Setup & Software Discussions
2018 13" Dell XPS 13 9370 [8th,4C,U] + RTX 2080 [email protected] (Razer Core X) + ...
 

2018 13" Dell XPS 13 9370 [8th,4C,U] + RTX 2080 [email protected] (Razer Core X) + Win10 [Jithin Jose] // VR Vive/VivePRO  

 of  2
  RSS

(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

System specs
CPU: Intel i7 8550 U 4C/8T
Ram: 16GB DDR4
Storage: 512GB M.2
Ports: 2x 4lane Thunderbolt 3, 1 USB c 3.0 port.
Laptop display: 1080p 60Hz
external Display: 1080p 60Hz

eGPU hardware

Razer Core X using RTX2080 Ti, Using thunderbolt cable comes with Razor Core X

 

Hardware pictures



 

Installation steps
Pretty much straightforward, Open eGpu >> Place GPU >> connect to laptop TB3 port >> boot system >> install GPU drivers >> Connect  external monitor to GPU

 

Benchmarks 

I am only adding a few benchmarks here, If any of you guys need more let me know.



 

Comments 

Very easy installation and setup process in windows. Because I am a developer, I used to work in Linux systems, For ubuntu installation was not straightforward, but got it working.

 

Gaming Experience: Finally I am able to play my favorite game DCS, a massive upgrade from 10fps to 100+ fps, 😀
CPU Bottleneck? Well, all the games I played kept CPU usage below 90%, I am not an expert here,  I think CPU is not a problem, for now, The main bottleneck is 1080p 60HZ monitor ;).
The curious case of the virtual link: I was very much curious about the virtual link port on RTX GPU. From my experiments, i found out it works like USB c 3.0 port. Now you can connect your mouse dongle to the GPU :), I have also connected External Disk, which also gives 400MB/s transfer rates(probably limited by SSD speed).
2080 Ti does not have a good FPS/$: True, but I care more about ML performance, for which 2080TI Does worth it(especially for Fp16). 

Please let me know if need more details/clarifications, I will answer my best. Also, let me know if any of you guys want to check for extra benchmarks/tests.

This topic was modified 6 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

Got 6769 score in 3DMark RayTracking Benchmarks, Unlike other benchmarks, I can see very high GPU usage.

https://www.3dmark.com/pr/2290

This post was modified 6 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
odin
 odin
(@odin)
Eminent Member
Joined: 12 months ago
 

Your rig completely destroyed my 15" notebook (Eurocom Tornado F5) with a desktop 7700K CPU and GTX 1080 in both TimeSpy and Heaven. I got significantly lower scores.

Can you run these same benchmarks on the internal display with the eGPU?

This post was modified 6 months ago

LG Gram 17 | Sonnet Breakaway Box 550 | Asus Strix RTX 2070 OC Edition | Win 10 Pro 1903 + Manjaro Deepin Dual Boot
Build Link


itsage liked
ReplyQuote
jerry Kansai
(@joi_kansai)
Estimable Member
Joined: 1 year ago
 

Yes you should get new monitor 😛

This post was modified 6 months ago

Razer Blade 15 2018, i7 8750H 16gb ddr4 2666mhz 1tb 970Evo
Razer Blade Stealth 12,5 2016 i7 7500U 8gb ddr3 1866mhz 500gb 970Evo
MacBook Pro 13 mid 2012 i5 3210M 12gb ddr3 1600mhz 500GB HDD
Razer Core V2 rtx 2070 Black 2040mhz Boost clock
Previous setting: rtx 2080ti Asus Turbo, rtx 2080 xc gaming, rtx 2080 Fe, Zotac Mini 1080ti, Evga 1080 FTW2 gaming.
Dual Monitor Predator XB241YU 165Hz 1440p, LG l24UD58 60Hz 4K
Portable Monitor Magedok 1440p HDR 60Hz usbA/C monitor


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

Sure, I have rerun for TimeSpy with and without External monitor, TimeSpy the difference is very low(2%).
First one is using internal monitor(8918) and the Second one is for external monitor(9178)

This post was modified 6 months ago

Dell XPS 13, 9370, RTX 2080 Ti


odin liked
ReplyQuote
odin
 odin
(@odin)
Eminent Member
Joined: 12 months ago
 
Posted by: Jithin Jose

Sure, I have rerun for TimeSpy with and without External monitor, TimeSpy the difference is very low(2%).
First one is using internal monitor(8918) and the Second one is for external monitor(9178)

Wow, that's pretty impressive. Thanks for posting that.

