Jump to content
HWBOT Community Forums

Recommended Posts

Guest george.kokovinis
Posted

Pieter,

 

There is an update.

I had some free time in the weekend, and decided to give this "Lasso app", a second chance.

Well, from my humble side, I can say that Z170 is not affected.

I also tried alternate, more aggressive intervention, than what Splave

suggested.

P95-bench.exe and perftune.exe were added to real time.

 

Nope. Nothing.

Of course Splave DID NOT mention Z170 on his initial post, so I would

expect his feedback on that.

Splave is one of my most respected overclockers here, and when he writes

he knows what he is saying.

 

I have to note though that XTU, as everyone knows, takes its measurement in the first 1 and 1/2 seconds of its run.

The rest is a stress test based on P95 algorithms.

My question is, since the measurement is over BEFORE P95 kicks in,

what is the profit of amending P95 to real time with Lasso ?

 

From what I noticed, with Lasso, when P95 kicks in, the blue line ( cpu usage ), remains linear without any ups and downs and cpu usage remains close to 90-95%.

 

Final result is the same though.

 

Maybe older versions of XTU are affected ?

Maybe older versions of Windows ?

Expecting Splave's respected insight.

 

George:)

Posted

I think perftune to low priority will help you George.

 

Can you track if PL indeed amends the priority of p95-bench.exe - it should state it in the log. I don't bench modern platforms anymore but I think it is either you doing something wrong or the OS build which is not that much affected with this bug.

Guest george.kokovinis
Posted
I think perftune to low priority will help you George.

 

Can you track if PL indeed amends the priority of p95-bench.exe - it should state it in the log. I don't bench modern platforms anymore but I think it is either you doing something wrong or the OS build which is not that much affected with this bug.

 

 

Thank you for the advice.:)

Indeed PL amends the priority of p95-bench.exe.

 

Nothing more from my side. No tweak expert here.

Let the true experts chime in.

Posted
Better than not knowing what scores are legit and what scores are cheated.

 

I don't know about that, I mean I spent months and tons of money on my 6700k xtu score and to have it erased because there is a tweak/bug that I didn't even know about would really suck. It would be nice if they could figure out a way to determine if it was used or not instead of deleting all points. I only have 4 xtu scores towards my global profile so it wouldn't hurt me that much, it just feels like the principle of the matter.

Guest george.kokovinis
Posted (edited)

As usual in life, same thing tends to happen here.

 

I explain myself.

There is a minority of cheaters ( I am not talking only about XTU ), and

a vast majority of members that play by the rules, investing time, money

and all available resources to achieve something good.

 

Who will ever forget how the winner of a popular competition, recently

got away with his cheating of ambient/sub-zero cooling methods, playing

on the edge of the rules of HWBOT, and against any logic of "ambient

interpretation".

 

Splave revealed to our eyes, a trick unknown to most. PL.

 

As the big picture suggests, XTU is a problematic benchmark.

At least vulnerable to heavy distortion of the final result, with "common"

or less common methods.

 

Instead of addressing the problem/s ( if possible ), we choose to address

the players of this benchmark, who are to my understanding 95% innocent.

 

OK.

I have spent two years and a ton of money, as Strong Island said, to reach almost the first place of the Enthusiast League ( missing it by 20 points ), WITHOUT ever using a single tweak.

 

Is it fair ? No.

Does it encourage me to move on to more effort ? No.

Does this whole situation makes the little bird in my head shout

" George, you are stupid " YES.

Edited by george.kokovinis
Posted (edited)

z170 It doesnt work for whatever reason all others are affected though, sorry could have saved you some time in trying z170 but was away.

 

I just think it will be hard to convince intel to fix this bug on older platforms

Edited by Splave
Posted
I don't know about that, I mean I spent months and tons of money on my 6700k xtu score and to have it erased because there is a tweak/bug that I didn't even know about would really suck. It would be nice if they could figure out a way to determine if it was used or not instead of deleting all points. I only have 4 xtu scores towards my global profile so it wouldn't hurt me that much, it just feels like the principle of the matter.

