M17x R2 7970m CrossfireX Ultimate Installation and Tweaking Guide

Okay some feedback... luckily its good

Re-installed Windows. Installed 13.4. Booted TR. Ran benchmark. FPS looking normal at 59 average. Alt-Tab. BSOD

Uninstalled 13.4. Ran AMD Cleanup Utility. Installed 13.6 Beta2. Ran TR benchmark.. 59FPS average. Alt-Tab. No BSOD!!

Just to confirm 13.4 was the problem... uninstalled 13.6 Beta2. Ran AMD Cleanup Utility. Installed 13.4. Ran TR benchmark 59FPS again. Alt-Tab. No BSOD

I think 13.6 Beta 2 installed something else that I'm not quite sure of... I'm leaning towards the MS C++ Distributable (2012) but I'm not sure. Either way I'm happy there's no BSOD's atm when Alt-Tabbing.

@AlienHack I did try to use TRIXX previously and I kept getting flickering on the screen so I stopped using it.

@pino7_92 I'll check the load on the GPU's and if theres any throttling once I get this Windows installation sorted. Check previous pages for some benches and FPS numbers in games.

@tron2^ I'll have a look at that when I repaste. I'll install one card and check the voltage before installing the other. Hopefully I can get to it today.

 

Very interesting.
Does your hdmi work? Do your fans work on auto?

 

With dell vBios HDMI should work (mine does) but fan remains stuck at around 2000 rpm (unnoticeable).
With RJTech vBios HDMI should lose functionality (probably) but fan control is enabled.

Optimal solution so far is to flash Dell vBios and use HWINFO on autostart as fan control with custom profiles set. Just make sure you open the HWINFO window once after system starts to actually enable the fan control. I've noticed if I don't do that it forgets to do so until i actually open the fan control window within HWINFO.

@Nospheratu: using WHQL drivers, naughty naughty I swear, for AMD, the beta drivers are always the good and stable ones, while the whql ones are always buggy lol.
Does Bioshock still BSOD on Alt-Tab?

 

So it looks like we might have a totally stable r2 with 7970cf and just one 240w psu?

Damn... I think I might order the secong 7970m

 

My fans work on auto. I cant test HDMi atm but once I can I'll provide feedback.

Fan work weirdly though... when CPU/GPU's cool enough fan switches off. As soon as CPU hit 53c fans switches on and cools CPU to around 43c and switches off again. Its winter here though so these may be the lowest temps I'm gonna get.

Afaik Dell cards have proper working HDMi, my Clevo has HDMi video but no audio. Dell cards cannot gain auto fans by flashing to a Clevo vBIOS. Clevo cards HDMi doesnt work at all with a Dell vBIOS but fans still work on auto.

Lol yeah I stopped using beta's some time ago. I still had the 13.5 beta though which I used to test the BSOD issue but didn't have any luck.

I got another BSOD Alt-tabbing out of TR again, but tried it a couple more times and couldn't get it again Lol... cant believe I'm trying to get another BSOD

No BSOD on Bioshock now. Also ran the "built in" benchmark (Its in the exe's folder) and frames are looking decent for the ultra setting at native res. Without looking the numbers and just sitting through the benchmark it appears smooth. I'll post the results up soon.

 

I still have to test the throttling issue so everything might not be quite perfect but its looking mighty fine so far

 

Regarding Alt-Tab, BSOD's when using Crossfire is a common issue and has been for years now, on the desktop scene as well as mobile. Generally (and especially with the machines with RGBLED screens) we should be using Fullscreen Windowed mode all the time. There's many games that aren't properly optimized and don't handle the resource re-allocation when alt-tabbing well at all.
I'm curious about throttling and possible CPU overclocks.

