Jump to content
HWBOT Community Forums
Leeghoofd

NEW XTU 64BIT VERSION LOCK

Recommended Posts

Seems Intel has a new 64 bit release out of the blue that seriously improves the output, However these new scores will not be compatible/comparable with exisiting ones.

My preference would be to go for a seperate ranking with XTU 64bit versions only.

 

Keep you posted what gives with the developpers

Share this post


Link to post
Share on other sites

This Version has been out since 27 -3-19.It"s not like it was release this week.If it give more point we should just rebench it with our hardware.

 

 

Share this post


Link to post
Share on other sites
Posted (edited)

XTU always had a 32 and 64 bit inner benchmark executable that was correctly chosen by the bitness of the OS. So that has nothing to do with the improved performance.

The reason for increased scores is that this new version now uses Prime95 29.4 while the old uses the slower 27.7.

Regarding benchmark security and timer reliability Intel opened up another can of whoop ass by using QueryPerformanceCounter() for timing. As I have not seen any counter measures that take care of timer skewing, changing your bclock in Win 8/10 should change the score. The old version had a "fix" for that behaviour by calculating the TSC frequency on a regular basis, so it successfully adapted to bclock changes as well.

Also, NOT A SINGLE ONE of my flaws was addressed with the new version: https://www.overclockers.at/articles/intels-xtu-analyzed

Edited by _mat_
  • Like 5
  • Haha 1

Share this post


Link to post
Share on other sites

At the very least remove points and make this benchmark truly pointless.

  • Like 3

Share this post


Link to post
Share on other sites

Hence why i said it was a new 64 bit one, not a new XTU version now in 64bit version… Frederik is looking into a  seperate ranking for this one , however seems not an easy task… maybe time to implement the mandatory OS field... however that's not bulletproof either… another moderation nitemare

Share this post


Link to post
Share on other sites
Posted (edited)

i checked the most popular bench on bot.(last one month submissions)

xtu submissions are staggering -5518 submissions (april 2019). 184 submissions per day.

(just to cross check, march 2019 xtu submissions are close to 6200.)

comparitively other benchs nos. as follows (per month)

1. super pi 32m -339

2. cinebench r15 -452

3. x265 1080p - 217

4. wprime 32m - 274

5. geekbench 3 multicore -217

few 3d benchs

1. firestrike - 211

2. 3dmark 11 performance - 155

3. timespy - 263

4. 3dmark vantage (p) - 112

5. aquamark - 108

xtu is more popular with masses especially rookies, novices. it's much easier to run. so points are required for this bench. now i have seen 4-4.5% (some cases almost 10 percent) gain with this version which is not acceptable to rebench old hardware with this version. 

xtu older versions can't be updated to this version. intel recommends uninstalling old version and installing this new version & further updates will be pushed automatically.

this is a new benchmark & needs a separate category.

anyway globals are removed from this bench. what we have is only hw points. keeping hw points will make rookies, novices happy.

personally i don't like this bench much and neutral in my opinion to keep hw points for this bench. 

Edited by ksateaaa23
  • Thanks 1

Share this post


Link to post
Share on other sites

To be fair, when the new OpenCL came out for GPU-Pi, we didn't remove all old scores back then or create a seperate ranking, we just accepted that it was part of the benchmark and got on with it.

I also don't know how you'll go about moderating the XTU scores if you have two seperate rankings considering it's the most popular benchmark by far. Unless the submission system can detect which version has been used and therefor always put you in the right catagory by default, I see XTU being completely unmanigable thereafter.

Share this post


Link to post
Share on other sites

Hence why I want to learn from the errors made in the past... it is stupid to have all the hard work and efforts people put in, to get just erased. 

Developpers just change  (part of) the benchmark making it produce way higher than before scores . Ofcourse they don't care about the DBase we all have build up over the years.

It is just plain stupid to see people score higher with far lower clocks due to the fact they got hold of an old CPU and rebench it on the latest software...

 

And yes it will be a moderation nitemare (again)

 

 

  • Like 1

Share this post


Link to post
Share on other sites
18 hours ago, yosarianilives said:

Better idea, how about just no xtu?

Gets my vote.

Share this post


Link to post
Share on other sites

remove points and give a currently "unpointed" bench a shot with points

 

$0.02

  • Like 1

Share this post


Link to post
Share on other sites

Don't make any changes.

If novice & rookie seemingly do better than the old scores it's a win for them " to keep trying overclocking "

Share this post


Link to post
Share on other sites
On 5/3/2019 at 6:29 PM, ObscureParadox said:

Unless the submission system can detect which version has been used 

I'm pretty sure this is the exact mechanism used to exclude results from old versions on other benches with datafile required, e.g. GPUpi and x265

Share this post


Link to post
Share on other sites

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...