 

Unfortunately for many the same happened to UCBench and PCMark 05, just the nature of the game unfortunately, when a bench breaking bug like this is found with no fix (not saying that is the case here) then you have no choice but to simply remove the points as it becomes impossible to moderate, especially if people can start being clever about breaking it to make it look legit.

Posted
I don't know about that, I mean I spent months and tons of money on my 6700k xtu score and to have it erased because there is a tweak/bug that I didn't even know about would really suck. It would be nice if they could figure out a way to determine if it was used or not instead of deleting all points. I only have 4 xtu scores towards my global profile so it wouldn't hurt me that much, it just feels like the principle of the matter.

 

Welcome to overclocking please enjoy your stay.

Guest george.kokovinis
Posted
Unfortunately for many the same happened to UCBench and PCMark 05, just the nature of the game unfortunately, when a bench breaking bug like this is found with no fix (not saying that is the case here) then you have no choice but to simply remove the points as it becomes impossible to moderate, especially if people can start being clever about breaking it to make it look legit.

 

 

With all respect, I do not agree.

 

What is a benchmark ?

It is an application.

Written by a human or humans. A series of 0 and 1.

Anything, even CIA encrypted 1024bit systems are vulnerable.

 

It is only a matter of knowledge and time to break any code.

 

Since moderation is done from a distance, no one can REALLY tell

what the user is doing at his home.

 

There is not even a single one benchmark application that can't be

modified in a way to look legit.

Simply, there is no financial interest to spend time to do that.

 

I am not so sure though if the same thing applies to big vendors who

have every reason to make their hardware LOOK better than the competition.

 

It is called insider info.

So, XTU or whatever XTU is vulnerable.

XTU today, another one in six months and finally we are out of benchmarks.

 

There are solutions to hold things tight and nice, but it is not the place nor the time to open such a discussion.

Guest Bullant
Posted
I don't know about that, I mean I spent months and tons of money on my 6700k xtu score and to have it erased because there is a tweak/bug that I didn't even know about would really suck. It would be nice if they could figure out a way to determine if it was used or not instead of deleting all points. I only have 4 xtu scores towards my global profile so it wouldn't hurt me that much, it just feels like the principle of the matter.

 

Regardless of the out come mate I know what it's like to work hard on something,I for one appreciate the effort you put in as I can see you've worked hard on it.Hopefully something can be resolved....

 

From what I hear skylake is fine? It's just the older platforms ?

  • 1 month later...
  • 1 month later...
  • Crew
Posted

this happens because XTU isn't even a real benchmark, i been saying this for years

 

i appreciate there is a component of XTU that is a benchmark, but I get the impression the measured CPU MHz is included in the benchmark calculation

 

we have seen repeatedly over the life of XTU, when the GHz measurement is off, the score is skewed and this issue keeps coming up

 

we were playing with this during country cup and were able to almost skew any score we wanted by fiddling with that realtime, but there is always the telling factor of the GHz measurement being wrong

 

in my long standing opinion, XTU should be removed from pointed/competition benchmarks until Intel revise the mathematics behind it. it may be to give some sort of "efficiency" calculation, which im all for, but it shouldn't impact on the score at all

Posted
Can you PMe the exact instructions so we can replicate and patch?

 

Thanks :celebration:

 

Any updates?

 

Just like @george\.kokovinis I'm having troubles reproducing this bug. With a detailed procedure we can report to Intel and get it fixed.

 

:)

  • Crew
Posted
But it gives me many fancy points! also, intel & hwbot dont give no bunnys. And as such, that was a waste of your keystrokes.

 

im not attacking, im just legitimately saying im concerned with the way it calculates points

Posted
:)

 

From the OP:

 

And here is a step by step way to do it.

 

1. Download process lasso

2. Open

3. Disable Pro balance garbage File->Pro Balance

4. Next you are going to go to default process priorities

5. Type in p95-bench.exe and select realtime from the drop down and add to list

6. Loop XTU till you get a WR.

 

You'll notice that the rotation above the XTU process bar will be glitchy fast then slow etc that's how you know its working. Also the Mhz will be super high but you can make them look more normal by under clocking or even using slow mode to get a nice round ln2 able figure.

 

Better yet you dont need a screen shot so who cares if it says 10 GHZ, your validation file will work just fine.

 

Is this not working for you?

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...