The official HWBOT Team Cup 2015 - SC2: GPU Challenge thread.

[Marquzz]

Well, I went through 5 rage 128 Pro cards until I actually found one that was identified as a Fury in GPU-Z

[Massman]

@Strunkenbold, what's the verdict on the Fury cards?

 

Looks like the old ATI Fury is as problematic as the new AMD Fury

[Strunkenbold]

New Fury is problematic? This must be a joke.

 

These old ones are a mess.

There was Rage Fury based on 128GL and a Rage Fury Pro based on 128Pro. Thats all we definitely know.

The existence of Rage Fury Pro category is actually a mistake. In the same way we would have to create a XPERT 2000 PRO or a Rage Magnum category. But that wont make sense.

If you ask me either allow only Rage Fury Maxx to make sure only "real" rage cards do participate or allow Rage 128 and Rage 128 Pro with the slight possibility that some non Fury card slip through into the competition.

 

Work is in progress and I really like to say thanks to havli for helping me with research.

I just created a mess in the atom section and need to clean that first after going on with Rage.

[viper-rd]

can anyone remove the fake results ??

http://oc-esports.io/#!/round/team_cup_2015_sc2/2366/catzilla_576p_lga775_igp

the first two results are not made in the 775 integrated graphics

Fur test (GPU) the result is too large.

In this case, only the tab graphics card in CPU-Z guarantees that only an integrated graphics card

[xxbassplayerxx]

can anyone remove the fake results ??

http://oc-esports.io/#!/round/team_cup_2015_sc2/2366/catzilla_576p_lga775_igp

the first two results are not made in the 775 integrated graphics

Fur test (GPU) the result is too large.

In this case, only the tab graphics card in CPU-Z guarantees that only an integrated graphics card

 

What makes you think Rasparthe's result is invalid?

 

http://hwbot.org/submission/2971828

[Rasparthe]

can anyone remove the fake results ??

http://oc-esports.io/#!/round/team_cup_2015_sc2/2366/catzilla_576p_lga775_igp

the first two results are not made in the 775 integrated graphics

Fur test (GPU) the result is too large.

In this case, only the tab graphics card in CPU-Z guarantees that only an integrated graphics card

 

Why go right to the cheat accusations? I can assure you that I have no time for cheating, not in my submissions or my teams. I moderate them all the time. I'm sure the mods can attest to that.

 

Catzilla was a bad choice of benchmark for this stage since it barely, barely runs on 775 IGP. Maybe its throwing up borked runs, I don't know, there are no existing runs for comparison. I saw I was higher than the runs that were (AwardFabrik in 2nd) there but I was using better hardware (Quad compared to Dual) and higher clocked RAM (important for IGP runs) so didn't see a problem. I have other runs in the 300s as well. I didn't get close to 1st place but needed more time to work out the runs I figured.

 

There is no need to head straight to cheat accusations. You reported the run, the mods can decided. If its not valid, I'll gladly take it down, I'm interested in seeing your runs you are basing the accusations on.

[Bones]

Looks legit to me - I checked that and yes, the board model he used does have onboard graphics.

His GPU-Z reading shows he did use the onboard IGP so from here it looks valid.

 

However let the mods do their job, I do believe this one will stand but if it's borked somehow, I'm sure they'll say so.

[Johan45]

If any of them are bugged it's the 2332 run that's in first right now. That looks very out of place and not all the test scores are showing.

[viper-rd]

Rasparthe

OK

Tell us - how you got on the GPU test values at the level of GeForce GTX TITAN+:

Hardware (GPU+CPU) - 196

Physics (CPU) - 68

Fur (GPU) - 5843

Fluid (GPU) - 14551

Raymarch (GPU) - 4044

I specifically looked at the results of more than a dozen - on any weak graphics card is no such values.

if you claim the correctness of your result - prove that the system is no more cards.

Make unwrapped screenshot or video.

you're right - I apologize to you.

 

i3 2120/8gb/intel HD graphics = 554

Hardware (GPU+CPU) - 576

Physics (CPU) - 66

Fur (GPU) - 42

Fluid (GPU) - 32

Raymarch (GPU) - 21

 

I would like to have seen a response moderators...

[Rasparthe]

You have obviously done much more research on the subject than I have. You are correct though something doesn't look right. I stand by the fact I only put up the score because it appeared in line with the ones that were posted, but I didn't look closely at the subscores. I will gladly take it down.

