Going to test clevo 980m in alienware m17x R4, m18x R2, AW17

[svl7]

Loss of performance is a relative thing, you're talking about a very specific task, or rather benchmark. Keep in mind that for other CPU benchmarks Win7 performs worse than XP, while for others Win7 is faster. Generally Win 8 and 8.1 is a superior OS to Win7 in many ways.

As for the legacy issue - I'm looking into this. Currently having trouble with my AW17 though and I have no clue what the cause is, really frustrating.

[mw86]

Loss of performance is a relative thing, you're talking about a very specific task, or rather benchmark. Keep in mind that for other CPU benchmarks Win7 performs worse than XP, while for others Win7 is faster. Generally Win 8 and 8.1 is a superior OS to Win7 in many ways.

As for the legacy issue - I'm looking into this. Currently having trouble with my AW17 though and I have no clue what the cause is, really frustrating.

What is not working correctly with your Alien 17 svl7?

[Player2]

Has anyone had a 980m running in there alienware 17 with optimus for any length of time to know if it works proper. Like running games and doing different bechmarks. I know several people have done it and it works but does it work proper. I am of thinking of going this route because there's no way I'm installing windows 8.

I heard one person that did it but it wouldn't function properly but it did function.

[Paolo11]

Anyone with an AW18 tried with 980m sli?

[Guest]

A lot of people are asking the same question. If I decide to upgrade my M18xR2 to 980M SLI I will test it in the Alienware 18 first to answer that question. There's no reason to suspect it would not work with exactly the same type of manual graphics switching setup.

[Zyzyx]

Hey all, got a Alienware m17x R4, : 32gb ram, i7 3820QM, OS installed on an 840 EVO SSD ------- I've got my hands on a Clevo 980m card, would like to install it.

58 pages on this thread, too many to comb through (did skim through about 20 or so) but can someone post a rundown on how to go about installing a 980?

A list of thee steps would be nice (step 1 : ____, step 2 : ___, ). I'm fairly handy with computers, but I'm no Sheldon, so if the steps could be written simple enough for a mid-to-moderate level computer user, that'd be great!

Assuming I can get some 'how to' steps on installing this, i intend to post pics/walkthrough here on TI

I've come across of number of folk online saying they've successfully installed the 980m onto their m17x r4, just haven't come across any clear-cut step-by-step guides.

imgbox - fast, simple image host

imgbox - fast, simple image host

[Guest]

OK, it looks like if you jack up the CPU voltage higher than necessary you can maybe get a little bit better physics performance from Windows 8. Still falls a bit short of Windows 7, but much closer. This Physics score was with Flex VID set at 110 and a mild GPU overclock with one AC adapter. This makes me wonder what kind of asinine shenanigans Windows 8 is doing behind the scenes to try to reduce power consumption.

NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-3920XM Processor Extreme Edition,Alienware M18xR2

[godfafa]

OK, it looks like if you jack up the CPU voltage higher than necessary you can maybe get a little bit better physics performance from Windows 8. Still falls a bit short of Windows 7, but much closer. This Physics score was with Flex VID set at 110 and a mild GPU overclock with one AC adapter. This makes me wonder what kind of asinine shenanigans Windows 8 is doing behind the scenes to try to reduce power consumption.

NVIDIA GeForce GTX 780M video card benchmark result - Intel Core i7-3920XM Processor Extreme Edition,Alienware M18xR2

Hello hyung nim.

I think if you have done the same thing ie. increase voltage in windows 7, you'd have gotten better physics.

[Robbo]

[MENTION=119]Mr. Fox[/MENTION] and @godfafa, how does raising the voltage result in better scores, isn't performance just based on the clock speeds that are set during your overclocking (e.g. 4.7GHz in Mr Fox's tests). My understanding was that you just needed to supply enough voltage to allow whatever clocks you're running to be stable, ie not throwing out any computational errors.

[mw86]

[MENTION=119]Mr. Fox[/MENTION] and @godfafa, how does raising the voltage result in better scores, isn't performance just based on the clock speeds that are set during your overclocking (e.g. 4.7GHz in Mr Fox's tests). My understanding was that you just needed to supply enough voltage to allow whatever clocks you're running to be stable, ie not throwing out any computational errors.

maybe thats what they mean perhaps to a small unnoticable percent at slightly lower volt there are some computational errors slowing it down and adding the extra volt fixed that for Mr Fox... maybe

i actually did witness that in XTU clocking up a certain volt did give a better score but more heat. This was on Windows 7 but not to say Win8 doesnt do some shenanigans. personally example in firestrike on the combined test i watch my power draw and seems to me Win7 or Win 8 there maybe some kind of power balancing shenanigans going on because i saw higher fps at first in the combined test sometimes and with no change in clocks i see power draw drop and fps drop to a different level through the rest of the run. there may be something in the motherboards themselves on M18x R2 and Alien 18 motherboard.

