M17x R2 7970m CrossfireX Ultimate Installation and Tweaking Guide

id didnt get stuck with the 1.1Vddc vbios.

i didnt install any driver after the flashing
i will try to reinstall ccc though

its strange that it doesnt do that to the dell (main) card

by the way, ULPS is suposed to do what? disable the lower power settings ? the 0.825V 300gpu 150 mem one?

i remind you that i havent put thermal pads on the power mosfets. I wonder if that is bad or good. But as i told you my temps now in gaming are 76-80 max with a 30C ambient temp (on winter with 17 ambient they will sure get lower to sub 70C)

 

I'll drop you a PM now. You're damn right you're lucky We have to resort to all strange workarounds that don't even make sense sometimes

Lol, mines is messy too. It looks like a frankenstein stitched up GPU made from thermal pad

... I think they're jealous from the pure awesomeness in this thread. Well either that or its cos the three same people keep posting here every single day

 

It's not ideal but neither have I
ULPS shuts off your cards when not used, and is particularly bothersome since it interferes with Crossfire / Overclocking / Sleep states / Sensors etc. Disable it (unless you want "decent" battery life lol )

 

my thermal paste job on my cards is not the best too, dont forget that while i was trying to fix my cards i exchanged them from main to second and back almost 10-15 times... all this time i used the paste they had on, i didnt repaste, and the pads are a mess, with cuts and point of no connect. i will order new ones and make a proper repaste now that nospheratu fixed my vddc issue.


PS. have you tried this on the rift?
Blue Marble: The Rift's First Space Audio Player

one of the best demos i have tried. gives me no naussea and its awosome, just imagine these space views through the cockpit of a 300i in star citizen!!!

 

Can't help incrementing the awesome though I was running an eye over the new machines benchmark threads for those equipped with 780m and having myself a little chuckle, although I'll admit that we can't pull of a P370SM or Alienware 18 feat and stick 2x780m SLI in here But maybe R9-9970 will be fun.

Until then, our machines are powerhouses even by desktop standards.

ALL MY WANT. You're making me want to get my own dev-kit, and that's very evil of you.

 

so the final verdict is that 780m cant run on m17x r2??

i love it more day by day. the nausea was the thing that made me think of selling it, BUT its getting better day by day. half life is the hardest for me to play it made me sick in just 2 mins the first day. now i can play for 15mins and dont feel very sick. I hope to get better like people playing its for 2-3 hours nonstop.

Other demos and games with more open end enviroments are not causing me great nausia. so i can play them for 1 hour at least.
i hope that the case with star citizen.

My verdict on The rift? well its a game changer, and the future surely BUT it really needs a better screen and positional tracking.

 

i think i did somthing wrong. the checksum on the vbioses are not the same.

 

@Nospheratu, AlienHack, sangemaru you all got a rep from me

thx really. I really had no hope for 7970m crossfire to work in my 17 r2.
you guys keep this alive thats nice^^.
I read everyday whats written here but I wished i could be of some help.

 

i got it to work!!! but it shuts down as soon as i start a game. im guessing my thermal pads needs to be adjusted. either way it works on idle.

 

download a program to monitor the cards temperatures.
If it reaches 90C it shuts down.

Do you overclock your cpu? are you running 2 7970m in crossfire? maybe the shutdowns is because of your psu (power brick) or software issue

 

@pinoy_92 YGPM

 

im only running 1 7970m and my core temp is reading 37c according to hwinfo. also there is a coil whine whenever i start a game.

is there another software that let me check temperature other than the gpu core on a gpu?

 

Here's to keeping the machine alive and ruling Let's see if pinoy joins our success story, and maybe you will too Thank you.

Please enable logging in HwInfo and then run an application that causes the card to shut down. It should generate a .csv file that you can put on pastebin or upload for us to check out.

With just one card, I'd have to either rule for thermal shutdown on the GPU, or for improper heatsink contact somewhere (maybe vrm's).
Also, please share as much detail as possible:

Card Manufacturer, vBios used, type of thermal pads (with pics if possible) and positioning, etc. I once had a shutdown with just one card as well, while overclocking and benching it. Thought it was a PSU issue, but it was just overheated. I haven't had a PSU shutdown yet.
On a different machine, I had thermal shutdowns but not because of the GPU core, but because the VRM's weren't cooled and were overheating. Could very well be that. Applying pads on them fixed my issues.

