*OFFICIAL* Alienware "Graphics Amplifier" Owner's Lounge and Benchmark Thread (All 13, 15 and 17)

I really do hope that rumor is true. it will amuse me greatly.

I literally just tried to do this last night and it would not fit. The PSU also wouldn't power on, but the screws werent log enough. I was super disappointed. Did you have to use different screws?

 

Greetings all. So I've been away from the forums, but still using my Alienware 17 R3 w/ 970M and the AGA with an ASUS Geforce GTX 1080. Up until this morning, it had been rock solid and running great.

I bought NBA 2K18 for the PC, figuring I could play in 4K on a slightly older Samsung 27" (I think) display I use hooked up to the 1080 in the AGA with a DisplayPort cable, and the internal LCD in the laptop disabled in the Windows 10 properties. I'd run this several times, just playing the game in the MyGM mode. I have the resolution set to 4K, and I'm currently running the Nvidia .41 drivers installed that worked fine with Geforce Experience installed and updating my drivers. However, there is a newer Nvidia driver that was released 9/20, and I couldn't install that. Saw the suggestion a few pages back to uninstall the Geforce Experience and install the latest driver. Gonna try that shortly. I'm also NOT using the latest BIOS I found on the Dell site. Since I use and love my AGA, wanted to look here and see if there was any reason not to install it.

For the record, the game isn't Blue Screening the PC. Instead, the external display just "goes black", and disconnecting the DisplayPort cable and other attempts to get a screen back up just fail. Instead, I have to hold the power button and wait for the system to turn off. I turn it back on, and it works for a while, but once I run and use the game for a bit it crashes again. Which is why I'm going to update to the latest Nvidia drivers. Hoping not to have to "hack" them, as I haven't had to with this current setup before, but will if that's what it takes to get things working again.

Any thoughts appreciated. Love these forums for the active and helpful community! Thanks.

 

Bah. Still won't install. Guess I'd need to hack it to get it to work.

http://nvidia.custhelp.com/app/answers/detail/a_id/3814

 

Could be the monitor. I personally hate display port. Try totally power the monitor off, leave it for a second to clear everything out and try it again. I had to do that to get the AGA and my DP monitor to play nice apparently it's a common problem with DP monitors.

 

I used the Dell screws and it worked. I got the one closest to the power plug started but not tight. Then did the top ones, they were easier, then the one on the right (when facing the back). Once those were all good I went back and tightened the one by the plug - it gave the most resistance. The back plastic did flex ever so slightly but nothing that is visible or has me worried. Are you using the same power supply?

I really wanted to use the Corsair SF450 but just couldn't imagine how I could make that work cleanly - power plug is in the wrong place. A fully modular PSU would have been ideal but I only have one unused connector which is not bad (the EPS 12V). I only used the CX650m because it had a rebate making it cheaper than the CX550m at the time (bought at MicroCenter). They are the same dimensions though.

If I could not get the new PSU in well, I was ready to just pull out the PCBs and rebuild in a case of my choosing (but sacrificing a lot of space in the process). All in, I am at about $250 before the GPU which I am pretty happy with compared to other eGPU enclosures.

 

I used the CX450m ... Same PSU, different wattage. It was on sale for 27 bucks after MIR, so I figured it would be good.

Any ways yeah the clearance was just barely not enough. Either way, It's on it's way back to newegg. I'm gearing up to sell my EVGA GTX 1080 Ti Founder's Edition so I can buy hybrid and then It wont matter lol. I'll just unplug the loud ass internal fan.

Still gonna probably try and replace the PSU, but I will be a lot pickier about which one I select, the coil whine on the PSU inside it is awful. I've got a Supernova 750 P2 sitting around doing nothing, Super disappointed I can't use that one lmao.

 

Out of curiosity, why are you guys changing the power supplies?

Just making sure I didn't miss something somewhere. I'm using the stock one with Titan X (Pascal).



Excuse the mess, I am in the process of moving.

 

Horrible Noise, Horrible Coil whine, Obviously the lowest ****tiest PSU quality possible.

It also has no additional peripheral power if you decide to use another fan inside there because the default fan is loud as hell and has low CFM. Many people might suggest modding the build in wire but that requires cutting, and I would prefer less destruction of the product, as would many others.