Robbo and Godfafa is right testing win 7 and 8 on same test same mhz across hardware and same volt would be a true indicator of Win 8 shenanigans.

[Darcoder]

Can someone confirm if i put a 980m in the alienware 17 (60Hz), is it possible to play games? I'm a bit confused with people saying it only works in optimus mode. Does the fn+f5 works for choosing hd4600 or gtx980m? Thanks

[Cloudfire]

To avoid confusion like above, this should be sticky:

There are currently two ways to get GTX 980M to work:

#1: With any type of boot (Legacy or UEFI):

This method only works with Optimus Enabled. If you disable it (FN+F5) you wont be able to boot up the notebook.

Works on both Windows 7&8. This method require to use the INF mod to download drivers.

There are some posts where they said the computer would crash every now and then. So stability is unknown.

Does not require a special vbios.

This method is only useful for Alienware 17 and older M17x models. Does not work with Alienware 18 or M18x due to not having Optimus.

#2: With UEFI enabled and Legacy disabled

This method only works under Windows 8. Any attempt to use Windows7 will result in a freeze during boot.

Both Optimus and No Optimus works with this method, so you get full normal support like you are used to.

Does not require any special vbios and works both with AW17/M17x and M18x (AW18 is currently untested but should work), single card and SLI.

FN+F5 switching works perfectly.

Like method #1, you will still need to modify INF to be able do download drivers from Nvidia.

This is currently the best method, due to full functionality and perfect stability. The only thing thats missing which is being worked on at the moment by svl, is support for Windows 7.

[Guest]

Hello hyung nim.

I think if you have done the same thing ie. increase voltage in windows 7, you'd have gotten better physics.

Yes, but there is a point of diminishing returns when using too much voltage actually degrades performance and induces throttling and/or overheating, or just good old fashioned instability. I have my overclock dialed in for Windows 7 but CPU performance using the same identical settings is diminished when booting into Windows 8. There is also a load throttling with Windows 8 that does not happen with Windows 7 and I cannot figure out why Windows 8 does that. It always has though... Windows 8, 8.1 and 8.1.1 and Windows 10 exhibit this undesirable behavior. Increasing the voltage beyond what delivers ideal performance under Windows 7 seems to somewhat minimize the Windows 8 performance drop.

maybe thats what they mean perhaps to a small unnoticable percent at slightly lower volt there are some computational errors slowing it down and adding the extra volt fixed that for Mr Fox... maybe

i actually did witness that in XTU clocking up a certain volt did give a better score but more heat. This was on Windows 7 but not to say Win8 doesnt do some shenanigans. personally example in firestrike on the combined test i watch my power draw and seems to me Win7 or Win 8 there maybe some kind of power balancing shenanigans going on because i saw higher fps at first in the combined test sometimes and with no change in clocks i see power draw drop and fps drop to a different level through the rest of the run. there may be something in the motherboards themselves on M18x R2 and Alien 18 motherboard.

Robbo and Godfafa is right testing win 7 and 8 on same test same mhz across hardware and same volt would be a true indicator of Win 8 shenanigans.

Yes, exactly... same machine, same identical settings, dual booting Windows 7/8.1 proves there are some icky things going on with Windows 8 "optimization" crap.

It doesn't fix it, but using more voltage than necessary under Windows 7 made my results better in Windows 8, closing the gap to some extent. I can't prove it, and don't know what causes Windows 8 to suck in this regard, but my guess is that Windows 8 does something to try to be "efficient" (save energy) and screws things up in the process. This load throttling "adaptive" nonsense is something we expect to find with NVIDIA MXM cards without a vBIOS mod to correct it. The behavior Windows 8 exhibits is similar to that. I am under the impression Micro$haft has intentionally done something similarly retarded, in concept, with Windows 8. By my standards the 2.5% to 5% decrease is a significant hit on performance.

As an FYI, I even use the exact same folder/installation path for ThrottleStop, so my Windows 7 settings are identical when ThrottleStop loads up under Windows 8 as they are under Windows 7. I also have a shared XTU profile folder and applying the same profiles to both OSes, Windows 8 CPU performance is always lower. Unparking the CPU and setting HPET in Windows 8 does help, but not to a meaningful extent.