 

In the future, you may want to start with asking mods to look at a score, give a reason why you think their is an issue and then ramp up to cheat accusations. Starting the other way just looks like sour grapes. I still look forward to seeing your scores though I imagine all that research will pay off.

[viper-rd]

you are my questions are not answered.

what is the point to argue with you?

let them decide moderators.

[TASOS]

There is really no need for cheat accusations.

 

Rasparthe is a valued member of this community and if his score is out of line that would be an honest mistake and nothing else.

[viper-rd]

TASOS

My offer is valid.

Show the result with an open tab graphics cards in CPU-Z.

@you're right - I apologize to you@

[xxbassplayerxx]

viper, have you tried running Catzilla on 775 IGP yet? Give it a try and report back with your findings.

 

Quick hint, the benchmark won't even open if your memory isn't above 600MHz. Not to criticize whoever chose this stage but Catzilla is about the worst benchmark possible to run on IGP. I'm sure it's bugging on nearly every run.

[I.nfraR.ed]

No need for accusations in cheating.

Sure, the run is bugged, but you know what... it happened to me several times on various cards (much newer and powerful than this).

It is most probably a stability problem. At a certain point in the benchmark the system fails and the points for subtests get inflated a lot.

 

I can imagine on that slow system you can't really tell if the benchmark froze or still running normally. Don't have such system, but I bet it is pushing it very hard.

 

As it is said: you're innocent by default, until proven guilty.

Edited September 9, 2015 by I.nfraR.ed

[Aleslammer]

Bench does throw a lot of whacky scores and a great number of folks haven’t ran it that much so don’t know that much about it. That’s one good thing about it being used in the TC by letting folks know what is generally used for determining bugged runs.

 

Personally have had three or four runs way out of line with cards that others had posted scores with so had something to compare to. Ran a Quadro FX 570 three or so months ago which had a high Raymarch compared to the other sub tests but was present on every run, did note it in the remarks section. Cards before and after seemed in line same hardware and OS.

[ObscureParadox]

viper may I suggest you take a breathe and calm down for a second, as others have stated if any mistake was made then it would be an innocent one and not intended by Rasparthe.

 

I suggest you let the mods do their work and not start a witch hunt in the forums

[Rasparthe]

New Fury is problematic? This must be a joke.

 

These old ones are a mess.

There was Rage Fury based on 128GL and a Rage Fury Pro based on 128Pro. Thats all we definitely know.

The existence of Rage Fury Pro category is actually a mistake. In the same way we would have to create a XPERT 2000 PRO or a Rage Magnum category. But that wont make sense.

If you ask me either allow only Rage Fury Maxx to make sure only "real" rage cards do participate or allow Rage 128 and Rage 128 Pro with the slight possibility that some non Fury card slip through into the competition.

 

Work is in progress and I really like to say thanks to havli for helping me with research.

I just created a mess in the atom section and need to clean that first after going on with Rage.

 

So what is the official ruling on this? Are Rage Pro cards allowed? Only MAXX? Only 4 days left in this stage so would like the clarification. Personally I can't seem to find a Rage Fury Pro, only Rage 128 Pros.

[Strunkenbold]

You can submit with your 128 Pro card in the Rage Fury Pro category as I probably will merge them later.

[TASOS]

... Personally I can't seem to find a Rage Fury Pro, only Rage 128 Pros.

 

Rage 128 Pro is Fury Pro.

[Strunkenbold]

Ok I finally sorted those Rage cards. Rage Fury cards where always 128bit and 32Mb Memory. Means, you can submit with:

Rage 128 (128bit, 32Mb)

Rage 128 pro (128bit, 32Mb)

And dont forget to attach a GPU-Z or Everest / AIDA window! Best is both.

 

Massman, please also allow Rage (128bit, 32Mb).

[TASOS]

Current version of gpu-z , is practically useless.

The only thing it can read correct , is the device-id.

 

It wont read the correct bus width.

[Strunkenbold]

It may display wrong things but it still shows useful information.

If there is SDR and 64bit displayed means: SDR 128bit

If there is DDR and 64bit displayed means: SDR 64bit

 

To be really sure you have to use everest / aida. Powerstrip is not enough.

I had to drop many results in the 128bit section because of no validation though I know that all score under 1000points are basically made with 64bit memory interface.

[Spiedie]

Just to be sure before i start a run with aquamark, windows 8.1 IS allowed for unoverclockable systems right? (Dell venue 8 pro with Atom Z3740D in this case)

[xxbassplayerxx]

Honestly, I think not. It may be logical but it's still outside of the rules.