*OFFICIAL* Alienware 17 R5 Owner's Lounge

I mean the 2242 slot is Sata.

 

I didn't find the Killer 1550i anywhere...

The Intel 9260 and killer 1550 are the same cards, now the Intel 9560 is the newest one and works with m.2 CNViO that I don't know if this Alienware 17 R5 is already able to use that.... That's was my main question... So if anyone knows if I will be able to upgrade from killer 1435 to 9560?

Enviado desde mi Pixel 2 XL mediante Tapatalk

 

I apologize if this was not clear - "The Intel 9560 and Killer 1550i are the same, but neither will work in an Intel 9260 or Killer 1535, or earlier version, slot." Therefore you will not be able to upgrade from the Killer 1435 to the Intel 9560. "If you have a Killer 1435 you could only upgrade to the Intel 9260 or Killer 1535."

 

Yes I understood your message. But I can't find that killer model, the 1550i, just curious I haven't hear of it, can you give a link to that card please?

Enviado desde mi Pixel 2 XL mediante Tapatalk

 

Use Google. It’s your best friend.

 

I did, and I did went directly to the killer website and there was no info about the "i" variant... Thanks anyway!

Now why will not work on the i9 8950hk? In the Intel website it says this CPU is CNViO ready, that it have that module...

Just an answer... Thanks in advance!

Enviado desde mi Pixel 2 XL mediante Tapatalk

 

The point is, the sockets are different, so neither the Intel 9560 nor the Killer 1550i will fit in your socket.

We offer the Intel 9560 and Killer 1550i on several of our models, but the Alienware 17 does not have the proper socket for them.

 

You can start read here... https://www.intel.com/content/www/u...6155/network-and-i-o/wireless-networking.html

Read the Note.

 

Yeah I am not sure either. I thought at first the "i" was just a variant for certain countries (frequency differences? don't know if that's a thing with WiFi like with other radio devices including cell phones - but I would assume so), and since we have the 1550 in ours, then the 9560 would work. Sounds like Donald knows though from being in the business of customizing and building these and the other brands.

I wanted to report though, my Killer has gone 24 hours now with no issues. I almost find it hard to believe, but perhaps the full uninstall/reinstall that Killer recommended to me really does work. I am copying and pasting their email below. I had done the full uninstall on my own twice before, but not to the depth of running their own custom uninstaller, as well as deleting temp files and checking the other Killer directory was gone (it was already removed). I am still thinking it could be total bandwidth related before it goes out, but this is a good sign. I still can't get the download to work in Microsoft Store for the Control Center, but I sent them an advanced diagnostics output from their own debugging tool they linked me to and will report back if they have further troubleshooting info. At this point, I don't really care about the Control Center, I am not even on that high speed a connection anyway (15-20 mbps down, 0.5-2 mbps up!), so I pause my onedrive/dropbox manually when needed already.

Here are the steps they recommended:

The Killer Control Center in the Windows Store relies on certain extension drivers being installed correctly from the installer on our website, or from somewhere else. You may be able to resolve this by clean installing the latest version of the Killer Control Center, using our uninstaller first to remove any previous installs. Here are the steps:

1. Download the latest copy of the correct installer and put it someplace handy - https://www.killernetworking.com/killersupport/driver-downloads/kb/faq/22-update-killer-control-center
2. Download the Killer Software Uninstaller Tool - https://www.killernetworking.com/support/KillerSoftwareUninstaller.exe. Make sure you re-download this, even if you have downloaded it in the past, as we are constantly tweaking it. Put this file somewhere handy.
3. Restart your computer by clicking Start > Power > Restart
4. Once the computer has restarted, run the Killer Software Uninstaller, and click Remove Killer Software. Once the tool is finished, it will say "REBOOT REQUIRED."
5. Restart your computer by clicking Start > Power > Restart
6. Once your machine has booted back up, access the temp directory by clicking Start and typing %temp% and press Enter.
7. Delete the contents of this folder. These are temporary files that are meant to be deleted from time to time - you cannot harm anything by deleting all of these files. You may not be able to delete some of these files. It is okay to "Skip All" for these files.
8. Using the File Manager, delete the folder that contained the Killer Control Center. By default, this folder is C:\Program Files\Killer Networking. You can delete the entire folder Killer Networking. If you cannot find this folder, it just means that the Killer Software Uninstaller already deleted it, so you can skip this step.
9. Double-click the installer for the latest Killer Control Center. and follow the prompts to complete install.
10. Restart your machine.
11. Test to see if the issue is resolved.