I've also noticed that the card cares about how much pressure it receives in certain areas, or at least my first 7970m did. If there was too much pressure on some IC's, there would be corruption and instability (and by pressure I mean physical pressure, such as using thermal pads on components that don't need it).

Unfortunately the only temp you can check afaik is the core.

 

i figured it out. i stacked the thermal pads too high, so that the die is barely making any contact with the copper on the heatsink. its working fine now and the coil whine seems to be almost completely gone.

this was basically what was happening. VID_20130806_084749_215.mp4 - YouTube

it takes a while, probably cuz im still using a hdd. it shuts down, then it restarts itself. its working fine right now. ill be running more tests make sure everything is fine.

thanks for the help everybody!!!

 

Lol! Dude that whine was hectic!!! You shouldn't have run that machine at all like that! I'm glad you fixed it though. Good job on the troubleshooting :thumbsup:

I should have mentioned you want maximum core contact and only stack a minimum amount of layers to make contact on the heatsink and component below. The thermal pads should not in any way cause your heatsink to sit incorrectly. Actually the heatsink should sit as if you didn't have any thermal pads or paste for optimal temps.

 

My god, that whine is like nails on high-speed chalkboard. You're lucky your card didn't die lol.

 

it was obvious it was the vga though because of the screen graphics coruption just before shutdown

by the way great temps while LTC mining with cards at 99% and hashing 2 x 300Kh/s and they can go to 330 each!!

 

Awesome stuff AlienHack

Is that screenprint from an external screen?

 

how u get yours running on 1v? mine goes up to .975v, but my old 260m goes up to 1v. also any of u guys tried overclocking one 7970m?






btw, been running some games lately, ive only played some old games like bf3, da2, settlers 7 and guild wars 2.

in guild wars 2, my gpu is still bottlenecking my cpu in lion's arch on ultra settings. everywhere else i get a nice 60fps avg.

in bf3, almost the same thing. 7970m is at 99% usage and the cpu avg usage is around 60%. however, the temps on the cpu is around 95c and the gpu is around 70c. still avg around 40-60 fps on a 64 player game.

thanks again to you guys for helping me get the gpu running. :thumbsup:

 

Eh, I never ever had any problem getting volts to stick, so I'm not good for advice on that Flash a 1V vbios or edit the profiles.xml file, I'd say

Yea, the 7970m is an awesome monster, but there's still a slot free there Heheh.

In an unrelated notice, I've determined a few things:

1. I'm an idiot.
2. CrossfireX doesn't work in windowed mode (utilization sometimes goes up to 40-50% for the second card but I have no idea for what reason, but it definitely doesn't contribute to gameplay speed).
3. Fullscreen gaming on the latest 13.8 drivers is the awesomest thing.
4. My PSU isn't strong enough to truly deliver with both cards running Running Bioshock: Infinte in fullscreen mode with Crossfire enabled shuts me down after a few minutes, and the CPU is already not running high-clocked due to temps. Seeing as how I'm going on vacation tomorrow and the 330W PSU should be arriving next week, that really isn't a problem, but I feel stooooooooopid. I only noticed because I was having particularly bad performance in bioshock in one area (with absolutely every detail imaginable maxed out of course ) and checked to see and saw cross wasn't running. Setting it to fullscreen made my eyes melt in awesome.

Sooo... are you guys playing fullscreen?

EDIT: Also, Bioshock: Infinte is the most awesome game of all time ever and you must must must play and finish it. (I played the ending sequence like 15 times while experimenting with different clocks and voltages for crossfire, and it pretty much made me want to cry like a little girl every time lol )

 

Lol Shame on you sangemaru!

