GALAX GOC 2018 online qualifier thread

[Leeghoofd]

K guys could be we have to run the full timespy benchmarks iso just the GTs... Thx Wizerty for testing, we will await the final call by mr Mad

hope to have news ASAP

Edited November 2, 2018 by Leeghoofd

[Micka]

I have no other comments，I hope that the final decision as soon as possible.

[Wizerty]

@Mickahave you tried to submit GT only on  3DMark  ?  When I try it I have this error :

"Your result has the following problem(s): Non-default settings were used in benchmark. Score could not be calculated. (What is this?)"

It's a problem, I think, because this message is also showing when you tweak driver...
That mean we won't be able to see if OCer did GT only or also tweak driver.

I don't care, I just need to know how to fight and I don't want an unfair battle, but I'm agree that final decision should be quick !!!!

As we don't need high CPU frequency and don't need tight mem for 3D bench (GT score) then doing full run is not difficult, it's only 30sec and with CPU et MEM not on the edge it's easy to handle... I guess doing full run is safer until final decision

Edited November 2, 2018 by Wizerty

[Micka]

I have been running the entire test，I just hope to make the final decision earlier.

[FireKillerGR]

Did the same yesterday and had the same issue.

Best will be to run default/full test and hit escape during loading of phyrics.

Then you only get Graphics Score that is valid.

[Leeghoofd]

For the Qualifier both Timespy stages require a full run, the janitor has decided!!!

The Total Score counts for the compo and it has to be validated and be valid at UL website too (don't forget that!!!)

Edited November 4, 2018 by Leeghoofd

[FireKillerGR]

So after running with CPU Test disabled ("custom run"), then full run with cancelling the physics test (says score is valid); we got to re-run for the 3rd time cause physics is needed?
Ez 30+L wasted

[Wizerty]

Can you please confirm that we need  fullrun valid, but only GT score is used for ranking ?

I hope so because if not, that mean all people with less than 6c/12t CPU have big disadvantage now while it didn't matter before... SuperPi don't need cores not HT, same for mem frequency, same for gpupi, same for Time Spy (X).. but if overall and not gt score for ranking....

[Leeghoofd]

FULL RUN guys, like a normal TimeSpy submission for the Bot ranking... contest has been altered, sorry for any hassles...

[IvanCupa]

So, it will be the total TS & TSE score counts for the ranking? Just like normal HWBot ranks for TS & TSE?
Not the GT score only?!

This change makes the (my) setup very different, hahah ?

[ikki]

Change of rules in the remaining two weeks? What?
That's a hard time for me

[Leeghoofd]

if you already had benched everything, you only need to redo 2 benchmarks... in 12 days... don't shoot the piano player

[IvanCupa]

Hi, I seem to notice that CPU-Z 1.86 sometimes not show SPD (1.85 is fine), and even the validation sometimes doesn't show SPD on this CPU-Z version.

Since we have to validate on CPU-Z 1.86 from Maximum Memory clock stage, Anyone else got same problem of missing SPD?

 

 

[unityofsaints]

2 hours ago, IvanCupa said:

Hi, I seem to notice that CPU-Z 1.86 sometimes not show SPD (1.85 is fine), and even the validation sometimes doesn't show SPD on this CPU-Z version.

Since we have to validate on CPU-Z 1.86 from Maximum Memory clock stage, Anyone else got same problem of missing SPD?

 

 

I think one workaround worth trying is to launch 1.85 and select the SPD tab, then launch 1.86 - somehow this helps to unbug the SPD tab on some platforms You can then close 1.85 and take your screenshot.

Edited November 7, 2018 by unityofsaints

[bolc]

this bug can happen a lot. my workaround is to have 2 folders of cpu-z 1.86 (or keep the zip file closeby). delete the buggy cpu-z folder one and replace it with a fresh copy, then it will show the spd

somehow cpu-z gets "corrupted" and stops displaying spd tab or the real voltage and shows Vid instead

hope this works for you

[Splave]

make the cpu-z files read only  right click properties 

[pro]

hey Alby, Pete is just trying to submit his gpupi 2.3.4, its not accepting his submission, erroring with invalid version when submitted

this thread says 'Stage 3: GPUPI-1B version 3.2 or below' and the GPUPI rules page itself says 'V2.3 (mandatory from August 1, 2016)' http://www.hwbot.org/news/11973_application_145_rules/

can we please make sure this submission is accepted as all official stated rules were adhered too

thanks

[Leeghoofd]

plz send me the gpufile I'll have it checked

[websmile]

Seeing some of the results I was never more happy not to be a mod anymore...

[bob(nz)]

Haha - run for the hills

[Leeghoofd]

K I'm not at home atm, will moderate tomorrow evening... sorry for the delay

 

update: I pmed MAD already with the driver releated issues and the GPUPI which was not accepted by the Dbase.

Edited November 17, 2018 by Leeghoofd

[pro]

Thanks for working this out alby and rich

[pro]

11 hours ago, websmile said:

Seeing some of the results I was never more happy not to be a mod anymore...

Unfortunately zerodan and micka are using the wrong nvidia driver version so both will have to be disqualified

Both unfortunate as they put up great timespy scores

[websmile]

You are slooow to mention this xd - Albrecht will moderate tomorrow and this was reported already a while ago. I have no idea why the rule was made the way it is, for a new card and new whql drivers with perfect readout and no glitch, but it was officially stated and visible to anyone, so the results are invalid because driver is too old ....

[pro]

24 minutes ago, websmile said:

You are slooow to mention this xd - Albrecht will moderate tomorrow and this was reported already a while ago. I have no idea why the rule was made the way it is, for a new card and new whql drivers with perfect readout and no glitch, but it was officially stated and visible to anyone, so the results are invalid because driver is too old ....

just woke up and looked at everyone ??