Important Rules Update! 3DMark01 to be Replaced by 3DMark Vantage in G.SKILL OC World Cup Online Qualifier

A couple of days ago, HWBOT members reported a bug affecting one of the stages in the G.SKILL OC World Cup 2016 Online Qualifier. In 3DMark01, when the integrated graphics is cooled to a temperature close to the cold bug threshold, the Frames per Second increases significantly. Typically this kind of behavior is detected by Futuremark's System Info, but since 3DMark01 has been discontinued by Futuremark there is no way to validate this issue online. Following this finding, the competition organization has decided to replace the 3DMark01 stage with another stage.

To ensure a clean and fair contest, HWBOT and G.SKILL have decided to reset the 3DMark01 stage and include 3DMark Vantage in the World Cup Online Qualifier. The 3DMark01 benchmark will also not be accepted as community choice benchmark for the Live Qualifier. In addition, the qualification time has been extended to 18th of April.

We apologize for the delay on detecting this issue and all the inconvenience caused by this change. For further discussion and information, please visit the competition page and the event discussion page.


28

Belgium Massman says:

The competition page has been updated accordingly.

Sorry for the inconvenience and hopefully this does not affect too many people :(

Germany Dancop says:

I'm out after 3 dead igp! Honestly...this shouldn't be the main part of a memory vendor. Mem definitely pushes the score, but it also does it with the igp at stock. A max freq on igp kills them so freaking fast and without any reason. I know, this is personal bad luck, but I Def don't wanna kill more chips by now...

United States rtsurfer says:

1 dead IGP here. Could've been more if I had more Chips. It was definitely Core PLL in my Case. Need to be careful with that Voltage, as Elmor suggested. Sent from my A0001 using Tapatalk

Belgium leeghoofd says:

Ouchaaaa

Germany Dancop says:

1.4v here! That's not much....

United States rtsurfer says:

Welp 1.4V isn't that high. :( Sent from my A0001 using Tapatalk

Germany Dancop says:

One died at 1.65...I thought, well...let's go lower! Then 1.55...died in bios. Then 1.4...died in bios as well. All chips except the first one at 160 degrees. ..MINUS! Sure, maybe I did something wrong, but to prevent that, I won't play igp any longer!

United States GtiJason says:

Dancop said: One died at 1.65...I thought, well...let's go lower! Then 1.55...died in bios. Then 1.4...died in bios as well.
All chips except the first one at 160 degrees. ..MINUS!
Sure, maybe I did something wrong, but to prevent that, I won't play igp any longer!


Sorry to hear, we are having similar problems in the US. Was it Core PLL or the IGP voltage that causes this?

Germany Digg_de says:

I've pushed my iGPU Voltage up to 1.50v.. don't remember if i put LLC1 or LLC2 on it. But i have never pushed the PLL Voltage over Standard i guess.

Germany Dancop says:

The problem seems to be high cpu core, cpu voltage , pll and vgpu...all together makes them die like flies!

Poland Xtreme Addict says:

Massive graveyard ^^ soon I guess I will join the gentlemen's club :banana:

Indonesia Hazzan says:

4 CPU IGP die here...so many risk bench with IGP really

Germany Dancop says:

Dancop said: The problem seems to be high cpu core, cpu voltage , pll and vgpu...all together makes them die like flies!


Oh...forgot to say, one died at 5g cpu with 1.4vcore, 1.4 core pll, 1.5 vgpu. Was just testing it.

Romania Alex@ro says:

Igp fuck fest :D Sent from my SM-N910F using Tapatalk

Australia TeamAU says:

Watch the vpll, better to keep it low as possible

Just work it up as you would any unknown platform, keep taking scores as you go

We know vpll is an igp killer, so focus on the biggest bottlenecks like memory instead of pushing the igpu too hard

Indonesia Lucky_n00b says:

So many dead IGP :eek:
It's weird, during our local amateur competitions (using air/water) with i7-6700K with IGP(and 3DMark Fire Strike/Cloud Gate), I don't find dead IGP even though some of the contestant gives it 1.45-1.5V Vaxg(GT Voltage).
I wonder what volts under cold is causing this..I'll try benching one of my banana CPU under cold next week.

Belgium Massman says:

Guys, be careful when pushing something new. High voltage and current tolerance for CPU cores doesn't mean the IGP can also handle it ...

United States Splave says:

You need the pll voltage. On air it's 1300mhz vs 1400mhz Max. Its like saying don't oc so hard in an oc contest for 10k. Can we just say igp was a bad idea and move on? :)

Germany Dancop says:

Keep on going with IGP, BUT!!!! WITHOUT OVERCLOCKING IT!

Christian Ney says:

[IMG]https://media.giphy.com/media/GjYjLvGErsggg/giphy.gif[/IMG]

Germany BenchBros says:

Just find the max stable setting without killing it... its a new kind of low clock challenge :D

United States Gunslinger says:

Splave said: You need the pll voltage. On air it's 1300mhz vs 1400mhz Max. Its like saying don't oc so hard in an oc contest for 10k. Can we just say igp was a bad idea and move on? :)
IGP was a bad idea, good luck everyone! :D :celebration:

Indonesia Lucky_n00b says:

Splave said: You need the pll voltage. On air it's 1300mhz vs 1400mhz Max. Its like saying don't oc so hard in an oc contest for 10k.


Hi Man,
This 'pll voltage' you were talking about and giving the IGP more clocks, is it the Core PLL, PLL Bandwidth, PLL Termination, or Internal PLL Overvoltage? Interested in testing what is killing these IGPUs :)

Germany Dancop says:

Core pll plus high vcore from what I saw

FlanK3r says:

Its logic, the same danger is at APU. If you raise too much CPU vcore together with VDDNB, it you can kill the chip (air or LN2, the same results)

United States Splave says:

can we change the final to 2d please? this will be a boring live comp when half the people dont have scores

Germany Dancop says:

What splave said...+1 Lost chip #4 although I was careful with voltages. Those igp are so damn fragile! Or, if igp is a must, for a reason, I don't understand, limit it to 1400 mhz or something. That way, the mem is much more important than a nicely binned igp I've tried the difference between strong mem and 25mhz more on igp. That means... 1450 igp with 4000c12-11-11 vs. 1475 igp with 2133c15 stock mem. Gt3 in 11 equals in 15.42 vs. 15.82 fps. This benchmark is heavily mem related but the igp does its work much better ;)

Romania Alex@ro says:

You guys brought decimation on the poor IGP's :D

Please log in or register to comment.

Leave a Reply: (BBCODE allowed: [B], [QUOTE], [I], [URL], [IMG],...)