Yeah a lot of users think they fixed their crossfire issues when running in anything but proper fullscreen mode and then find out its working cos crossfire isn't even enabled

I managed to complete Bioshock Infinite with no shutdowns at all but I was running my CPU on stock Dell settings ie. 24/24/24/24 and 62/62 TDP/TDC and stock GPU's with my usual 0.975v and 1.00v on the cards. Framerate was always around 60fps so I never felt the need to oc the XM. I wasn't running the 13.8's though, that could be a huge factor. I also noticed that power draw did not exceed 240W and was mostly around 220W.

I'm trying to avoid any Bioshock spoilers in this next paragraph but if you are good at reading between the lines... stop right now. Its a brilliant game but the ending was a bit disappointing for me. Actually its a very good ending when compared to all the mediocre stories in other games. I just was a bit disappointed in myself for not figuring it out earlier and also that the ending doesn't make any sense to me since I'm a huge sci-fi fan and it doesn't fit into any of the concepts I've come to understand from the movies/tv series I watch. Yes, I understand what they tried to to do. I actually read a couple of reviews of the ending and it still doesn't make technical sense to me but I can appreciate the fact it was refreshing and they took a step outside the norm to bring us that intricate story.

I edited Alienhacks vBIOS to 1.00v

If I remember correctly you are running a Clevo card with the RJTECH/P270WM vBIOS? That's the exact same vBIOS and cards I'm running and I am working on creating a vBIOS pack specifically with that vBIOS for all us M17x R2 users who are running Clevo cards. Since that's the only safe vBIOS we can use with our machine without killing our screens we don't have any other options since svl7's mod pack and editor don't work with our vBIOS.

I'll also fix the "stuck at 0.975v" or "stuck at 1.00v" issue when running a default vBIOS which should be running at 1.05v (my personal issue as well).

Hopefully I'll be done with it by next week.

 

Yea, I personally believe one has to be pretty experienced on Ayahuasca rituals to be able to fully comprehend the magnitude of the Bioshock: Infinite endgame. And I'm proud to say that I could see it coming a few hours away, although not to the full level of artistic awesome it presented. Really, to me it's the best game I ever played so far.

Try the 13.8 drivers, see how you like them. If you upgrade from 13.1, the BSOD's seem to go away (full install). And I do think the consumption goes way higher than with 13.1 drivers. GPU's become properly utilized, when before they really weren't.

I run my CPU at extremely conservative settings (24/22/20/18 at 55W/40A), so yea, my PSU isn't any great performer. I even had shutdown with just one overclocked card, remember?

 

the bioshock saga never got me hooked. so i didnt even get bioshock infinite.

i push the limits of my alienware with hawken , company of heroes 2 , crysis 3 and waiting THE game, star citizen. (24th of august the hangar module gets released and hopefully by christmass we will have the alpha version.)



ps. as you remember from day one of me installing my 2x 7970s and ocing my cpu i had a psu shutdown only once although my cards were running at 1.1V
now with nosperatus edited bios at 1V i cant make my psu fail. the cpu oc fails first with a BSOD when over 25x multi (4 cores)

by the way i read the new m17 review....

Alienware 17: Performance

not Impressed neither by the design or the 3d performance . ofcourse Its a one gpu machine and gets almost 7000points
in 3dmark 11' but we have a 5 year old cpu and hull and manage to get 10.000+ points In 3dmark 11'

nospheratu. did we get a final verdict that 780m cant run on our systems?

 

Not that I'm aware of. Last I read flingin was still having problems with the white screen after the windows logo.

edit: It seems he's got it working albeit with some issues. He has to use an external screen. He's still working on it though.

edit 2: So my PSU shipment is officially lost Good news is the seller has shipped another out and tracking seems to be working quite well this time around. Also completed the Clevo vBIOS pack and uploaded it here.

 

So what's up about that clevo vbios pack?
I'm surprised it took flingin so long to use an external display, that suggestion was made weeks ago. Maybe there's something i'm missing (to be fair i'm not monitoring the thread lol).
Too bad about your PSU mate. Mine is supposed to arrive next week along with my heatsink. Can't wait.
I'll be modding it as a "travel" PSU and will probably convert my current 2x240 into a 480W supply for OC purposes Gotta break some records.
I'm pissed off about the new 3dmark demo not allowing individual tests. If i try to run the first 2 tests I'll get overheating shutdowns (too many thousands of frames per second, lol).

