MSI Big Bang Xpower discussion thread: Overclocking results and discussion

[Leeghoofd]

That's a secret tweak Bill, only works on MSI boards... Asus, nor eVGA ones don't boot then...

 

Nice results PJ, I want one sample to compare with the REX III later on... new REX bios shows big improvements for older bloomfield CPU's...

[miahallen]

Ooo...looks interesting, keep us posted Pieter

[Snoro]

Do you have any word on the release date of the xpower Massman.Need a x58 motherboard for this month with atleast 5 pcie slot and definitly interested in this one.Also, please tell me there is no nf200 in it.

[marmott]

no nf200 on it

[Massman]

Do you have any word on the release date of the xpower Massman.Need a x58 motherboard for this month with atleast 5 pcie slot and definitly interested in this one.Also, please tell me there is no nf200 in it.

 

- No NF200

 

- Prices are said to be ~300$, but I see that EU shops already go to 300€ and below, so $ might decrease

 

- Full retail vailability in two/three weeks. I was told early June.

[Massman]

Ooo...looks interesting, keep us posted Pieter

 

Will go to a friend's place to run 980X + Lightning on LN2 with the board. First some hardmods (will share if I don't kill the board) ...

[Hondacity]

the mobo needs mod? what is it for?

 

i like how they used a 3x2 connector for pcie power....more juice as 6 cables will provide more power than a puny molex cable

 

the low profile components with the cpu socket is a good thing too....big pots will fit i suppose...

 

good luck testing..and don't kill the board:D

[Snoro]

- No NF200

 

- Prices are said to be ~300$, but I see that EU shops already go to 300€ and below, so $ might decrease

 

- Full retail vailability in two/three weeks. I was told early June.

 

Hope so, would be my limit for release date, would make the perfect cruncher/folder/bencher combo in one mobo.

[Massman]

the mobo needs mod? what is it for?

 

I don't if they nééd the modifications, I just got them in the mailbox .

 

One's to increase OCP to 320A, the other one's to increase input power.

[Hondacity]

ahhh thats why you had problems with 1.9v....

 

i hope the mod fixes that issue...

 

increae input power..hmmm capacitor mod?

 

the xpower comes with the dashboard right? is it easy going thru the dashboard program?

[Massman]

The Vcore issue is totally CPU related as I had the exact same behavior with the UD7. It almost looks like the CPU isn't getting enough cold to deal with the high voltage properly. I mentioned the CPU was bad, no?

 

The Xpower comes with a dashboard indeed. It's the same as the one that came with the Trinergy, but with an updated firmware. So, if you want you old one to work, you need to get the new firmware. For anyone who needs it to test the board, let me know and I'll drop it all here. The Dashboard is working good enough if you are familiar with it. Now quite as elaborate as the OC app that Asus is giving to selected overclockers with the R3E, neither as flashy as the EVBot, but just a simple functional tool .

 

I'm trying to get MSI to add a couple more features for the Dashboard that should make it even better. Fingers crossed.

[Hondacity]

i didn't catch the cpu was voltage/cold deprived....hows leeghoofd's 980? better than yours?

 

*folds his fingers upside down....

 

i hope its still early so they can revise the dashboard.....

 

so with experience on ud7/r3e/xpower do you think the capacitor technology they employ sorta limits how low the temps could go? or the phases really help? were you able to clock your 980 on a 760/762?

 

thanks for your feedback...

[Massman]

Leeghoofd doesn't have a 980X (yet), but Gamer has. As far as I can see, his retail should indeed be better than this B1 ES.

 

The dashboard won't be revised of course, it's the still the same in the retail package. I might have send out the wrong message in my last post: the Dashboard works just fine for overclocking in windows. In fact, I don't use applications such as SetFSB anymore, which is quite a significant change in my overclocking behavior. The only thing it's not is look as fancy as EVBot and it's not as elabore as Asus' ROG panel. But who says that's necessary

 

It's too soon to be talking about technology limitations, at least for me. At the moment, I have only properly tested this 980X on UD7 and Xpower; the R3E was just a very, very quick session. I heard there's a new bios out that improves the overclocking, so I'll need to retest that board before jumping to conclusions. Haven't tested the 760 or 762.

[danit]

when is this board going to be availbe and for how in the uk or in usd dollars

[Hondacity]

whats the switching frequency range of the cpu pwm circuit?

[Massman]

There's no selectable range available (at least not in this bios)

[Massman]

Last update.