LG Gram 17 | Sonnet Breakaway Box 550 | Asus Strix RTX 2070 OC Edition | Win 10 Pro 1903 + Manjaro Deepin Dual Boot
Build Link


ReplyQuote
zamardii12
(@zamardii12)
New Member
Joined: 7 months ago
 

Hey. Could you do me a very quick favor? Can you just run a standard Firestrike benchmark for me on the external monitor? I'd super appreciate it. I'm trying to decide between this and the Dell XPS 15

Thanks!

Pending: Add my system information and expected eGPU configuration to my signature to give context to my posts


ReplyQuote
noxo
 noxo
(@noxo)
Active Member
Joined: 2 years ago
 

Nice setup, i have Matebook Pro X with same CPU (but 2-lane TB3) and GTX1080. I get only ~75% of your CPU score, in TimeSpy, on external monitor (1080p/60Hz). Is it the 2-lane TB3 becoming bottleneck, causing CPU to wait until data is exchanged with GPU or what is it.. ?

This post was modified 6 months ago

Matebook X Pro + MX150 + GTX 1080 + Zotac AMP Box


ReplyQuote
zamardii12
(@zamardii12)
New Member
Joined: 7 months ago
 

That 2-lane TB3 is most definitely a bottleneck... especially if you aren't running an external monitor.

Pending: Add my system information and expected eGPU configuration to my signature to give context to my posts


ReplyQuote
noxo
 noxo
(@noxo)
Active Member
Joined: 2 years ago
 
Posted by: zamardii12

That 2-lane TB3 is most definitely a bottleneck... especially if you aren't running an external monitor.

Yeah, I know that with internal monitor it requires more bandwidth. But in this case I've external monitor in use. Would nice to know what exactly happens  🙂 I guess, in my case, CPU is not actually overloaded, but it's sitting idle waiting for slow I/O over 2-lane Thunderbolt 3, causing lower CPU score, because the execution of benchmark will be delayed as much as CPU have to wait for data exchange over TB3?

In my opinion TimeSpy CPU benchmark gives sort of "wrong" score for CPU, because it could execute more, but just have to sit idle and wait for TB3.. if that's the case for my low CPU score in here  🙂 

This post was modified 6 months ago

Matebook X Pro + MX150 + GTX 1080 + Zotac AMP Box


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

You can check examples like,  https://www.3dmark.com/spy/4235100 which have my same laptop with same processor & 1080 egpu, The CPU score seems very much similar to mine. So I think 2 lanes might be the bottleneck here. Also, the other affecting factor will be TDP, thermal limits by the system. Also have a check you are experiencing thermal/power/current throttling during the benchmark. For mine, I can see CPU clock stays between 3.2 & 3.9.

This post was modified 6 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

Yup, finally upgraded to 4k IPS monitor. The funny thing is fps drop is very less in games. In DCS I was getting 90-140Hz in 1080p, But in 4K in the same setting, I am getting 70-90fps.

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
Shimmerene
(@shimmerene)
New Member
Joined: 6 months ago
 

I'm curious about the Razer Core X and the RTX (for me, the 2060) GPU. Razer Support couldn't say that the RTX series was supported by Razer Core X, but I'm guessing they work fine based on several builds I've found here. Seems like Razer hasn't updated their tech specs yet? Did you have any issues with the Core X and the RTX 2080 Ti GPU?

Razer Blade Stealth 13 2019 i7-8565U 16GB, Building first eGPU


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@zamardii12, I have done firestrike benchmark, on an external 4k monitor, with a score 17256

https://www.3dmark.com/3dm/33033519

firestrike.17256

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@shimmerene, Well RTX 2080 Ti that I am using also does not include in Razor Core X compatibility list. But it works without any problem. Also, sure RTX 2060 should also work well, the Only problem is Razor Core X you have to pay almost same price as RTX 2060  😀 

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

After upgrade to 4K monitor and GPU table mounting.