 

Yes, that is what I seemed to find when I quickly researched it a month ago - good for an m.2 SATA drive, but no NVMe in that half slot.

I wish there was a way to put a WWAN card in there. From what I understand though the antenna would need to be routed and would probably cause interference with the WiFi/Bluetooth. (it seems the machine needs to be designed to support WWAN - if anyone knows anything about that, would be curious)

 

That would be fantastic if it was pre-wired for WWAN but I'm not sure how effective adding in your own antenna array would be. With the right amount of research I'm sure one skilled with wireless communications could probably make it work rather well. Personally I'd be all for it if Dell offered that upgrade / pre-wired for WWAN like they do for many of their laptops on the consumer / business side of the house.

 

Do you mean the front corners (away from screen) can get messed up if they bang against something, or delicate in some other way?

I do like the coating actually but I see what you mean about it needing care in handling.

One other thing I realized I do like a lot (perhaps the ONLY thing about the Tobii that works well and is useful, beyond the Hello login) - I set it to dim my screen after just 5 seconds, and it's very nice to look away at another computer or paperwork, and not have the screen on so bright in my peripheral vision, then have it turn back up instantly and automatically when looking back at it. I'd really miss that feature I think. (maybe any Hello-compatible camera can do that though, and it's just new to me... but I like it!)

 

No, this is a Tobii feature. I haven't seen anything else on laptops that does it. Samsung Smart Stay does the same thing on my Note8.

 

Cool, that's good to know! Yeah I thought I had heard of one phone or another doing something like that. iPhone has done the "turn on screen if picked up" thing which is handy as well, but this face recognition version, as on your Note8, is really useful. To me, more useful that any voice recognition usually is - this just works, without even thinking about it.

 

I also came to post an update - my WiFi stopped working again and needed a restart. Oh well. I guess the complete uninstall Killer sent me didn't work after all. (although it did stay up the longest yet, weird) I still wonder if it's something else causing it though, since I can ping domain names - it's just anything else that doesn't work (different web browsers, one drive, dropbox, anything else Internet-connected).

 

The whole top surface, everything surrounding the keyboard, is made of a very delicate material that can be scratched very easily as you rest your hands on it, and is also very hard to clean it seems it stains in a very weird way sometimes, I do not know what material it is, I think some carbon fibre derivative. All in all I'd prefer brushed steel or aluminium alloy something.

If you can ping then you have wifi-connection which is what the wifi card is responsible for. Does it happen after a sleep ?

 

The palmrest itself is a form of ABS plastic with on the bottom an aluminum frame and a mixed steel+aluminum plate (the motherboard is mounted onto that aluminum frame). The soft touch finish itself is a coating.WIth my last AW15R2 it lasted quite a bit, after 2 years it started to become harder to clean. In general avoid strong solvents to keep the coating in good condition and avoid alcohol (alcohol also eats ABS slowly by the way)

 

Anyone with the new update have an issue with command center? seems i cannot change keyboard colors, so far is that plus the windows hello and brighness issue with the new update

 

On the wi-fi - no, I have it set to never sleep even with lid closed, and it's plugged in around the clock. I know, it's weird, the ping works, but http does not. I just tried flushing my dns and running some other reset commands, will see if it acts up this time after another day. I forget the details but originally, it definitely seemed like a Wifi card problem though; I really don't remember but I think the issues it was having were more constant then. For example, I know that the first week I had the laptop, Outlook just would never send out a message because the connection to smtp would be "interrupted". Yet after I did my first reinstall of the Killer driver and removed the control panel, that has worked flawlessly ever since. Maybe there are additional issues going on.