Btw, you totally stole my thunder I still don't have my crossfire up and running yet )

 

Haha don't worry bro you'll be up and running before you know it... we can always shine together man

 

Well done Nospheratu
I just received my Alienware M17xr2 today. Loaded windows 7 on it and now typing this message on it. Quite a beast and have a 7970m on it's way original is 4870m. woooot
Would you guys install a 7970m on the Alienware m17xr2 or a 680m as I have two cards coming. Have to think carefully and what to put on the Clevo p170HM as well... decisions decisions

Anyway this now holds great promise for all those wanting to crossfire on r2 and thought the psu could not do it.

 

@tron2^: single 680m is an option for heavy OC and SLI availability, in case 7970m crossfire isn't an option for some reason.. however I expect 680m SLI to be near-impossible to achieve. I expect using HyperSLI in software mode there might be a chance, but I wouldn't hold my breath on that.

 

Exactly what sangumaru said. Our machines can't do SLi for cards above the 285M and HyperSLi is a shot in the dark at best.

I think I can say with fair certainty that its the MS C++ Redistributable (2012) x86 and x64 thats included with the Cat 13.6 beta's that's solving my BSOD problem. I installed my commonly used apps and tried TR again and I received BSOD's again on alt-tab. Re-installed the 13.6's without uninstalling them and the installer said the C++ distributables were not installed (Even though Control Panel Programs says it is). Drivers and other AMD add-ons in the AMD installer confirmed that the same versions were installed... just not the C++ Redistributables. Suffice to say, I completed the install, rebooted and no more BSOD

Hopefully this can be of some use to someone who stumbles across this thread looking for answers to a simular problem.

edit: C++ redistributable solving problems is nonsense. it might help matters a little but see my next post.

 

Interesting reply.
The psu on the m17xr2 is rated at 240w the psu on the clevo is at 220W. One would gather that the 240w would be better suited for the higher o/c? so I should put 680m into alienware? I had originally thought to put the 680m into the clevo and o/c the crap out of it and put the 7970m inbound into this machine with possible crossfire as nospheratu has done but I will wait on that front.

I just wonder which will give higher 3dmark'11 the clevo with the 680m or the alienware with the 7970m? Time will soon reveal the results....

 

The higher overall score will be on the clevo due to the MUCH more powerful CPU
But the differences in graphics score will be negligible either way.

 

just in benches though....
in reality and everyday game use they will be the same.

 

guys can i ask you something else? the m17x r2 battery how many cells is it?


UPDATE: nevermind i found out, 9 cell 85Wh type F310J

i wanted to check wether there is a replacement on ebay with more cells

 

It's an 85 WH / 9 cell battery

 

Okay so my theory on the C++ redistributable is nonsense

I'm still getting bluescreens when an application looses fullscreen due to a crash or alt-tabbing or in the case of 3dmark11 between the tests (mostly after the physics test). Its totally random now as compared to before. I'll still keep looking for a fix though. I've tried disabling ULPS and still its not 100%

Here are some Bioshock and TR benches comparing stock CPU vs CPU OC at 24/24/24/24 with 80/72 TDP/TDC. Both games are run at native res on the Ultra settings.

CPU Stock


CPU OC


CPU Stock


CPU OC

 

CAP profiles reinstalled after installing windows So you never had a power trip?

 

Lol yeah CAPs installed this time around

I've only had BSODs so I'm not sure if this is my machines version of the power trip And the BSOD's are only when loosing Fullscreen. Another thing that might be helping me along is that this vBIOS is a Clevo version and it isn't modded. Its default volts for these cards are 1.000v but somehow my first card is running at 0.975v. Also, its pretty cold here now, its about 13c ambient which maybe helping the PSU stay stable.

edit: In TR benchmark primary card utilisation is 98% and secondary card is 75%. In Bioshock Bench both cards are at 99% utilisation. I'm using HWinfo to view these results so they aren't really accurate. I'll run a GPU-Z log for more accuracy. The thing is, this can't be used as any indication for another game as all games are coded differently.

 

You all realize what these results mean... the CPU MIGHT be a bottleneck in some games BUT it clearly depends on the game.
In Bioshock clearly the faster cpu you have the game eats it up and gives more FPS
BUT in tomb raider the cpu doesnt play as much a big role.

lets hope that most games wont be cpu intensive...