 

Ah yes the coil whine, I just ended up removing the whole top cover. No more fan noise, and it looks nice with the Titan X chilling outside.

 

If I had any craft skill at all I'd probably build some kind of enclosure. As it stands, I may just take the cuts out and mount them to something else. Having a fully exposed GPU in my house sounds like a nightmare (cockatiel) but it would look damn cool.

 

I was debating whether or not to dremel the excess part of the amplifier bottom part that exceeds pass the GPU. I decided to leave it as is, in case if I ever want to put the cover back on (even though the hinge is broken it still closes somewhat).

GPU runs cool on open air. I get mid 60s with my fan profile in any game that stresses the card full load.

 

>with my fan profile

you mean Jet Engine mode?

The FE coolers this generation was sexy af, but they loud AF too. my 1080 ti stays nice and cool at 80% fan speed, but holy **** is it loud.

 

Mine isn't that loud. At 70C the fans are at ~75%. But mine hover around 60-68C so it's lower normally. I think 80%+ is where it starts getting annoying. At 100% it's very loud (still no where near as loud as my 3x SLI GTX 480 setup).

BTW does anyone know how to stop the drivers from reinstallng EVERY time I switch from AMP to dGPU and vice versa. Literally all my profiles etc go away whenever I switch. I even tried modifying the INF so the drivers are from the exact same download but that didn't work either.

 

hi guys, i just bought the AGA today but unfortunately i dont have a GPU to test it out
i plugged it in and connect to my AW15 and the head logo turn ON for a second and the FAN also kicks in for a second
then BOTH turn OFF
i repeated by unplug and plug it in and still the same

BUT i tried plugging in my USB mouse and it WORKS

does it mean my AGA is half-broken ?
or can someone tried plugging in their AGA without GPU ?
please im so confused right now

 

Without a graphics card, the AGA acts like a USB Hub. The light and fan shouldn't be on, so this is normal.

 

Not trying to plug one brand or another but I have not had any issue with my setup at all. The only thing close to an issue was after updating the newest NVidia drivers I started having graphical corruption randomly on Prey and only Prey. Rolling them back seems to have fixed it.

Given what I've read I think I can say the best AGA config is cross-platform card combo. All the problems people have talked about are non-existent on my system and that's because I'm using an RX Vega on my amp rather than a duplicate NVidia card. My guess is that there isn't a way to prevent the driver reinstall unless you could somehow stop the new device manager process within Windows.

 

I think nVidia needs to fix the way their desktop and mobile drivers install and make one unified driver.

Have you tried making profiles for nVidia , switch to AMD then check nVidia to see if the profiles stick?

 

I’m not sure what you mean by “profiles”. If you mean overclocking profiles in Wattman, yes. If you mean game specific configuration profiles, no I haven’t tried that however they are there I just haven’t used them.

When I swap between them I use the undock button or shutdown options and I have yet to get a driver or config warning at any point. The only issue I’ve had with my GA was I managed to strip the retaining nut out of the unit (over tightened) and had to superglue the brass nut back into position so I could screw my card back down.

 

I always do a full shut down when adding or removing the dock and have not had a single driver issue with nvidia mobile and desktop GPUs. With an NVMe SSD the reboots are so fast it is not a burden.

 

Just an FYI, I recommend staying away from Radeon 17.9.3 if you have an R13 - as soon as I installed and rebooted I had nothing but issues getting my monitor to sync on from the AGA. After reseating all connections, and a few other things I rolled back. After two more reboots its now working again.

Oddly while troubleshooting I noticed if I powered my monitor off and back on, the laptop screen would go dark, yet the monitor would not sync. Powering off the monitor would then allow the laptop screen to come back on. *shrug* I'm going to assume its a software issue.

 

I also have the Titan X (Pascal). Can you please tell me how you updated your driver? I just keep getting the Nvidia error message about the AGA.