 

any improvements in the vbios other than the voltages? and whats overdrive?

btw awesome work!!

 

I assume Overdrive refers to the ATI overdrive feature. Oh, I'm using svl's modded 022 .075 vbioses now, too. Having constant access to clock control feels good with all those afterburner troubles.

 

Any of you guys getting random crashes when running evga precision or msi afterburner? Sometimes when I open evga precision it freezes the whole system. Its been doing that when I still had my 260m installed. Only on my 260m, it crashes every time. Any ideas?

Sent from my DROID RAZR using Tapatalk 4

 

There's a way to run them without freezing your whole system, but I never got the hang of it. In the end I just use the Overdrive-enabled 022 vbios. I don't know what's wrong with AF or Precision.

 

@ Sangemaru, is the 022 vbios you are currently using still the 0.975V undervolted version? I wasn't aware overdrive was enable...that's awesome. Is this the one from SVL7s TI forum vbios pack? I have second card on the way which should arrive this coming week. I just would like to clarify.

@ pinoy_92, I've never been able to run MSI afterburner on my R2. After installation, I can open it once but would consistently lock my system up on future attempts. Overclocking has only been achievable by installing MSI afterburner in conjunction with "-xcl" method to unlock overdrive in CCC. Even then, I find myself having to install and perform the registry changes in specific sequences in order to get it to work.

 

@batfuse: Yes, mine is the 022 .0975 modified to enable Overdrive by svl. It's not on the main page afaik, it's later in the same thread. There's a 1.05V and a 0.975V version posted. Page is here: http://forum.techinferno.com/general-notebook-discussions/1736-amd-7970m-modified-vbios-48.html
Afterburner is very finicky on our machines. Rarely ever works. That vbios is really useful.

 

Thanks. Will post once I the card arrives and get everything working.

 

Those vBIOS's are mainly for M17x R2 users with Clevo cards. Actually anyone with a Non-Enduro machine running the Clevo card can use it. Yeah 3Dmark is really annoying especially since you have to sit through those annoying "demos". I wouldn't mind if it ran the 3 separate tests only even though I only really want the Firestrike result but those demo's are just torture.

Overdrive is AMD's official overclocking functionality that built into Catalyst Control Center under the Performance tab. As sangemaru mentioned its much easier to overclock with Overdrive than using Afterburner's -xcl switch and jumping through loops to get it not to lock up your system.

There's nothing changed apart from voltages and Overdrive. The Forced vBIOS's have multiple voltage entries changed to "force" the card to use a particular voltage irrespective of what method it uses to determine what voltage state it should be in. As our cards only really have two performance states ie. 2D and 3D the other voltage entries aren't used to my knowledge. I've tried to play flash videos on YouTube to get the another state to trigger which possibly uses one of the other voltage values after activating an intermediate performance state, but the card remains in 2D state so its safe to say that those entries aren't really being used.

I've tried AlienHacks test vBIOS as well and while it does work and seems to fix the different default voltages on my cards, there's no sensor readings except for core and mem clocks. Without temp and vddc readings its a no go for my Clevo cards. Dell cards might fair better and since he has auto fans, full HDMi and DP I think it's worth a try.

 

I briefly skimmed that post and didn't notice that it was mostly for clevo cards. So, will flashing that vbios to a dell 7970m brick the card?

 

Not reading the entire thing here, but have u guys concidered the 330 upgrade with the fix from imsolidstate? Just to avoid having to bring 2 PSU's?
Going to get a new 330 and have an old 240 and will try to do the old mod, and the extra new mod for the 330 to work above 240.

 

No it won't brick the card. Though I would stick to Dell vBIOS for Dell cards and vice versa. Quoting Bioshock Infinite... "the universe does not like its peas mixed with its porridge"

Yeah it has been discussed here. Read a bit from there I think I mentioned how to create an adapter so you don't have to open the 330W. Thanks for the thought though, its appreciated.