 

ALSO now that i think about it...

ITS not that the 920xm isnt capable of giving great results when oced! i mean 24 multi (3.2GHZ) is pretty crazy for 4 cores , and if setup correctly (and you have a great chip too) the 920xm can have 4 cores at 25-26 and 2 cores at 27 , i personally run benches at 28-27-26-25 stable (ofcourse not with 7970 cf)

Nospheratu can you use something like 28(or 27 if you cpu cant go that high) - 27-26-24? and rerun the bioshock benches? i want to know wether Bioshock uses all 4 cores of the 920 (thus running all the cores at 3.2 ) or it uses 1-3 thus running them at 3.4Ghz to 3.7 Ghz...


i mean we say that the 920xm might be a bottleneck but if that chip can run all cores at 3.33 Ghz (25multi) or maybe even higher then i dont know what other mobile cpu (i7 or other) can surpass it... i mean yeah 3rd and 4th generation i7s are 6-20% faster than our chips in the same GHZ but there are problems ocing them with throttlestop and also the 4th gen have stability problems, many of them cant even run at the factory suposed multies... so forget about running all cores of a intel 47xx or 49xx series over 3-3.2 GHZ , so maybe we are still gonna be on top in terms of totall cpu power all is not lost brothers!!!

 

Exactly. I was for a long time comparing performance and clock numbers in different machines and the sacrifice required to actually overtake the 920/940xm was too big to justify.

 

Yes ThrottleStop turns our CPU's into something else really. I did some testing recently regarding our multi's in relation to how many threads an application uses.

Effectively only the 4th multi is worth paying attention to since that multi is used for 8 threads or more and in a Windows environment there are numerous threads running all over the place so the CPU is never going to be running one/two or three threads unless you physically go and set affinity for the application you're running which isn't practical .

It easy to set your multi high and think you're getting higher performance but that's not the case. I can only sustain 25x across all cores with a minimum TDP of 90w. To sustain a higher multi I would have to increase the TDP and then the temps are out of control. As it stands I can run benches at 25x on all cores but I would never game or video encode with those multi's as the temps reach ridiculous levels.

If you want to read about it in more detail and unclewebbs explanation see here.

 

Don't need to overcomplicate things. What I do is find the highest stable and temp-safe 4-core multi and then increase each other multiplier by a stepping of two.

So my multis at the moment are 27/25/23/21. This keeps my machine under 90C and still quite formidable.
Regarding load balancing across the cores, yes, the situation you mentioned is normal and expected but actually quite irrelevant in the long run, because the CPU will load-balance across multiple cores to maintain temperatures and once you eliminate prime, TSbench or other apps that WILL single-handedly stress one or more cores 100%, leaving just background tasks and games to stress the CPU, what happens is that there will be activity spikes with various cores gearing up for the workload while the other cores stay chill processing less important threads.
It all comes down to how a game is optimized, but you are very very RARELY going to see any game at all pushing load anywhere near 100% across 2 cores, let alone 4.

So if the game can't generate 100% thread workload across all cores, what should happen (and appears to happen in my case) is that the CPU automatically starts distributing work across the most efficient configuration and does so on the fly. The CPU dynamically turns cores on-and-off for fractions of a second in order to push the most efficient core workload configuration within the TDP and TDC limits depending on what load the applications put on it.

So a moderately well-optimized game that is CPU-bound may STILL show up logging frequencies way above those set by the 4-core multiplier, because the 4-core 100% load may only be for like 50ms, while for another 100ms the CPU needs to push 1 core an extra 600mhz to finish a complicated particle effect and push 2 other cores to calculate some collisions or whatever.

Bottom line is this: You're interested in finding the HIGHEST stable and temperature-safe 4-core multi. Once you found it, increment the 1-core multi to the highest stable multiplier. Usually the temps won't matter for a single core, because the CPU will automatically redistribute load to another core when it gets too hot.