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
Eightarmedpet
(@eightarmedpet)
Noble Member
Joined: 2 years ago

ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@eightarmedpetIt's LG 32UD89, 4k, 60fps now i connected though usb c of RTX 2080 Ti

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
Eightarmedpet
(@eightarmedpet)
Noble Member
Joined: 2 years ago

ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: zamardii12

That 2-lane TB3 is most definitely a bottleneck... especially if you aren't running an external monitor.

Exactly when you aren't running an external monitor. With external monitor it's hardly an issue.

Posted by: Jithin Jose

Yup, finally upgraded to 4k IPS monitor. The funny thing is fps drop is very less in games. In DCS I was getting 90-140Hz in 1080p, But in 4K in the same setting, I am getting 70-90fps.

This is not a surprise as the TB3 setup typically limits at high FPS. You won't get 400 fps with your RTX2080Ti even for old games.

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@oliverbWas it TB3 the limit there? 4k 60fps requires same bandwidth as 1080p 240 fps. I think limit might be the CPU.

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: Jithin Jose

@oliverbWas it TB3 the limit there? 4k 60fps requires same bandwidth as 1080p 240 fps. I think limit might be the CPU.

This is just wrong. The bandwidth Host2Device (CPU -> GPU) ist equal for any resolution. You are talking about the bandwidth GPU -> monitor and confusing apples with pears 😉 Only when you display back into the internal display this bandwidth comes into play.

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@oliverb, Thanks for that info, i am just a noobie here :D. I have one question there, if TB3 connection is the bottleneck, change from internal to external display should give significant fps gain, but mostly it is very less(~10%). Does TB3 have separate independent lanes for Device2Host and HostDevice?

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: Jithin Jose

@oliverb, Thanks for that info, i am just a noobie here :D. I have one question there, if TB3 connection is the bottleneck, change from internal to external display should give significant fps gain, but mostly it is very less(~10%). Does TB3 have separate independent lanes for Device2Host and HostDevice?

Actually this 10% is a value that is only valid for certain tests and parameter. If you try to display 4k back into the internal display or Tess with very large FPS, the loss ist much higher. I have made some tests in this regards, I will look for the them
 and post them here. About the data-lanes: Yes, they should be different.

If you look at this comparing internal with external display, it's quite clear that the performance of internal display gets worse with higher resolution compared to external display.

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@oliverbThanks for the link, that is very comprehensive benchmark results. But it does not match with the results I am getting. When we consider only TB3 External display case, You have Avg fps drop from 105 to 64 fps when moving from 1080p to 1440p that is almost than 40% drop. But, For me fps drop is from 110 - 80 when moving from 1080p to 4k which is ~25% even at 4k.

Also if possible can you please do the same benchmark in 4k High settings.

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: Jithin Jose

@oliverbThanks for the link, that is very comprehensive benchmark results. But it does not match with the results I am getting. When we consider only TB3 External display case, You have Avg fps drop from 105 to 64 fps when moving from 1080p to 1440p that is almost than 40% drop. But, For me fps drop is from 110 - 80 when moving from 1080p to 4k which is ~25% even at 4k.

Also if possible can you please do the same benchmark in 4k High settings.

It's very difficult to do comparing benchmarks with 4k Settings with the internal display. There are not many internal displays with 4k resolution. Perhaps with Timespy there is an option to make 4k Benchmarks even though the display is not 4k. I am not sure because I rarely use it. It needs too much time to start.

My question is:
How did you get a 4k Benchmark running in internal display? (So you could derive this 10% loss).

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@oliverb, Hi I am as mentioned in the comments referring to the external display, Not internal. And 10% loss is also happened for external display. not internal.

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: Jithin Jose

@oliverb, Thanks for that info, i am just a noobie here :D. I have one question there, if TB3 connection is the bottleneck, change from internal to external display should give significant fps gain, but mostly it is very less(~10%). Does TB3 have separate independent lanes for Device2Host and HostDevice?