On the deck material - maybe carbon fiber like Lenovo uses is better for longevity. I for one am happy to be away from Apple land where everything is cold or hot aluminum all the time (and slippery - so many times my MacBook would slide off the huge dashboard in my car with not even much acceleration; or off my lap, etc.). Aluminum looks nice in a photo but after a while I personally think it gets boring to look and is rarely a comfortable temperature. Plus I find with black decks and screen surrounds it's easier to actually focus on the task at hand... Rinneh's advice is good to avoid alcohol or strong solvents, that definitely makes sense.

It's like with anything in design though, everything is cyclical. Like with car wheels - every few years they'll change the rims from spokes to a more curvy design, then a few years later they'll go back and say spokes look better. Maybe down the road I'll be craving aluminum again too. (I doubt it though, for the temperature issues alone.)

 

Sorry to bother you again - did you change any other settings, or just the undervolt and multipliers? No core voltage change (default or 1.25v, I assume)?

I have the CPU core undervolted up to -170 mV and the cache at -142 mV and get the best results on Cinebench with multipliers at x43. I get a 1206 in Cinebench. Only one core seems to get to the low or mid 90s, all others are high 70s to mid 80s. The Throttle radio button does come on and off though in ThrottleStop.

I also know you said you set the AW Command Center to 5Ghz and -100 (I have mine set to 4.6 and -100, the best I could get before trying ThrottleStop); but the ThrottleStop overwrites whatever Command Center sets when it start up, so the Command Center settings are irrelevant, right?

Finally, one other weird thing - when I set different multiplers in TS for 6 cores vs single core, then run the single core test in Cinebench… the clock speed in Throttlestop is still the same across all cores, and it's limited to the 4.3 that I've set for six-core operation. It's not possible ThrottleStop isn't actually working properly in setting the different multipliers for different numbers of cores, is it? Or maybe whatever it uses to change the setting is being triggered even though Cinebench thinks it's doing a one-core test? Did you set the same multiplier across the board?

If there's no other settings, I guess I have to blame the thermals/cooling for the limit I am seeing, right...

 

I am not sure, is maybe the undervolt just an offset to some predetermined curve which is *different* for each CPU ? If yes then that means that the absolute UV value cannot be compared across machines.

 

Well, I did a lot more testing. Some of my findings:

1. Only one core gets to 92 and above, all others are at least 5, if not 10, degrees below that. But that one core causes thermal throttling still sometimes. I tried XTU separately from Throttlestop also - it has per core limits, but it doesn't really work. The CPU applies the lowest setting to all cores so that's not an easy fix. I was hoping to underclock that one hot core and let the rest run higher, but that won't work.

2. The highest I can get on mine is -148 mV on core/cache, and 4.0 for six cores active. Sometimes it will stay at 3.99 for the whole Cinebench test, other times it falters.

3. Looking at XTU, it seems to be power limit throttling that happens more often than thermal throttling. So is this is still an issue - it just won't allow enough power to flow? I couldn't even get a single core in the one-core test to get much above 4.15 Ghz. Power limit there too even though it was only pulling one core's worth of wattage.

It seems to me that even if I re-paste, if it's the power limit throttling happening, I am not going to fix that. Just looking at ThrottleStop, I was thinking, gee, maybe I should re-paste. But then from XTU showing the power limit throttling on 30-60% of the time I run the benchmark, that is the culprit.

Is there something I am missing? Otherwise the max Cinebench I have achieved is 1206, and for some reason now I can't get back to that, around the 1150-1180 range.

I can really appreciate now why folks say the 8950HK isn't worth it compared to the next one down. (this was all testing without heat from GPU either, of course...)

I mean, it's not so much the cost, the difference isn't that huge between CPUs, it's just that I was hoping to get 1500 one day like Aristotle did, hehe.

