MSI 580 Lightning

[loopy83]

I have heard a very strong crack and that was it, card never played again.

Sounds like the pot pops off the HS... this happens sometimes...
DCU have the same problem, ask stummerwinter

[slamms]

matose i hope your frequencies for Lightning on hwbot is no best?

[Kabauterman]

has anyone tested the difference between highest GPU Clock with Mem-OC and without Mem-OC ?

[slamms]

i am not tested, but think that difference will be

[matose]

Vid of our card was 1.13v

Are you sure, because that VID is huge? We have 1.00v VID

matose i hope your frequencies for Lightning on hwbot is no best?

From your comment I understand you have 1450-1500MHz card . Unfortunately I don't have much headroom left...

has anyone tested the difference between highest GPU Clock with Mem-OC and without Mem-OC ?

Yes, there is some difference but not big! It depend a lot of the temperature and voltage of the memories...

vPLL helped you guys? I didn't find any gain from high PLL...

[boblemagnifique]

our card's mem OC decreases much after GPU is totally maxed out. e.g. 1330 RAM -> 1220, though when GPU is not - RAM seems to be perfect.
Gained extra 30-50 MHz GPU and max stable RAM clocks after the mod.

Interessing

We have the same frequency RAM on air & ln2 = 1200Mhz

[matose]

Huh? I have 1300MHz on air and 1400MHz+ on LN2

[Kabauterman]

Are you sure, because that VID is huge? We have 1.00v VID



From your comment I understand you have 1450-1500MHz card . Unfortunately I don't have much headroom left...



Yes, there is some difference but not big! It depend a lot of the temperature and voltage of the memories...

vPLL helped you guys? I didn't find any gain from high PLL...

PLL doesn't change anything. Crazy to see DCU running best on 1.3V and Lightning on Stock 1.1V or something.

[boblemagnifique]

Huh? I have 1300MHz on air and 1400MHz+ on LN2

we have test other 1pc o of "Sample Press" -> 1170Mhz RAM on Air .... ahahahah

As often , Good Ram or not ... (if MSI can test before just the frequency ram max on air, it's more fair ...) after it's different for the GPU of course ...

Same for the Final French -> our HD6870 have 1250MHz RAM but the others teams 1150/1170MHz max ....

it's a life , many chance or not .... and i waiting to see the bclk max for the different 2600K with the multi 50x

[matose]

I had some 103-104MHz samples as well as 106-107MHz samples, so this is anything but fair not. The best tweak is frequency, so luck rulz everything in SPI32M

[ME4ME]

103x50 or 107x50, its 200mhz of a difference.

much better than "full out" and having 500mhz of a difference between chips.
I say it was a good call of MSI

[matose]

200MHz in SPI32m can make the difference between first place and last place . Don't you think so?

[Kabauterman]

I promise there will be some chips with a Clock wall arround 5200 to 5300 MHz

60% of the Sandy es we tested had this issue

[matose]

We had a couple of ES's from Intel (which we will get in Istanbul) and all of them were over 5400MHz (some with 2 cores but still that's ok)

[Alex-Ro]

Why are you guys complaining even though the competition hasn't started yet?

[Kabauterman]

nothin better to do bro

[loopy83]

Are there 1 or 2 CPUs per team?
With 2 CPUs the game is a bit more fair...

A good GPU is not enough. When you have a CPU maxed out at 5200MHz in 03, GPU over 1400-1450 is senseless.

[ME4ME]

200MHz in SPI32m can make the difference between first place and last place . Don't you think so?

Which place isnt important, since the competition system isn't based on places but on actual performance

So in worst case .. your 200mhz behind of the top. Thats what? .. 15sec? With good tweaking you may be able to reduce the gap to 10sec. In procentual performance, thats not very much, and with a good VGA card you'll be more than able to make up for it in the 3D part of the contest

[boblemagnifique]

i have a friend have a good 2600K on Bclk -> 109.8Mhz on air but the max frequency is 5.2GHz

It would have been fairer to limit the frequency, for example has 5000MHz and not multi 50x for the Spi 32M (as in Tweaking GBT) -> to have a little difference (the memory frequency / timing and tweak OS).

But it's the rules !! Fingers crossed for a chance with us

[massman]

I promise there will be some chips with a Clock wall arround 5200 to 5300 MHz

60% of the Sandy es we tested had this issue

Tested 50 chips of same batch:

~ 8% chance of < 5000 (bad)
~ 10% chance of > 5600 (good)

+50% chance of in between 5200-5400 MHz (average)

[matose]

A good GPU is not enough. When you have a CPU maxed out at 5200MHz in 03, GPU over 1400-1450 is senseless.

1450MHz GPU in 2003 is a good Lightning card from what I saw discussing with different MOA competitors. GT4 will scale over 1450MHz even at 5200MHz CPU

[Kabauterman]

1450MHz GPU in 2003 is a good Lightning card from what I saw discussing with different MOA competitors. GT4 will scale over 1450MHz even at 5200MHz CPU

It's far far away from Scaling with 5200 MHz on the CPU.
the Gain with every 50 MHz of the GPU get lower and lower but lets discuss after the competion.

A lot of stuff to talk about at the Evening dinner *jummy*

[Vivi]

elmor has won almost every competition, if he finds a way to win every time i think the hardware scaling is fair enough :P.

we don't want a low clock live competition.

Then should be called MSI master tweaking arena

[aristidis]

Are you sure, because that VID is huge? We have 1.00v VID

Yes it was 1.13V..Vid 1.00V is the best you can find but it;s not always important. Maybe a card with higher vid do better at cold. Who knows

[pastelidis]

Which place isnt important, since the competition system isn't based on places but on actual performance

So in worst case .. your 200mhz behind of the top. Thats what? .. 15sec? With good tweaking you may be able to reduce the gap to 10sec. In procentual performance, thats not very much, and with a good VGA card you'll be more than able to make up for it in the 3D part of the contest

i have a friend have a good 2600K on Bclk -> 109.8Mhz on air but the max frequency is 5.2GHz

It would have been fairer to limit the frequency, for example has 5000MHz and not multi 50x for the Spi 32M (as in Tweaking GBT) -> to have a little difference (the memory frequency / timing and tweak OS).

But it's the rules !! Fingers crossed for a chance with us

we have test other 1pc o of "Sample Press" -> 1170Mhz RAM on Air .... ahahahah

As often , Good Ram or not ... (if MSI can test before just the frequency ram max on air, it's more fair ...) after it's different for the GPU of course ...

Same for the Final French -> our HD6870 have 1250MHz RAM but the others teams 1150/1170MHz max ....

it's a life , many chance or not .... and i waiting to see the bclk max for the different 2600K with the multi 50x

PLL doesn't change anything. Crazy to see DCU running best on 1.3V and Lightning on Stock 1.1V or something.

103x50 or 107x50, its 200mhz of a difference.

much better than "full out" and having 500mhz of a difference between chips.
I say it was a good call of MSI

200MHz in SPI32m can make the difference between first place and last place . Don't you think so?

Guys dont forget we will have 2 cpus for one team