After you've done that, adjust the 2-core and 3-core multipliers accordingly, but keep in mind that 2 or 3 cores generate far more heat than a single one. The final important thing to keep in mind is that you must strike a balance where you don't keep clocks near their maximum safe thermal limits, because if 4-cores at 3.33GHz 100% load generate 95C for example, and a single-core clock-speed of let's say 3.8GHz goes above that temperature, when the CPU needs to switch core load to cool that core down, providing you don't have enough thermal headroom in your 2-core, 3-core and 4-core multis, the core may be unable to cool and you'll be flummoxed to find that even though 4-core 3.33GHz workload doesn't push over 95C in even the most demanding of stress tests, you'll somehow reach maximums of 100C for every single core while gaming.

Now I'm not particularly sure exactly how much sense what I just wrote makes But I want the takeaway to be this:

Raise the highest multis you can for 4-cores and 1-core, while trying to play it chill for 2 and 3-cores. There's a thermal balancing act in there that will allow you to make the most of your CPU.

 

Yes, I agree... thats helluva complicated

I do agree with your method though, you're trying to get maximum performance out of your chip. You will get higher overall frequencies if you increase the 1st-3rd multi's as when the CPU is loaded it will try its best to reach those multi's depending on the load of the CPU.

From my testing though it never reaches the multi's that you set. It will go above the 4th multi though so yes performance should be better.

When I set my multi's to 25/24/23/22 with a TDP of 90/82 and run a 3 threaded test you will see your overall multi at around 22.XX... not 23. Same goes for a 1 or 2 threaded test. Run a 1 threaded TSBench test and see if your CPU is hitting the 25x multi... it wont. 90 TDP is more than sufficient for those multi's so that's not a bottleneck.

I guess its personal preference really. I like to know that when loaded all cores and virtual cores are running at a certain speed consistantly rather than trying to reach the higher multiplier and bouncing around.

 

Off-topic as hell, but I think we're about 30 posts away from making the "Most active threads" list

 

Once more, what I mentioned above is not for benching or stress-test scenarios But set your multis to 27/25/23/22 with a TDP of 90 and a TDC of 62-65 (trust me you shouldn't need more than this) and do some gaming and log usage with hwinfo or something You might be in for a surprise.

By the way, I currently run something like 70W TDP / 55A TDC with multis 27/25/23/21, and it's cool as a cucumber and frequencies of 3.5GHz and above are common. Under extended load the cpu will clock down, but for short activity spikes (which is the more likely scenario), I'm running smooooooooooth.

I'm going to experiment with giving the cpu +75mV (since I already run stable 28x at stock anyway) and pushing the 1-core and 2-core multis around 30x while also giving it 5% OC, and keeping the TDP around 70 and TDC around 60.

I'm REALLY curious what's gonna happen I'll leave the 3-core multi low (and may even keep the 2-core multi a little more discrete) and watch the temps. I'm figuring that with 21x on the 4-core I may have enough thermal headroom especially at that low TDP to be able to consistently run one core at 4.2GHz Think that's too hopeful?

 

Will do I like those multi's actually, its reasonable for everyday use. Hopefully I get some good results.

edit: I wouldn't do the extreme clocks personally Especially the voltage increase. But I'm a bit conservative with my clocks and if there weren't people like you pushing the boundaries I wouldn't have two 7970M's in my machine right now

 

i personally use everyday (and in gaming) these multies 27/26/25/25 . The laptop is stable even after playing BF3 for hours. The cpu temps are 78-95 (my 4 cores have wildly different temps, i will do a repaste)

90W TDP / 75A TDC


I dont think that i need something faster for everyday. I mean come on!! 4 cores (8 threads) running at 3.3Ghz to 3.6 is high enough!!! for whatever you throw at them!!! even my desktop i5 @ 2.8 (3.3 turbo on one core) is superfine for my needs

 

Wow, those are some nice temps for those clocks. Very impressive, imo. And yes, my temps fluctuate wildly as well.
I plan to buy a second heatsink and try and lap it, to get to equalize the temps, and see if I can do a retention mod to really put the pressure on.
This entire machine has so much hidden potential that it's hilarious.