 

Yesterday, Madshrimps OC Team had a small bench session / barbecue with 180L LN2. We mainly focussed on two VGA cards (GTX280 and R5870) of Gamer to see if we can push the team total a bit further. Some notes:

 

- The coldbug/coldbootbug/overclockability was better than on R3E: with Xpower we managed -145°C coldbug, -127°C coldbootbug and 107°C imc coldbootbug

- This retail was pretested at 5.8GHz 3DMark06, but yesterday we managed 6GHz flat at 1.83V through 3DMark06 CPU test. With a bit of extra finetuning, I think it's possible to get 6GHz for Vantage

- The SuperPI 1M problem with Gulftown looks to be cache related - 6GHz CPU test, no 6Ghz 1M

 

Results:

 

- http://hwbot.org/community/submission/1008065_gamer_superpi_core_i7_980x_6sec_781ms

- http://hwbot.org/community/submission/1008056_gamer_3dmark_2005_radeon_hd_5870_44475_marks

- http://hwbot.org/community/submission/1008061_gamer_pifast_core_i7_980x_13.5_sec

- http://hwbot.org/community/submission/1008058_gamer_aquamark_radeon_hd_5870_438484_marks

- http://hwbot.org/community/submission/1008057_gamer_3dmark_2006_radeon_hd_5870_34120_marks

- http://hwbot.org/community/submission/1008067_gamer_wprime_32m_core_i7_980x_2sec_656ms

[Hondacity]

xpower better than r3e?

 

or r3e better than xpower?

 

is that what you meant?

 

what was the vtt?

 

vcore vtt adjustments ...available on the dash board?

 

great teamwork...i love the aquamark run..i little more cpu speed and your nearing 450k

 

******ocp mod performed? if successful can i see pics?

Edited May 25, 2010 by Hondacity

[Massman]

Xpower > R3E.

 

It was quite amusing to see how I had to force both Gamer and Leeghoofd to try out the Xpower and in the end they were both like "okay ... this was good", hehe. The genuine amazed overclocker look. The R3E was bugging at -90°C, whereas the Xpower went to -145°C load (which is like Classified apparently). Together with the lower temps, the chip was also producing higher MHz on the CPU and thus higher scores.

 

VTT was 1.53V - in contrary to my ES, this retail was capable of booting with 1.5V+ VTT.

 

Yes, VTT is adjustable in the OC Dashboard

 

By the way, the only modification I did was increasign the OCP on the board.

[Hondacity]

which bios of the r3e did you use?

 

when you change vcore inside windows...how fast does it change compared to the r3e? i remember the r3e vcore change was almost instantaneous

 

were you able to check max bclk on the retail 980x?

[Hondacity]

 

did this help with max bclk?

[BenchZowner]

Hondacity that switch ( jumper-set ) is an enable/disable switch for the LoadLine calibration function for the UnCore ( Vtt aka Vdimm/Vqpi ) voltage.

With the jumper @ disabled the QPI Voltage drops & droops.

[SF3D]

Xpower > R3E.

The R3E was bugging at -90°C, whereas the Xpower went to -145°C load (which is like Classified apparently). Together with the lower temps, the chip was also producing higher MHz on the CPU and thus higher scores.

 

That is simply because you got pre production (not retail) sample which have different chill controller. That is the cause of those weird temp issues.

 

Last weekend we did some runs with no cold bug ES and no cold bug retail My retail was able to handle -160C too.

 

So, those judgements are solely based on your personal experience and I think it should be mentioned. Now it looks like it is a common trend to have worse scores with R3E

 

I think R3E is very hard board to beat in any aspect. That is my opinion and thats all

[Massman]

******ocp mod performed? if successful can i see pics?

 

I'll take some tonight/tomorrow. It's a very simple mod.

 

which bios of the r3e did you use?

 

when you change vcore inside windows...how fast does it change compared to the r3e? i remember the r3e vcore change was almost instantaneous

 

were you able to check max bclk on the retail 980x?

 

- Bios: can't remember

 

- Changing the voltages in Windows goes quite fluent, no problems. Maybe a few milliseconds delay.

 

- Max BCLK: haven't tested it. Do note that there were some issues with the Bloomfield: 220MHz was bootable, but above that it's very hard to get boot-to-windows stable. Leeghoofd's testing the completely retail PCB with fully retail Bloomfield, but there's still a new bios left to test. More on that soon, I think.

 

That is simply because you got pre production (not retail) sample which have different chill controller. That is the cause of those weird temp issues.

 

Last weekend we did some runs with no cold bug ES and no cold bug retail My retail was able to handle -160C too.

 

It's very possible that the sample is to blame for the bad results (under cold) - my ES had this problem as well (although same overclocking results). I guess it's maybe a good idea not to send these kind of non-retail media samples to reviewing websites (madshrimps) that actually test boards what they are made and marketed for.

 

So, those judgements are solely based on your personal experience and I think it should be mentioned. Now it looks like it is a common trend to have worse scores with R3E

 

I ALWAYS post my personal findings and experience in threads. And that experience tells me that both with the ES and the retail 980X, the R3E does indeed give me lower overall clocks. Whether that is due to the sample or even the overclocker is a topic I'm not touching. I think that recent overclocking results with R3E from various sources have proven that the R3E is a very decent board as well.

 

People who read my posts as truth and only available truth are ignorant and shouldn't be reading it to begin with .

[Hondacity]

you guys should keep posting

 

but yeah i do process your information... i don't accept them as is

 

the coldbug is the hindering issue...if we're able to fix it like some ****ish guy mentioned it will be awesome..