This also goes for anyone who might have the answer. I have a Titan X (Pascal) in my AGA, and it has worked fine for a year. A game I'm trying to play is stuttering, so I figured I'd update the driver as I'm still running 376.33. However, I now get an error from Nvidia when I try to do so. It tells me I need to use a driver based on my laptop's graphics card, which is a 980m. Why would I want to do that, though?!? That seems like it somewhat defeats the purpose of the AGA.

Thanks in advance for any assistance!

 

Use DDU in safe mode to remove the current driver then it'll restart, install the latest driver after.

 

Worked like a charm. Thanks, ssj92!

 

Greetings all. I've been posting in another thread about my Alienware 17 R3, my AGA, my ASUS 1080 GTX (Founder's Edition), and the system going from solid and stable for months, to suddenly "black screening" (one minute windows is working fine, the next the video just goes black on whichever monitor I'm using). I have to hold the power button, wait a few moments, power back on, and usually it boots back normally. Sometimes it says I'm not using the proper Nvidia or ATI video card, and won't continue the boot process till I press a key.

I was running NBA 2K18 for four days before this all happened, so I wondered if updates to that game or something might be involved. I've installed ALL the Dell drivers I'd been avoiding previously, went to the VERY LATEST version of the BIOS (1.3.12 I think), started using HWiNFO64 to control my cooling fans and make sure the laptop and CPU aren't getting too hot, and yesterday looked up the stock frequencies and voltages for my video card, and backed those down from "boosted" to "standard" speeds and such. Only after leaving the laptop on all day while gone to work, coming home starting up NBA 2K18, switching to the browser briefly, going back to NBA 2K18, and not but a few seconds passed and it happened again.

I'm not sure exactly what I'm supposed to be looking at, but because my Nvidia 1080 GTX seems to be involved, I'm wondering if I might be having a power supply problem. I've noticed at least a few posts here from people swapping the Dell stock unit with other options like a Corsair. Was there a specific way you knew you needed a AGA power supply swap? What can I look for to see if I might need it to?

I've really enjoyed my Alienware 17 R3, but I feel like all I've done for two weeks is try to fix problems with this laptop now, and it's REALLY getting old. I'm not stupid, but I don't know what voltages and frequencies everything should be running at, so if there is an obvious problem in my monitoring, I'm not seeing it. Any thoughts or suggestions are appreciated.

 

Often, the above problem occurs when I'm trying to run NBA 2K18 in 4K resolution. Wanting to take it easier on the system, I've tried running it in 1080p, and I've generally had better success. Did that now, saved the changes, went to look at my roster of my "MyGM" mode game, and the system black screened, the desktop video card fans when from not running or REALLY low to maxxed out, and everything stayed that way until I powered off and on again.

So frustrating. There must be a log file or something I can look at to help diagnose this issue!

 

So I was nearly done typing my question and the browser F'd up. Grrr!

Anyone else with an Alienware 17 R3 having trouble loading the Nvidia 385.69 driver with your AGA connected?

If I disconnect the AGA, Nvidia's site finds and downloads the mobile 385.69 driver and it installs fine.

But if I connect the AGA, Nvidia finds and tells me to use the 385.69 desktop driver. Only when I run it, it says it's not compatible with all my devices, and exits out without installing anything.

I was reading in another thread folks create their own "versions" of drivers to support things like CUDA. But I neither have the account to log in elsewhere, or the invite needed to access those drivers. And I suspect they aren't "exactly" what I'm looking for, even though someone must have a tool or the drivers that make 69 work like the desktop drivers did before this one.

As I'm STILL trying to figure out if I have a driver problem, or a hardware problem with my AGA and ASUS 1080 GTX, any help is appreciated!

 

This is what I would do, step by step:
#1 Download both driver packages
#2 Unzip the desktop driver package if possible and make note of the folder.
#3 Disconnect all internet sources
#4 Go into the device manage, uninstall both graphics adaptors and make sure to select "remove driver software from PC"
#5 Install the mobile drivers without the AGA and shut down
#6 Reinstall the AGA and boot
#7 Go into the device manager, locate the 1080, select update driver, use specific driver, have disc, specify the location as the .INF file you unzipped for the desktop driver package. Install.