 

ohhhh guys.... dont push me.... :thumbsup:

in 2001-2004 i was in the greek national oc team (well there wasnt technically one but i was in the team that did the extreme cooling (from water back then, to chiller to liquid nitrogen , in Greece and later took some international prizes) . We breaked several records withwhatever we had back then , p4 prescots (mine was a 2.4 one that was running 24 hours at 4.2 with a waterchiller). My 9800 pro ATI was the fastest in the world back then (in 3d mark)

SO dont give me ideas to get out of the closet my peltiers abnd my chillers... i always wanted a peltier laptop. :yes:

 

Do eeeeeeeeeeeeeeeeeeet

EDIT: Actually I took a look over the theory of thermoelectric cooling and from what I'm seeing it has only about 1/3rd of the efficiency of vapor compression techniques?

How would that compare to our current heatpipe system? Any ideas?

 

look the peltiers werent exacty efficient BUT they did their work. they would surely drop the temp on the cpu by 20-30 C maybe even more. But you would need an extra 110W power brick to feed the peltier.

THE GREAT thing about them is that they can fit anywhere, i mean i could just put a peltier in my r2 right now. it would sit between my cpu core and our heatsink exactly like a thermal pad. But two small cables would come out of it. When you power them..... nice things hapen!!!



liquid nitrogen is a no go for the laptop (for mounting and using reasons, only maybe just for extreme benches.

And watercooling would be something to see!! (although its doable i am sure, i havent seen another watercooled laptop)


BUT i am sure that there are surely many extreme mods we can do for our laptops. I mean i just thought of one. Taking out the current fans and changing them with ones that have maximum 6000 rpm not just 4000 like our current ones. That surely will drop temps by 5-10 C (although produce very loud noises)

also i just has another idea. we could handmake (i have a friend who works with metals) an external heatsink it might even be detachable. whenever you have your laptop on your office it would "attach" with this external piece of metal (it might even have fans on it) , wich would transfer heat out of the laptop and into the enviroment

 

Well, if you say peltier is that much more efficient compared to heatpipes, that would be great. 20-30C is gigantic. I wouldn't mind doing the chip mod to place the PSU ID Chip directly on the mobo, and use an aftermarket 300-400W PSU to power the machine and peltier cooling, so long as it can be done in a non-invasive manner. I don't exactly want to butcher my machine

Watercooling has been done in the m18x R1 afaik, as a demo by a cooling product company. Not sold, though.

Regarding fans, what we need is not necessarily more rotations, but a higher cfm. However, seeing as how our fans are 5V versions, I've checked the net for better models that might fit the bill, but it appears they're surprisingly high-quality ones for 5V. Still, I wouldn't exactly like a turbine in my machine.

Regarding the external heatsink, in order for it to be useful, it would have to:

1. be well connected to the heat transmission method of choice (either heatpipes, the fins of the current heatsink or the Peltier dissipation module).
2. it would have to represent the bottleneck in the machine (e.g.: the heatpipes would be able to cool the machine faster if the heatsink could dissipate it faster). If dissipation is not the bottleneck but transfer is, then the external heatsink would be useless.

As far as I can tell, the supreme bottleneck regarding cooling in the machine is the heatsink itself, specifically at the point of contact with the core. That's the reason people focus so much on lapping, pressure, retention mods. The heatsink and fans can already dissipate a lot of heat. Sure, it can be improved, but the real gains will be found at core-contact-level.