Another thing we don't know is exactly how much lower the performance is because the Windows 8.X.X RTC error that makes the results unreliable also tends to slightly overstate the results in benchmarks. The drop in performance feels like more than 2.5% to 5% and may actually be more than what it appears to be from looking at inaccurate benchmark results due to the RTC error. I think HWBOT disqualifying some benchmarks run using Windows 8 is for very good reason. 3DMark 11 provides the most extreme example of the performance hit. In some tests it appears to be closer to the same, but it's really hard to say if you can't trust the numbers to be accurate.

It's not only CPU performance. Memory and SSD read/write performance is typically a little better with Windows 7 as well. Windows 8 has basically gone through two Service Packs and now Windows 10 (SP3) is continuing the tradition of lackluster performance. I don't expect Micro$hat will ever fix this... in fact, I doubt they even care. There's essentially zero public pressure for them to fix it, as most Windows 8 users are totally oblivious to these problems.

[Schumsta]

Does not require any special vbios and works both with AW17/M17x and M18x (AW18 is currently untested but should work), single card and SLI.

i asked a second time HIDevolution, cause they have problems with AW18 + 980m. I asked if they made BIOS Changes etc. like in Peter's guidance andI told them that some guys here got it to work with a m18x r2 and thats the answer:

Hello Rene. Yes, I have heard about the M18x R2 working with the SLI 980M. So yes, the problem occurs only with the 18.

-Ted

The Alienware 18 - the BIOS does not recognize them.

[Robbo]

@Mr. Fox, I don't think there are Windows 8 Power Optimisations that you talk of that are fixed by increasing voltage of your CPU, because you'd think if that were the case then increasing the voltage would simply mean that Windows 8 would try & kick in the power optimisations/restrictions even harder. Performance is related to the clock speed, not the voltage, unless you get power throttling which would be worse with increasing the voltage anyway. Did you ever try running the increased voltage level that works so well in Windows 8 on your Windows 7 to see if you also get a performance increase from increased voltage in Windows 7 - it makes me think that the overclock is borderline stable/unstable, throwing out errors & causing slow downs in the program software somehow? (I'm just curious about the overclocking of these new CPUs so that I can do this more sucessfully in the future if I ever buy a system with overclockable CPU).

[Cloudfire]

i asked second time HIDevolution, cause they have problems with AW18 + 980m. I asked if they made BIOS Changes etc. like in Peter's guidance andI told them that some guys here got it to work with a m18x r2 and thats the answer:

The Alienware 18 - the BIOS does not recognize them.

I dont trust them until I or anyone have tested it out.

AW18 have UEFI too. Makes no sense that only M18x works.

[Khenglish]

BIOS recognition and functioning are not the same thing. A BIOS can not recognize a card, but it can work fine anyway. This is the case when a system needs the driver .inf to be modified to install a driver. I highly doubt that the m18x R2 BIOS recognizes the 980m.

[Robbo]

BIOS recognition and functioning are not the same thing. A BIOS can not recognize a card, but it can work fine anyway. This is the case when a system needs the driver .inf to be modified to install a driver. I highly doubt that the m18x R2 BIOS recognizes the 980m.

When you say "recognize", do you mean being able to go into the BIOS screen and see the card listed as "980M"? (My laptop recognises my 670MX even though it's from a later generation if the definition is being able to see it named correctly in the BIOS screen).

EDIT: Actually, I'm mistaken, just checked my BIOS screen and it just says NVidia GFX, so I guess that's what you mean by the BIOS not recognising a card but still working?

[Quark]

When you say "recognize", do you mean being able to go into the BIOS screen and see the card listed as "980M"? (My laptop recognises my 670MX even though it's from a later generation if the definition is being able to see it named correctly in the BIOS screen).

EDIT: Actually, I'm mistaken, just checked my BIOS screen and it just says NVidia GFX, so I guess that's what you mean by the BIOS not recognising a card but still working?

No, that doesn't mean BIOS doesn't recognize it, actually it does. AW BIOS recognize as NVidia GFX 8xx series too, as far I know.

[Khenglish]

When you say "recognize", do you mean being able to go into the BIOS screen and see the card listed as "980M"? (My laptop recognises my 670MX even though it's from a later generation if the definition is being able to see it named correctly in the BIOS screen).