If that doesn't work then it's either a BIOS issue, a hardware issue, or some weird compatibility issue with dual NVidia cards. Either case buy an RX Vega or similar RX card. I would then go to bestbuy or similar and buy an AMD card, *any* AMD card and see if I could reproduce the issue; if not then you've at least isolated it to that card or the power supply. If it happens again now you're looking at the power supply or the game itself. Fortunately they sell PSUs at BestBuy as well.

However if you're able to get either an RX Vega or an RX580 and it works flawlessly I'd pin the issue on a driver issue.

 

Good advice man. I basically did exactly what you said, going from 385.41 which worked perfectly for Destiny 2 and NBA 2K18 at I've pointed out before, to 385.28. Did a clean install, and set the 1080 GTX to be the primary video card though the Nvidia Control Panel. Wouldn't you know it, I've played Halo 2 without a single glitch, and just spent maybe an hour playing NBA 2K18 without a SINGLE problem. No lockups, no black screen, nothing!

I'd be lying if I said I understood what the hell happened, and I'd be curious to know if I hacked 385.69 to work with both desktop and laptop video cards, if I could install that on the laptop and how that would work. But for now I'm thrilled that I can play games on my primary gaming rig again! And I guess the Corsair PS can go back to Best Buy unopened in the next couple days.

 

try modded driver from @j95

 

Glad to hear man! It's just a driver conflict because windows is ****. What is happening is you have two device IDs with two incompatible software and Windows is getting confused which driver package to apply. What you did with what I outlined and set specific boundries and guidelines of use this when card A is installed and use B when card B is installed rather than whatever available driver package is convenient. From here on out you shouldn't have any further issues as each device ID on your system is locked to a specific driver package so install away!

 

I wish it was true. So I played for several hours last night. Exited the app normally, shut down the PC, went to bed, work this morning, came home, powered up, and played for about 2 hours until...

BLACK SCREEN!

Couldn't Alt-Tab, couldn't do anything. Held the power button, turned the system back on, and now I'm posting here again.

I had NOT returned that Power Supply. I have no proof that might be the problem, but here's what I've found while troubleshooting.

1) The laptop itself is VERY clean. No dust in the fans or headsinks that I can see. This is kinda surprising to me, as I've had two ASUS gaming laptops in the past (aka Desktop Replacements). They were solid, fast, and rarely had these sorts of issues (except for the onboard video card BIOS needing to be updated after 6 months, which I didn't even know you could do to a system like that!). So while original thoughts of overheating because it might be a little dusty and need to be blown out weren't a bad thought, there was nothing to suggest that's the problem.

2) The AGA, and the ASUS 1080 GTX video card, and all it's parts seem very clean. When I opened it, the front fan was working, and without taking it out and looking on the under side, all the parts of the "motherboard" looked clean and in good shape. No obvious capacitors or burns on the board making me think the Graphics Amp is the issue. Though that doesn't mean its not.

3) The only SUSPECT in the AGA was the power supply fan. I've read here it's normally loud and annoying. I left the AGA open while testing the video card before switching to the latest driver, and the power supply not only wasn't making ANY noise, but I didn't feel air being drawn into the Power Supply or pushed out the back. That said, I wouldn't say the power supply felt hot, so I didn't freak out about it not running hot, loud, etc. But if EVERYONE else is really having issues, maybe it does need to change.

One more bit of info for now. Even before I switched to the new video driver, I tried NOT running HWiNFO64, in case there was some part of that app that was interfering with the video driver, etc. But after the system "black screening", booting back up, and coming here, the system wasn't keeping up with my typing. It felt hurky jerky. So I just started HWiNFO64 up, went to the sensor tab, looked at the fan settings, and the motherboard or whatever fan(s) are coming on with the CPU settings I set, they came up fast. The system cooled off within a couple minutes, the fans have slowed back down, and I'm able to type normally again. Which still concerns me about the BIOS and fan management in this system. Is it not cooling enough without running FAN PERFORMANCE MODE in the BIOS and ALWAYS having the fans run full speed? That certainly feels like overkill, but I don't like having to run extra 3rd party apps just to keep my system cool enough not to have issues either.