In other news, the 330W wasn't lost It's been sitting at my Post Office from the 25th of July Going to fetch it after work today. Weird thing is USPS still hasn't updated the tracking and our local international tracking system doesn't reflect the package either... yet I received a notification in my post box ( which I almost never check ) at home to fetch a parcel from the local Post Office. It can only be the PSU since I haven't ordered anything else... and the seller has already shipped another

 

@Nospheratu: Wait for it ....

...
..
.

...

MWAHAHHAHAHAHAHAHAHHAHAHAHAHhaahhAHHAahhaHHAhha. Typical post office behavior. Were they a private enterprise, they would be both bankrupt and out of business.

 

Lmao!! Yeah now the dilemma of what to do with the 2nd PSU that incoming...

Picked it up now and this thing is quite large. Double the size of the 240W height wise... cant wait to start modding

 

You could mod it and re-sell it to an interested party Or go for the 660W mod

 

Haha I thought about the 660W mod as well But it would be pointless really, our XM's would burn up before drawing anywhere near that kind of power

I need to contact the seller and find out what they want to do. They provided nothing short of excellent customer service sending out another without delay and I can't in good conscience keep quiet about it. I also don't want to give S.A a bad reputation and shoot myself in the foot when sellers stop shipping here.

I still need a few parts for the adapter though before I crack open the 240W. I figure since I'm making the adapter I might as well make it a right angle connector to reduce strain on the mobo connection. I was thinking getting something like this and then modding that into what I need. It has the center ID pin, and the measurements of the connector tip line up with ours but I'm not sure whether it will handle 300W as the cable is most likely thinner than our 240W's.

 

You could take a look at a mod cable i built for my 150w adapter

http://forum.notebookreview.com/alienware-m15x/727220-psu-modification-cable.html

 

You could take a look at a mod cable i built for my 150w adapter

http://forum.notebookreview.com/alienware-m15x/727220-psu-modification-cable.html

 

Thanks for that @NiveQ Awesome work!

edit: Your diagram is exactly what I had in mind for my adapter except the green ID line coming from the 240W in your diagram will be grounded for my purposes (the 330W won't deliver more than 240W otherwise).

 

Hey guys just a word of warning!

"Current kills" so be xtra careful with those psu mods"

 

I can't wait for my own PSU to arrive this week You're lucky Nospheratu, your PSU can mostly handle full gaming with Xfire so far I've had to play Bioshock: Infinite and Mass Effect 2 in Windowed mode until now...

EDIT: So my PSU has arrived, heatsink not yet. I'll see if I can drop by the office to pick it up today.
EDIT2: On second thought, let's enjoy the perks of modern economics. I'll just have it shipped directly to my door

 

Exciting times!!! Cant wait to see what you can achieve with that combo!

Yeah I think I've got a good PSU, even with its continuous buzzing it still works well. It's a pity I have to crack it open for the ID chip

 

@nospheratu

is it better to use the vbios with stock voltage or the vbios with the over voltage? i only have one card so far, so im not sure which one would benefit me more.

 

@pinoy_92: There's absolutely no point in using even the stock voltage, let alone the overvolted vbios, unless you plan to overclock. Even stock voltage may yield quite a bit of OC headroom, with overvolt taking it further. However, unless you're an overclocker at heart, you'd be best served by finding the lowest possible stable voltage vbios, to minimize heat and wear and maximize the card's lifespan.

 

well i flashed the stock voltage vbios. i dont plan on overclocking. mainly because my gpu is already hitting temperatures on the low 90s. however, i keep getting crashes on games that are gpu heavy. extended bf3 sessions, few minutes of crysis 3, and some far cry 3, all of which usually ended with a crash. while games like dragon age 2, it rarely crashed. any ideas why? kind of off topic. sorry.

 

Then you should try for the moment and flash one of the 1V or 0.975V vbioses in order to lower heat. You honestly should not be running that card if you're getting temperatures like that. Either your heatsink is defective or you did not obtain proper card-to-heatsink contact and you should really re-check your card installation. I consider my load temps in heavy games of ~82C to be on the very high end. These cards cause problems over 85C in these machines, which should not be much of an issue because if you install it properly you should not go higher than 82-83C.