EDIT: Actually, I'm mistaken, just checked my BIOS screen and it just says NVidia GFX, so I guess that's what you mean by the BIOS not recognising a card but still working?

Yup.

[Guest]

@Mr. Fox, I don't think there are Windows 8 Power Optimisations that you talk of that are fixed by increasing voltage of your CPU, because you'd think if that were the case then increasing the voltage would simply mean that Windows 8 would try & kick in the power optimisations/restrictions even harder. Performance is related to the clock speed, not the voltage, unless you get power throttling which would be worse with increasing the voltage anyway. Did you ever try running the increased voltage level that works so well in Windows 8 on your Windows 7 to see if you also get a performance increase from increased voltage in Windows 7 - it makes me think that the overclock is borderline stable/unstable, throwing out errors & causing slow downs in the program software somehow? (I'm just curious about the overclocking of these new CPUs so that I can do this more sucessfully in the future if I ever buy a system with overclockable CPU).

Really? Can you imagine that someone like myself or Johnksss would not have already tried that, LOL? C'mon, I don't know anywhere close to everything, but at least give me a little credit for not being stupid, bro. I've spent hundreds of hours testing to get my CPU overclocks dialed in for benchmarks, otherwise I wouldn't have any kind of status at HWBOT. Windows 7 CPU and GPU settings don't always work as great with Windows 8. It's super easy to test with a dual boot system.

The only thing that has changed now is I am trying harder to get good CPU performance results from Windows 8 to determine if I am going to bother with a 980M SLI upgrade. If I can get equal or better CPU performance with Windows 8, I will ditch Windows 7 and jump on the 980M upgrade bandwagon... but, so far that doesn't seem possible. This isn't my first attempt to coax Windows 8 into working well. It was too much trouble and I gave up before because I don't like wasting my time on a what seems like a hopeless case. So, now I am starting over, trying things that didn't work great for Windows 7 to see what Windows 8 likes best. There was simply no reason for me to care that Windows 8 sucked until now because there was no reason for me to use it. The more I try to force myself to use it for benching, the more I hate Windows 8. For normal everyday use and gaming, Windows 8 is fine. It does 3DMark 2013 better that Windows 7... and, well, so far, that seems to be about the extent of the value of running Windows 8 for benching.

Since you don't have an unlocked CPU, here is a clue for when you get one... Voltage and power (watts) both affects core behavior. More is not always better. Too much or not enough of either one reduces performance and stability. Yes, performance is related to core speed, but voltage and power are the keys to achieving high core speed with stability and no throttling. You cannot separate them like your comments suggest. Simply dumping more voltage on the CPU doesn't (usually) produce good results. For some reason, Windows 8 behaves differently than Windows 7 with increased CPU core voltage.

[Robbo]

Really? Can you imagine that someone like myself or Johnksss would not have already tried that, LOL? C'mon, I don't know anywhere close to everything, but at least give me a little credit for not being stupid, bro. I've spent hundreds of hours testing to get my CPU overclocks dialed in for benchmarks, otherwise I wouldn't have any kind of status at HWBOT. Windows 7 CPU and GPU settings don't always work as great with Windows 8. It's super easy to test with a dual boot system.

The only thing that has changed now is I am trying harder to get good CPU performance results from Windows 8 to determine if I am going to bother with a 980M SLI upgrade. If I can get equal or better CPU performance with Windows 8, I will ditch Windows 7 and jump on the 980M upgrade bandwagon... but, so far that doesn't seem possible. This isn't my first attempt to coax Windows 8 into working well. It was too much trouble and I gave up before because I don't like wasting my time on a what seems like a hopeless case. So, now I am starting over, trying things that didn't work great for Windows 7 to see what Windows 8 likes best. There was simply no reason for me to care that Windows 8 sucked until now because there was no reason for me to use it. The more I try to force myself to use it for benching, the more I hate Windows 8. For normal everyday use and gaming, Windows 8 is fine. It does 3DMark 2013 better that Windows 7... and, well, so far, that seems to be about the extent of the value of running Windows 8 for benching.

Since you don't have an unlocked CPU, here is a clue for when you get one... Voltage and power (watts) both affects core behavior. More is not always better. Too much or not enough of either one reduces performance and stability. Yes, performance is related to core speed, but voltage and power are the keys to achieving high core speed with stability and no throttling. You cannot separate them like your comments suggest. Simply dumping more voltage on the CPU doesn't (usually) produce good results. For some reason, Windows 8 behaves differently than Windows 7 with increased CPU core voltage.