And I haven't even tried running NBA 2K18 again after it locked up. I had traded for the best pick to hopefully earn the #1 overall in the lottery. Drafted a kid who has the ceiling of Labron supposedly. And I had signed my rookie, moved into the pre-legal signing period, and was making offers to players to bring them into my Golden State Warriors team. Durant declined his final year option, so I'm making him a deal, while trying to bring in a 80+ Center, and hoping to resign Klay Thompson, all while looking for 2nd and 3rd stringers. And because there is a salary cap and such, it's all pretty crazy. Sure hope I didn't loose too much work from the crash.

So yeah, back to troubleshooting I guess.

 

Alright. So I came home from work, disconnected the AGA, and opened it up. Didn't realize I needed to remove a metal piece toward the front of the case, but until I did the Corsair power cable wouldn't have anywhere to go. Between a flat head screwdriver, and then a pair of pliers, I pulled on it till I got it free of all the plastic holding it in place. I'd already unscrewed and removed the old power supply, so I put the new one in place, eventually getting it in place enough to use the old Dell PS screws and mounting the new unit in place. Cleaned up the power supply to the motherboard cable and the unused "MB" cable with a Velcro strip keeping it all as tidy as I could. Then I used just one of the Corsair cables to power the video card off the power supply. Put it all back together, booted it up, and so far... so good. No signs of issues, played NBA 2K18 twice, each time no glitches or lockups.

Still using the 385.28 driver. I'd like to get the j95 driver so I can use 385.69 with my system, but I'm still not clear where to start. Do I PM j95? Do I go somewhere outside Notebook Review? Sorry, sometimes I'm still a newb about stuff.

 

So it was power limitation of PSU.
Here's the link nVIDIA GeForce Drivers v385.69 WHQL Findings & Fixes

 

I'm really struggling to find some benchmarks that compare a 1080 ti in the AGA to a desktop.

Can anyone show me some?

I'm waiting for my 13 r3 with 7700hq to arrive, and currently have half a pc that I was intending to upgrade with a 8700k.

At the moment, i'm just wondering if it's really worth buying a z370 motherboard and a 8700k or if I should just sell my other parts (minus my 1080 ti) and get an AGA.

I currently play at 1440p144hz, so I feel like the bottleneck is likely to have a larger impact (I want that max framerate), but it seems silly to have an entire separate machine with 1k worth of parts just for an extra 5 percent extra performance.

if however it's more likely to be a 10 percent or higher loss of performance, i'll go ahead with my separate desktop.

Any thoughts?

 

Probably power, but an outside chance it was heat. The video card never lacked it's fans, and with the large grills on the AGA that probably wasn't it. But the fact that I never felt the fan in the PS "ejecting" hot air out the back of the AGA was weird.

Thanks for the link!

 

So I downloaded the 385.69 file, and the zip. Extracted the 69 files from the exe to a folder. Copied the 4 files from the ZIP, and pasted them into the Display.Driver folder I'd created when I extracted the 385.69 files, overwriting them as I did so. I unblocked the setup file in the 385.69 folder, ran it as admin, told it to do a clean install... and it failed.

Not sure what I'm missing. Was I not supposed to use DDU to remove Geforce Experience and 385.28 using the DDU in Safe Mode before the install? If I wasn't supposed to do that, maybe that's the issue.

 

Think I got it. Ran the command line thing about Legacy with admin rights, and I booted and chose to ignore driver signatures or however it referred to it. Trying the install again now.

 

If you're wanting 1440P @ 144 FPS or max framerate you've picked the wrong laptop to do that. The loss of overhead coming from 16X to 4X won't be nearly as large as the loss of CPU clock from a non-HK processor, you really should have looked at either the 15 or the 17" unless you absolutely wanted the 13" form factor. You're looking at a processor that's locked at 3.7 versus one that is easily overclocked to 5+ so no amount of graphics power is going to overcome that especially with you only at QHD. If you want the small form factor and portability then you're going to have to accept that 1080P will likely be the resolution you play at when running 144hz

 

Thanks, thays all i needed to know.

I didnt buy the 13 specifically for playing at 1440 144, it just ended up being the laptop i went for for playing games while travelling.

Sounds like im befter off sticking to a seperate PC for gaming at home, so ill just go ahead and get my new mobo and cpu for my desktop.

