey guys same thing happened today, E5800 and Rampage Formula on SS, when using setfsb to go over post +10mhz it shows lower multiplier higher i get, how come this wasnt sorted in 2024 ?

Any advice on this use case ?

 

dejan

 

 

EDIT:

Question: Can I use an older CPUZ version on legacy hardware?

Answer: Yes, HWBOT recognises that on LGA1156 and older hardware newer CPU versions can lead to wrong multiplier detection and even BSODs while just opening CPUZ. Therefore scores are allowed on these Legacy platforms with CPUZ version1.59 or CPUZ version 1.77

 

I found the answer in FAQ lol, will try tommorow thank you

Edited by dejan_bin_laden
possible answer

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...