Cheers for the response, as I was saying I'm just trying to understand the quirks of overclocking the newer CPUs (out of interest), I know most of what I highlighted in your post in bold already, the only question I've asked you a couple of times is why you get more performance by upping the voltage while clocks stay the same. I know you've talked about vague hunches about Windows 8 power optimisations being the issue, and I was hypothesising that it's instead a borderline stable/unstable overclock throwing out errors and slowing down the program software. I suppose when it comes down to it, neither of us really knows why you're getting increased performance through added voltage in Windows 8. We can definitely leave it at that, I was just trying to find out if you knew why. Good to discuss it though, hopefully the mods won't mind this much off-topic, I'm ok leaving as it is though, don't feel like I need to ask you anymore about it. Thanks, and also apologies if I insulted your experience by asking you if you'd run the same increased voltage in Windows 7, 'twas not my intention.

[mikecacho]

@Fox

If this is what you: Want: Alienware ALX-18 - Core i7-5960X | Unlocked Dual BIOS | 3xGTX 980M SLI | 64GB DDR4-4266 Quad Channel| 6x1TB mSATA RAID0 | 3K IPS| 1kW AC Adapter

And this is what you are thinking of: Have: Alienware M18xR2 - Core i7-3920XM | 2xGTX 980M SLI | 32GB DDR3-2133 | 2x256GB SSD RAID0 | 2x2TB HDD RAID0 | 250GB mSATA | Dual 330W AC Adapter

It might be best for you to just get them 980M cards for your current system, trade in the 780M cards and hold out for another few years since that "want" system is a pipe dream at this point. with AW.

[Cloudfire]

Really? Can you imagine that someone like myself or Johnksss would not have already tried that, LOL? C'mon, I don't know anywhere close to everything, but at least give me a little credit for not being stupid, bro. I've spent hundreds of hours testing to get my CPU overclocks dialed in for benchmarks, otherwise I wouldn't have any kind of status at HWBOT. Windows 7 CPU and GPU settings don't always work as great with Windows 8. It's super easy to test with a dual boot system.

The only thing that has changed now is I am trying harder to get good CPU performance results from Windows 8 to determine if I am going to bother with a 980M SLI upgrade. If I can get equal or better CPU performance with Windows 8, I will ditch Windows 7 and jump on the 980M upgrade bandwagon... but, so far that doesn't seem possible. This isn't my first attempt to coax Windows 8 into working well. It was too much trouble and I gave up before because I don't like wasting my time on a what seems like a hopeless case. So, now I am starting over, trying things that didn't work great for Windows 7 to see what Windows 8 likes best. There was simply no reason for me to care that Windows 8 sucked until now because there was no reason for me to use it. The more I try to force myself to use it for benching, the more I hate Windows 8. For normal everyday use and gaming, Windows 8 is fine. It does 3DMark 2013 better that Windows 7... and, well, so far, that seems to be about the extent of the value of running Windows 8 for benching.

Since you don't have an unlocked CPU, here is a clue for when you get one... Voltage and power (watts) both affects core behavior. More is not always better. Too much or not enough of either one reduces performance and stability. Yes, performance is related to core speed, but voltage and power are the keys to achieving high core speed with stability and no throttling. You cannot separate them like your comments suggest. Simply dumping more voltage on the CPU doesn't (usually) produce good results. For some reason, Windows 8 behaves differently than Windows 7 with increased CPU core voltage.

You will probably explode now in pure anger, but I`m getting a notebook next week with 4940MX which I plan to downclock and undervolt with Intel XTU...

Yeah I know, you are like now.

I will of course play around with overclocking (A lot of it), but when I`m not doing benches or need greater CPU performance, I`m gonna make it spew out as little heat as possible when I found the CPU speed needed for my tasks. I plan on getting a pair of GTX 970M which will do nicely with my 100W heatsinks, so someone will finally test out AW18 with Maxwell to find out if AW18 works with them. If someone doesnt test them before me that is

[Paolo11]

You will probably explode now in pure anger, but I`m getting a notebook next week with 4940MX which I plan to downclock and undervolt with Intel XTU...

Yeah I know, you are like now.

I will of course play around with overclocking (A lot of it), but when I`m not doing benches or need greater CPU performance, I`m gonna make it spew out as little heat as possible when I found the CPU speed needed for my tasks. I plan on getting a pair of GTX 970M which will do nicely with my 100W heatsinks, so someone will finally test out AW18 with Maxwell to find out if AW18 works with them. If someone doesnt test them before me that is

What notebook are you getting?