Thanks for the info.

 

Oh computers. How you torment me. Let me count the ways...

So let me start THIS post saying I JUST double-checked my Windows 10 Advanced Hardware settings, and I HAVE set drivers to NOT update automatically.

So I'm up early. NBA 2K18 hadn't had an issue since the new PS was installed. I haven't been running HWiNFO64, because I hadn't seen any reason to. I'm surfing the web, and all of the sudden the screen went black. I'm like... no way. I'm not even gaming!!! I wait a moment, and the screen comes back. But it's clear something happened. My GPU Tweak II software from ASUS isn't showing quite right, and when I check what driver I'm using, I'm somehow using 382.05. I had figured out how to get the custom driver installed, and it HAD been working great. Not sure WTF happened.

For now, I'm gonna boot to safe mode, remove the GeForce Experience and any Nvidia drivers, and then install the Destiny 2 driver that HAD worked up until this mess started. Then I'll see if this looks like an on-going issue or what.

 

Guys,

I'm into buying a 1080Ti for the AGA.
Do you think is a good time or should I wait for the 11XX line? Any news regarding the release date!?

I'm currently running a 17 R2 with 4980HQ/980M.

 

The 1080 Ti is a solid card and will not be less so when some future generation is released. I would just say be patient and buy something near MSRP not one of the crazy overclocked configs or boosted prices due to mining. I got an Asus 1080 Ti from $700 from MicroCenter I have been happy with.

 

Realistically speaking you aren't going to see much of a difference. A 970M is roughly equivalent to a 1060 and a 980 is roughly about the same as a 1060 non mobile. If you bought an AGA and then bought a 1080Ti you'd be roughly at twice where you are now at best. Is that enough for you?

As for the 11 series I wouldn't expect much; most likely what you will see is a die-shunk version of the 10 series and again not much upside until Volta but that is unlikely to launch at all during 2018 until closer to Navi's launch. If double what you have now is something you are OK with then go for it. But bear in mind it will cost you almost $800 when you factor in the cost of the AGA and the card.

 

Ok, so I just noticed that I can't update my drivers past 385.41 when the AGA is connected. What gives? Is it because of my graphics card (1080FE) or does this happen to all cards?

 

It seems like a lot of the driver related problems are coming from users of older laptops where the PCIe lanes come off of the PCH and not off of the CPU. Anyone else noticed this?

On my Alienware 15 R3, it looks like the 16 lanes off of the CPU are divided as x8 for GTX 1070, x4 for ThunderBolt, x4 for AGA so nothing overlaps and you can have both ThunderBolt and the AGA in use at the same time without competing for resources. Also this design removes the need for any expensive and hot PCIe switch chip. Seems like they realized the shortcomings of the original design and improved it with Skylake and newer laptops.


Also, although I have the Noctua fan installed, I currently have it unplugged, it is just too loud and I am not really convinced it is necessary.

 

That may be the case but I see far more correlation between people who have issues using both an onboard NVidia GPU and an external one. While I am sure I am in the minority I have not had one driver related issue with my configuration 1060 - > Vega56/RX580

 

Another possibility is that people who have one generation of Nvidia GPU in the laptop and a newer one in the AGA could cause driver conflicts because different parts of the driver are being called on? For example, I have a 1070 in my laptop and a 1080 Ti in my AGA and have not seen any issues yet (both Pascal). Most users with issues have 9x0m series GPUs and are trying a 10x0 GPU in the AGA. Just throwing out ideas right now but it occurred to me.

 

Is this why nvidia is no longer adding the drivers in the latest update?

Sent from my ONEPLUS A3000 using Tapatalk

 

Good catch!

No, NVidia tends to take planned obsolescence quite seriously and so as soon as a new generation arrives almost all but bare minimum support is given.

 

This should not really be a problem though. By the time a new gen comes out, the existing gen hardware is pretty solidly supported in the drivers. The only reason new drivers would make a huge impact is things like SLI profiles for new games but that is less and less relevant with a) being on laptops that don't have SLI b) the direction Nvidia is taking on SLI going forward (lot less of it happening) and c) DX12 and games making explicit SLI support less of a required thing