 

I looked in your footer links and couldn't find it, but I was wondering - what is the single core Cinebench you get on your P870? I assume 280 or thereabouts. Highest I can get with the power limit is 192 it appears. Thanks...

 

At least this made me feel a little better - I searched for what Cinebench score the MacBook with the i9 gets, and it's only at 1017, with Dell XPS 15 at 1107. (kind of annoying the XPS gets a higher score out of the box than I did with this AW 17 R5 though! I was only getting 980 or something like that before I touched any overclocking controls)

https://appleinsider.com/articles/1...he-2018-15-inch-macbook-pro-with-dells-xps-15

 

Is there a solution for this Command Center bug?

The games, that are manually added in Alienware library, get minimized when escape key is pressed when opened through command center. If you close the command center, escape key works fine but your settings are reverted.

Steps to recreate issue
1. Manually add a game
2. Apply Full Speed to thermals for the game and Quiet for the System
3. Open the game and press escape key, the game should minimize.
4. Close the Alienware Command Center
5. Escape key doesn't minimize the game but the thermals should go back to Quiet settings as per system.
6. You can no longer alt-tab to desktop.

 

Well, I am not feeling so great - after five or so hours last night of overclocking, and thinking at least I had close to a 1200 Cinebench score in the bag, GTA 5 crashed within three minutes of playing. I lowered my undervolt to -120ish, and it crashed again! So I went back to just using the Command Center at -100, and it played fine for a couple of hours again. So basically with the -100 undervolt, I get the same score as an XPS 15 out of the box, which doesn't really make sense given the extra size and thickness.

I can try re-pasting, but is it going to help with maximum CPU speed, if the XTU tool is indicating a lot of power limit throtting going on? I guess I am not sure if more cooling via re-pasting means I can run with no undervolting, and that will solve the power limit issue; or is the power limit issue the 65w or whatever it is that everyone is saying Dell has put into the BIOS and which there's no way around? If it's the latter, I don't see how re-pasting will get me more speed, right?

1100 vs the 1200 is a big difference, and especially 1100 vs the 1500 score in-thread that Aristotle got. (and of course 1100 vs 1700 on the Clevo chassis - 50% increase - now that's something to consider in the long run again)

 

Yes. I have come across, and documented here in this topic, if you can find it amongst so many pages, maybe my post should be pinned, that there are two throttles (at least) that come into play very often, and these are:

1) Thermal throttle when the CPU package reaches 99C+. As you can imagine the frequency then drops immediately to stop the CPU from burning.
2) Power throttle starting at 93C-94C. It is a power throttle that kicks in once the CPU has exceeded 93C for a few seconds and it is designed to vary the power so as to keep the CPU below 93C. In the XTU charts you will see a thermal spike for a few seconds and thermal throttling (1), then followed by ABSENSE of any further thermal throttling but introduction of a power throttle (2) that comes on and off as required and for the duration. The actual power level varies depending on conditions, it could be 40W or 60W, and the only constant is the temp staying just below 93C. I think this is BIOS 1.2.1 and maybe even 1.1.6.

 

The proper way to do this is to open it up very carefully and remember there are plastic lugs and catches on the chassis and every time you open the laptop they degrade. Sad construction. You can remove the thermal paste and replace with graphite pads. I did it and got fantastic grouping of CPU temps amongst all cores. But then there are also the VRMs that get hot and throttle badly, down to 800MHz. I could not ever get them right because the thermal pads required are all of differing thicknesses and should all be compressible. It is not an easy job and not something you can test easily afterwards. You got to allow the GPU to run at 180W and generate a lot of heat, and then and only then test the CPU. I had the situation where my CPU + GPU were both below 80C, but the CPU was throttling down to 800MHz in a cyclical manner and the culprit were the VRMs just below the GPU reaching 104C.

 

Thats odd, the CPU shouldnt throttle because of GPU VRM's. GPU VRM's overheating normally results in in game stuttering because the GPU alternates between idle speeds and normal speeds.