 

the connection of an external heatsink is the point. Because if you have an external heatsink you can play with its size and the fans it has on so you can maximize dissipation, BUT the point is how is it going to pick the heat from your current heatsink?

also as you mentioned contact and heat transfer between the core and our current heatsink is very important BUT its not the most importand.
In thermodynamic equation the things that mater are the materials , the contact area, the heat transmited (temperature of heat source) and the temperature of the enviroment
Especially on the last two numbers, what is important is the difference of the temperature between the heatsink and the enviroment. The biger the difference the quicker the heat is transfered from the heatsink to the air (or from the core to the heat sink , its the same idea)

SO , its logical that if you run your laptop outside at winter at -10 C then the heatsink and cpu will be greatly cooled and so you will see like 60C max cpu temp

BUT there is another way we can do this . if we manage to "attach" something (like two heatpipes) to our current heatsinks (making 2 holes on the plastic casing of the laptop maybe?) and on the other side of these heatsinks we attach a piece of coper that is submerged in water then beleive me the temps of our inlaptop heatsink and of our cpus will drop. how much you ask? well i cant say for sure it really depends BUT it could be as high as 10 C i beleive

NOW if this piece of outside copper is submerged in COOLED water then the temp drop in cpu will surely be 20C at least

BUT the reason i hesitate to try these stuff is that they all need our laptop chasis to be "edited" , either small holes for the heatpipes or cuting the grills to fit the external heatsinks fins... So i am not sure i want to be so intrusive . Thats why i think either peltier would work nice (as it just fits as is and just has 2 small cables that will go outside the laptop.

 

I have to admit, modifying the chassis is something I'm not particularly interested in, not to mention that in order to make the addition of external heatpipes useful, the interface between them and our current heatpipes would have to be soldered for optimal thermal conductivity, otherwise we would suffer the same issue as when using a heatsink without thermal interface.
I have to say, the most interesting out of the mods you mentioned remains the Peltier option.

What exactly is it about it that is better than traditional heatpipes? Does it conduct heat faster?

 

i dont think the thermoelectric cooling would work on a laptop, because one side of it would be very very cold while the other would be very very hot. which means ur gonna need to fit very conductive material from a small gap between the keyboard to the outside of the laptop. and ur gonna need a beefy heatsink. i dont think thats plausible.

 

We must learn the cooling capacity of the current heatsink and fan we have. If its for example capable of cooling 70 w then we can use a peltier of up to 70w with no prob. The problem is if you use a peltier biger in w than what your heat system can handle. Then you would have heat excess with no way to expel it and everything would melt (well not in reality. The cpu temp would just skyrocket and shut down

 

Now that i think of it and i try to remember whats right.... MAYBE the cooling capacity of our heatsink should be able to cool the produced cpu heat together with the peltier produced heat.... hmmmm... its been over 10 years that i used peltiers... i must refresh my intel...

 

So I played Crysis 2 now since .Cameron. reported it shutting down for him constantly on his 330W mod (I purchased the cards from him). Stock Dell settings on the CPU which is 24/24/24/24 with 62/62 TDP/TDC and stock GPU's.

Settings were 1920x1200, Ultra system setting with DirectX11 and High resolution textures. I played from the beginning to the next video scene after you climb up into the building below the subway and the screen fades to black while walking up the stairs, the first "chapter" if you will. Framerate was good with a minimum of around 38fps and average of about 58fps. I was using RadeonPro and glancing at the FPS display so its not really accurate. PSU became quite hot. Not to the point where you cant hold your hand against it though. And I'm happy to say it didn't shutdown.

I'm still getting BSOD's though when alt-tabbing and when Bioshock crashes on 13.6 beta2 so I reverted to 13.1 and it seems to be working well for me so far. Bioshock just crashes harmlessly to the desktop.

 

so bioshock apart (i never likes the series anyways) , you have a totally stable R2 with 7970cf??

damn that sounds cool. Is there anyway to try crysis 3 again to see the max cpu oc you can get it stable at?

with 24/24/24/24 multies what temp does the 920xm get to when playing crysis 3?


