I do not envy your jobs

Somebody has to do it

 

And we are kind of crazy (understand stupid)

We work for potatoes from our master ^^

I deleted my bugged results, as a note.

 

Discovered that I have a 980x result that is actually legit, I think. Pretty stoked about that.

I've sorted out mine as well, except there are 2-3 submissions that have the legit score inside, but need edit from a moderator.

If you're going to make an exception for such kind of scores, I'd like to keep 'em. Otherwise, I can remove them too.

We will try to edit those exceptions when we see them. But out of 13k results, I am pretty sure we will overlook a few and block them instead. If you think it shouldn't have been blocked and could have been edited, please let us know.

 

@Pieter, when are you going to hire me full time?

Deleted Madshrimps bugged scores... all for the masters!!

 

However why is this one being blocked ? http://hwbot.org/submission/2499007_leeghoofd_ucbench_2011_core_i5_3230m_430.3_mpt_score

 

The passwords line up, so what's the issue now?

Edited September 7, 201410 yr by Leeghoofd

I've got some of my ucbench submissions blocked. Last one was modified and get value achieved with 4 threads (for 4 thread cpu).

Does that means that blocked results can be modified for right thread score and get back to the ranking?

For example this score was blocked but it has 4 thread score that should be legit.

What about dual core processors? Can they run just the default selected threads or would that run be considered a bugged run?

Let me see if i understand it right...

 

This is ok because pwds are in ascending order even considering that i used 56/60/64 threads: http://hwbot.org/submission/2323148_noms_ucbench_2011_fx_6300_977.7_mpt_score

And this is bugged because pwds is out of order http://hwbot.org/submission/2318196_noms_ucbench_2011_athlon_64_le_1640_(orleans)_83.7_mpt_score

 

Or all this bunnyextraction are bugged because 56/60/64 threads regardless of the pwds order?

Edited September 8, 201410 yr by NoMS

I've got some of my ucbench submissions blocked. Last one was modified and get value achieved with 4 threads (for 4 thread cpu).

Does that means that blocked results can be modified for right thread score and get back to the ranking?

For example this score was blocked but it has 4 thread score that should be legit.

 

Edited to the 4 threads score, thanks for reporting.

 

@yamaharacer: default threads from now on.

@NoMS: yes, any thread that is > than the default threads set is bugged anyway.

make it clear now my 3 core wr is banned, settings i used genieben give me and score was ok

i dont get it!

Where's UCbench?

Yeah whats going on my 4670k Run now 1552 insteeed of 1900, i dont get it

Finally been removed

Will be back, when cleaned.

Ctrl+A, delete all.

Panic mode on!)

Will be back, when cleaned.

 

Oh..that's cool

 

No word from the author or anything?

 

I would like for it to be reliable enough for globals, personally.

 

Hats off to whoever's doing the cleaning (probly Genieben)

I can never remember what constitutes a bugged run

Edited September 8, 201410 yr by Schmuckley

Edited to the 4 threads score, thanks for reporting.

 

@yamaharacer: default threads from now on.

@NoMS: yes, any thread that is > than the default threads set is bugged anyway.

 

Thanks!

Already warned my teammates to not run this benchmark anymore until the new version is out.

 

Will start to delete my bugged subs as soon as i get time to do this.

yeah Genieben and I will clean the mess.

As for bugged runs, as long as you use default thread count and/or lower or equal thread count than the cpu total its fine. Just dont set higher thread count than cpu total.

yeah Genieben and I will clean the mess.

As for bugged runs, as long as you use default thread count and/or lower or equal thread count than the cpu total its fine. Just dont set higher thread count than cpu total.

 

But am I right in saying that under certain circumstances (I believe with AMD 3 and 5 core CPUs) a stock run will still bring up a bugged result due to an odd number of threads used??

Two days ago we started cleaning all of our results (HWBOX Team) . I believe we didn't miss anything .

I started looking for bugged scores from many people and reported many scores (please don't take it personally) .

 

Thanks , and feel free to check and report if you find a missed/bugged score from my team .

yeah Genieben and I will clean the mess.

As for bugged runs, as long as you use default thread count and/or lower or equal thread count than the cpu total its fine. Just dont set higher thread count than cpu total.

 

*testing 5960X - i get bug result every run if t-count used is 8/16 ..

Can you please show us a screenshot. Thanks