 

GPU VRM's used to trigger 'BD PROCHOT', not PROCHOT in my laptop as long as I had poor GPU VRM pad vs. heat-sink contact.

BD PROCHOT would down-clock CPU only. Since I got GPU VRMs cooling right, there was no more CPU throttling. @propeldragon had the same problem.

See excellent pic. by @propeldragon.

 

Most likely overheating GPU VRMs send signals to EC who then tell the Cpu to slow down. This is meant to lower the temp on the whole platform where the GPU VRM's is a part of (shared pipes - The cooling can do the job if most components reduce the heat output). Remember Notebook manufacurers engineers implement weirds set of throttling algorithm. Nothing new there.

From my thread...
How Dell cripple performance explained by...

" How Dell's Dynamic Power Mode helps
"Each notebook has its own design and has to be individually optimized for maximum performance while remaining within the skin temperature limits. Instead of just setting functions based on worst-case thermal limits, power limits to electronic components can be dynamically adjusted based on temperature feedback from thermal sensors placed at specific locations in the system. Correlation between skin temperature readings, power draw, operating frequency, device temperature, and other factors are established that enable the Dell Dynamic Power Mode to predict the surface temperature. This results in a matrix of power limits and thermal rules that are applied in real-time to adjust power limits to various components."

 

Not merely 82C, I was hitting 104C in that location with GPU/CPU at about 70C-75C

 

LOL @propeldragon and also me, both, had no proper pad contact with heat-sink on the 'GPU south of heatsink' place because of too 2 thin Dell stock pads there. With Gelids the chips are fully swimming in the pad like strawberries in gelly lol.

Therefore the reading from 82C heat-sink is so 'low' during the moment Propeldragon measured temperatures with IR, the VRMs 'under' without proper padding under were definitely MUCH hotter. The specific VRMs caused CPU to throttle. But, that was about 6 months ago, now it is all good. I still keep BD PROCHOT unchecked in Throttlestop just to make sure no throttling occurs, even if I have no throttling with it when activated.

 

I never experienced drop sin that way myself.

But what I can think off that the PCH might be overheating and causing the 800mhz drops. That behaviour is very similar to the Haswell throttling which also ran into too high PCH temps. Intel didnt rework their PCH controller for 6core CPU;s so they become vastly more hot than the Kaby Lake counterparts.

 

Interesting. thanks for the info. Never could have imagined that the GPU VRM's would trigger anything. Because the Skylake motherboards just locked up because the GPU would basically stop working when the VRMs overheated and dropped to 0% usage. After a second of cooling down it would work again and it would go back and forth constantly.

 

As I said... Several ways to throttle down Cpu clock speed. See this DELL mess.

DIMM throttling:
So the 9570 has issues with DIMM temperatures. Once it measures above 62C it throttles the CPU to 800MHz. Putting thermal pad between the ram and the case helped me to avoid throttling but there isn't much headroom. Other users are cooling the sensor itself to avoid throttling.

http://forum.notebookreview.com/threads/dell-xps-15-9570-benchmarks-temps.817970/

 

Everything is so tightly packed on the motherboard that the IR thermometer may not be the best tool, I use a probe thermometer to be certain.

 

So you basically manually touch the VRM with the probe? This is cool.

 

@equalizer2000 , I have the EXACT same issue. This is super annoying. My AW17R5 keeps dropping the Wi-Fi. I've tried the latest drivers from Killer as well with no luck. This also effects the Ethernet port (since that's killer as well). Has anyone found a solution?

To recap, I leave my machine on 24/7 and every couple days my internet stops. I can still ping which is weird, but I cannot access sites or anything (no matter what browser). I tried Wi-Fi and Ethernet and it still happens. Only fix is a quick reboot, but i usually have stuff opened and is kind of annoying.

Anyone else experience this with their AW17R5?

Thanks
Jeff.

 

Well I guess I was having that problem too. At the time I associated it with my firewall that I built and thought it was a hardware issue with the port I used on it, but I sent it back before realizing it was another fault that computer had. None of my other machines have an issue with network including the Area 51 that took my m17's place.