ALSO! a question. People say that the 7970m when slightly oced is similar in performance as a 7870 desktop , so 7970m cf should be like 7870 crossfired

so... acording to this : http://www.youtube.com/watch?v=7BeKb9M5zV0

if you look at the benches the 7870 cf is always slightly better than a 6990! and almost double the power of a 6970 (which i own on my second desktop) , thats almost as powerfull as a desktop 680gtx!!!

 

I checked max temps after Crysis 2 and one core hit 80c with the others in the high 70c region. This was with TDP of 62W though since I left the CPU at Dell stock settings. I haven't installed Crysis 3 yet after the Windows re-install. I'll get that done today.

I don't want to stress my system too much, all I wanted is smooth game play at everyday sustainable clocks without damaging my hardware. I love this machine, I honestly don't know what other laptop I could replace this with should anything go wrong. Good secondhand R2's with my spec and condition are rare now and parts seem to be mostly secondhand. I was actually planning to underclock these cards to 700/1000 so the fact that they are working at default frequencies is a massive bonus

Theoretically it seems to be true but I have a feeling if you run a desktop machine with a GTX 680 or 7870 Xfire next to this system the desktop cards will put the smack down on these 7970M's

 

Assuming perfect scaling, the crossfire should be anywhere between 15-30% faster than the GTX-680. I'm guessing that with an average of 50% scaling, they would be pretty much equal. Anything above 50% scaling should start overtaking the desktop GTX-680 stock.
And if we could actually OC the cards, we would devastate the GTX-680 and at least equal the 7870 xfire.
A worthy thought in the event that we find a good 300W+ PSU to use as alternate with the on-mobo ID chip mod.

EDIT: Also, don't forget to setup RadeonPro profiles for the games in order to eliminate microstuttering Read this for more info.

With proper optimization, AMD manages to beat nvidia in Crossfire vs SLI frame latency and microstuttering. Only the drivers need to catch up.

God, I wish there was an nVidia fanboy in this thread

 

what? 70-80C??? mine at 2726/25/25 hit 80-95C ! Oh i forgot you said its winter there! here we have 38C enviroment!

ALSO guys i need your help.
The seller i bought the card tried to flash it with the bios you told me from here Downloads | Tech|Inferno
he used the normal bios file , no oc no underclock

BUT he msged me that now the card wont show anything on the screen.
I asked him what laptop he uses it in cause maybe this video works only in the m17x r2??

the card he flashed was suposed to be Alienware AMD HD DELL 7970M 2GB GDDR5 M15x, M17x ,M18x Dell P/N # 747M2
arent these the ones that work eith our bios?

 

Pretty much any vbios except the Clevo HM version should work with any card except the ones with .022 vbios.
It pretty much depends what machine he has. The chances that he actually damaged the card are infinitesimal.
He should be able to blind flash it back to a dell vbios and it would display for him. You shouldn't really need to have the card pre-flashed anyway. Even if it used the wrong vbios (except 022), as a slave card it would still be detected and display in your machine just fine, and you could flash it yourself to anything you wanted.

 

Phe... hate you guys. Your temps are SOOOOOO much better than mine...

 

i told him to flash cause i have sold my cards (2x 5870) and now the R2 is without any so i cant flash it (except with a blind flash which i dont wanna try)


SO if he had a dell card with the 022 bios what does that mean? that this card was a uefi card? and that it wont accept our bioses?

 

Yes, .022 vbios means UEFI card (and we don't have UEFI option in bios afaik).

However, since he advertised the card as working on m15x it shouldn't be an UEFI card, since m15x doesn't support them either (i think).

Still, I don't think you should worry overly much about it. If you're worried you can ask him to flash the Dell stock vbios listed here. It should probably give him output, and also works on your machine.

Another thing is, you can tell him to use the VGA output port on his laptop. The card should display using it.

 

which bios do you mean? the link you gave me is the same .zip file as the one i gave you. or not?

 

Yeah sangemaru's right. M17x R2 and the M15x doesn't support UEFI either. It all comes down to what machine is he testing it on. But if the flash completed without any errors I also think its good to go in your machine.