BEWARE ! of validation

[kick]

Watch out for cpu-z validation rubbish
Ive seen some sites where people dont accept opinions/advice etc unless they see the cpu-z validation link.
The link itself is rubbish tho ,mho.

I think,if i worked at it a little I could get 3.7Ghz on air cooling for a validation for my cpu,but it would be like playig a con game.
Mainly beacause the test takes only a few minutes.

Im including a screenshot ,that has gimped memory timings and using very hi voltages .
As u can see,one worker fails after 14 min but for "validation" that wouldnt show!

[SolidBladez]

Re: BEWARE ! of validation

The reason for a validation link is so that you can show your highest overclock (or current for that matter). 80% of the time they're just for show and not 24/7 stable.. maybe bench stable but not stable enough to run LinX or Prime for a sustained time frame.

For example, I'm sure a lot of Lsdmeasap's overclocks are barely enough to run a SuperPI. When you're running 1.5V+, it can get a little hot and unless you have an endless supply of LN2, DICE, or a cascade I doubt you could run at those speeds 24/7.

Also, the validation link prevents the user from editing and falsifying his own clock.

[Lsdmeasap]

Re: BEWARE ! of validation

Yes, alot of validations are made solely to show a "MAX" Something, they are not always "Stable" as one might say they are. Some people, like myself, often push for a MAX of something only, and then later if time permits test things a bit lower for stability. I was doing Dice runs this past week so I was limited by time for sure so I did not have any time for stability testing as my ice was disappearing and I was just learning the ropes

As said above alot of people, including myself, often post a "MAX" something meant to be just that, a Max they reached for bragging rights or just to show what they pushed there hardware to and they do not at all imply any stability.

Now if someone is claiming something is stable per some software but only showing a validation then you need to ask where the stability screenshots are at!

Validation would show those memory timings though, in case you were unaware of that

As linked above, my max Blck of 237 was for sure not something stable, nor did I even test it as I knew better

232 Blck I did a small test a few times, SuperPI 1MB but of course that can't be considered stable by any means either, only 1MB SuperPI Stable and that is all I will claim as that is all I tested at the time (Just in case you was referring to my validation posts) I have many Validations at through 225-238/4.7-5Ghz I collected while trying to push max Blck, and max CPU Freq.

Here is that screenshot of said stability

Click to Enlarge


Validations I made at that speed, with different timings though as I was changing many things very quickly to try to find some things out.



CPU-Z Validator 3.1

[kick]

Re: BEWARE ! of validation

Well perhaps I should have said beware of using others validations as a guide to what your pc should run at.
Fine for showing a max youve done or what pc will run at for a few min as long as people dont get the idea that theyre what all should aim for:
As I said Ive seen sites where the "validation link" is expected for conservative advice and the validation itself is only a "snapshot".

[Psycho101]

Re: BEWARE ! of validation

I concur.

I can see how finding an absolute maximum could be fun. It's a great way of pushing your components to the max. Unfortunately I haven't had an exceptional socket 775 CPU to do such things with. Both the chips I've owned are so called "budget" models, so aren't exactly top binned silicon.

Definitely don't take validation links as a definite indication of what to expect from your own components, be they identical or similar. There are posts on every forum of people being disapointed that for example their E5400 won't do 4.2GHz, and they believe it should because someone else posted a validation. CPU's are as different as people. There are so many different variations and characteristics that every single one is different.

[Acebmxer]

Re: BEWARE ! of validation

What I hate about Cpu-z validation is that it blanks out the core voltage. It would at least give you a better idea at what they had to through at it to get it to that max speed.

[gampamu]

Re: BEWARE ! of validation

I concur.

I can see how finding an absolute maximum could be fun. It's a great way of pushing your components to the max. Unfortunately I haven't had an exceptional socket 775 CPU to do such things with. Both the chips I've owned are so called "budget" models, so aren't exactly top binned silicon.

Definitely don't take validation links as a definite indication of what to expect from your own components, be they identical or similar. There are posts on every forum of people being disapointed that for example their E5400 won't do 4.2GHz, and they believe it should because someone else posted a validation. CPU's are as different as people. There are so many different variations and characteristics that every single one is different.

Yes, reaching the max CPU-Z validation is the stuff of bragging rights and not much more. But it can be a lot of fun. Can anyone with a Q6600 beat my 4275 score?

CPU-Z Validator 3.1

[Lsdmeasap]

Re: BEWARE ! of validation

What I hate about Cpu-z validation is that it blanks out the core voltage. It would at least give you a better idea at what they had to through at it to get it to that max speed.

I agree, I HATE that as well. I think Franck@x86 has said why this is but sorry I could not find the thread. I do believe if I remember correctly it is so that users do not look at/through uploaded validations and try the higher voltages shown on some validations and end up killing their CPU's.

Yes, reaching the max CPU-Z validation is the stuff of bragging rights and not much more. But it can be a lot of fun. Can anyone with a Q6600 beat my 4275 score?

CPU-Z Validator 3.1

Yeap, A member here Fatguy1992

545 FSB
Q6600 @ 535FSB 4 Cores / 545FSB 2 Cores - XtremeSystems Forums

4.545 (Phase) + 4.1 (Air)
Golden Q6600 - XtremeSystems Forums


Sorry, you asked and I just had looked at one of his posts at XS so I had to reply for him since he may not read this post to brag about his setup

Don't feel bad, I can't win for trying either But you are right, it is fun and nice to be able to brag about a high score on anything, even if you are not the winner. It is also nice to see others validations to give yourself something to work towards, or sometimes to see that you have beat someones score

[gampamu]

Re: BEWARE ! of validation

Yeah, fatguy1992 has a very nice chip. It was actually his results that inspired me to push for the high clock.

[creamsoda]

Re: BEWARE ! of validation

I wouldn't brag about an overclock unless it was stable in Prime95 small, large & blend.

[gampamu]

Re: BEWARE ! of validation

It's a different kind of overclocking and it can be a lot of fun.

[creamsoda]

Re: BEWARE ! of validation

On youtube, I've seen ppl get 6-7ghz overclocks using liquid nitro & dry ice.

I've always thought it was kinda pointless.

[Acebmxer]

Re: BEWARE ! of validation

I agree, I HATE that as well. I think Franck@x86 has said why this is but sorry I could not find the thread. I do believe if I remember correctly it is so that users do not look at/through uploaded validations and try the higher voltages shown on some validations and end up killing their CPU's.

well I guess with the new Cpu-z they fixed the missing Vcore in the validation.

CPU-Z Validator 3.1

[demonsavatar]

Re: BEWARE ! of validation

What I hate about Cpu-z validation is that it blanks out the core voltage. It would at least give you a better idea at what they had to through at it to get it to that max speed.

Yeah I'm not sure why they do that! Very annoying!

well I guess with the new Cpu-z they fixed the missing Vcore in the validation.

CPU-Z Validator 3.1

Good to hear :)