It's a shame those laptops have so many issues.

Sent from my Pixel 2 using Tapatalk

 

I agree but in fairness this one doesn’t seem to be Dellienwares fault as I’m seeing posts that MSI and ASUS laptops with Killer are having the same issue.


Sent from my iPhone using Tapatalk

 

Sounds like DNS resolution if you can ping....try setting a static DNS either to 1.1.1.1 or 8.8.8.8 or both could be something to do with DHCP losing your DNS setting when your lease runs out

 

Hi @Maldax , Yes, I did this as well. That was the first thing that came to mind. Made no difference, nothing worked other than a reboot So far. I did hear some anecdotal evidence that deleting the killer suite software solves the problem. I'll be trying that tonight. I did update the driver from Killer's site, but did not blow away the suite software -- not sure that's going to make a difference but I'm willing to try anything.

 

The time it takes for this to occur is very unlikely to be expired leases. Especially since my network is static.

Sent from my Pixel 2 using Tapatalk

 

I touch the probe on the heatsink metal just above the VRM or the VRM itself if accessible from the sides. The probe is around 0.5mm diameter metal ball (a thermocouple) at the end of a thin wire (two wires). It is very fiddly and the laptop is on its side (on the left) and you have to be extra careful else you may short something...

 

Yes, I think that removing the suite definitely helped in my case [EDIT: by help, I mean, I seem to remember the issue happened a lot more frequently at first until I removed that. And now, since it won't download from the Windows Store, I can't even re-install their control center even if I wanted to.]. I am still waiting to see how it holds up. I've been restarting a lot due to overclocking tests so it hasn't been 24 hours yet. Same exact symptoms as you - nothing HTTP would work (two browsers, Dropbox, Store, OneDrive, Steam, etc.), but ping would work.

Killer sent my some reset instructions with netsh to flush the dns cache, advised to use Google's 8.8.8.8 instead of CloudFlare's 1.1.1.1 (which I had already tried but hadn't flushed the cache). I've tested with two different routers I have. Switching connections when the issue occurs doesn't work. Only a reboot remedies it. Killer said if I can ping out, then it's not really their issue, and beyond changing the DNS, they had no other advice, so that's pretty much the end of the road with getting help from them. (They were friendly and helpful, at least; but unfortunately this doesn't seem to be a recognized problem.)

If my connection stays up after the last round of flushing I did, I'll share an update!

 

Thanks for your detailed feedback. This makes more sense then. I was thinking it was some built-in power limiting function, but the fact that it's tied to a different temperature level makes sense. And yep, it seems to happen right at 93 on that one core!

I haven't researched using pads too much; Aristotle said their machine improved just by re-pasting. If you use pads, are you saying that there is no need for ANY paste? Just scrape off the old paste, insert the spongy types of pads, and that's it? (I know, I need to research more, but it's good to know exactly what worked for someone.) I thought the pads sounded more elaborate when I read about them earlier in the thread, but if that circumvents the need to re-paste, it actually sounds easier. (at least I can remove and re-pack if I don't see much of a difference)

Also, there is no need to de-lid on these BGA chips right? (hopefully this is a dumb question as I think it is, but I thought I'd ask to be sure) It doesn't appear there is a heat spreader on them, I mean, in pictures I've seen.

Great advice on the tabs that get worn down with each disassembly too. I felt like even the ones on the bottom cover were different after one removal to install drives.

At least I have some hope again... would really like to see this thing maintain 4.1 at least on all six cores under load and get a 1250+ on Cinebench. Makes sense how this can be thermally related after all.

 

I would not even consider de-lidding the CPU unless you’re out of warranty. With these laptops there’s just too high of a risk of needing Dell to replace something.




Sent from my iPhone using Tapatalk

 

BGA cannot be delidded due to it being delidded already. Only LGA comes with a Heatspreader

 

Cool, thanks for confirming that - I was pretty sure that was the situation.