@jithin_jose, didn't you actually write "change from internal to external display"?

So you it looks like you were comparing internal and external display and this is all I did and was referring to. And for this is answer is no: You will not lose just 10% in 4k if you compare external with internal eGPU acceleration. (or "gain" if you compare internal with external, which you wrote about).
It's much more, of course. This is pure numbers and mathematics.

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 
Posted by: OliverB
Posted by: Jithin Jose

@oliverbThanks for the link, that is very comprehensive benchmark results. But it does not match with the results I am getting. When we consider only TB3 External display case, You have Avg fps drop from 105 to 64 fps when moving from 1080p to 1440p that is almost than 40% drop. But, For me fps drop is from 110 - 80 when moving from 1080p to 4k which is ~25% even at 4k.

Also if possible can you please do the same benchmark in 4k High settings.

It's very difficult to do comparing benchmarks with 4k Settings with the internal display. There are not many internal displays with 4k resolution. Perhaps with Timespy there is an option to make 4k Benchmarks even though the display is not 4k. I am not sure because I rarely use it. It needs too much time to start.

My question is:
How did you get a 4k Benchmark running in internal display? (So you could derive this 10% loss).

@oliverb, Okay, confusion happend here, Did't you ask "How did you get a 4k Benchmark running in internal display? (So you could derive this 10% loss)."? Well, where did I say I got a 10% loss on 4k? 10% loss happened when moving from External to Internal display at 1080p. This is nothing related to 4k.

So, I will repeat my question as in my final comment, This is the case for only External display(Don't get confused there ;)). Problem is your benchmark results no way matching with result i am getting in games. Your benchmark result says when changing from 1080p to 1440p fps drop is almost 40%. But from my experience in games, fps drop is only 25% when changing from 1080p to 4k. 

So that is the reason I request you to run the same benchmark in 4k at high quality. Then we can have apple to apple comparison and can prove the benchmark you have run do not represent real-world performance in games.

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 
Posted by: Jithin Jose

 

@oliverb, Okay, confusion happend here, Did't you ask "How did you get a 4k Benchmark running in internal display? (So you could derive this 10% loss)."? Well, where did I say I got a 10% loss on 4k? 10% loss happened when moving from External to Internal display at 1080p. This is nothing related to 4k.

Ah, ok. 10% los for 1080p is totally normal from external to internal display. This loss for internal display raises a lot with higher resolutions.

So, I will repeat my question as in my final comment, This is the case for only External display(Don't get confused there ;)). Problem is your benchmark results no way matching with result i am getting in games. Your benchmark result says when changing from 1080p to 1440p fps drop is almost 40%. But from my experience in games, fps drop is only 25% when changing from 1080p to 4k. 

 

Really, it drops only 25%? Which benchmark? On Unigine benchmarks it's normally significant higher.
Could you please Unigine Superposition benchmark?

1) 1080p Preset High
2) 4k (2160p) Preset High

(Important: Don't use "4k-optimizied" for 4k, the details be the exactly same of course for each resolution). If it really goes down only 25% you may have the best GPU I have ever seen.
You may use Unigine Valley or Heaven as well, if you see it fit.

Just for note, the results of My GTX1080Ti for 1080p (FullHD) and 2160p (4k):

 

  GTX1080Ti FullHD GTX1080Ti 4k
Unigine Heaven
Heaven MBP2018 external

This was the wrong benchmark. It can't be used for comparison.

FPS drops from 122 FPS to under 60 FPS, which is a loss of over 50%. This was to be expected.

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

@oliverb, 25% drop happened in the game, DCS. which I play the most(yes I know it is the same quality setting and same scene). 

But when I run TIMESPY in 1080p and 4k the drop was almost 40%. That's why I believe benchmark we are running is baseless and it completely depends on the game/application we use.

And yes, I will post the Superposition benchmark.

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 

Another important aspect is the fact, that eGPU setups with external monitor perform best on higher resolutions compared to the desktop counterparts. The less FPS are display the less the bandwidth plays a role.

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 
Posted by: OliverB

Just for note, the results of My GTX1080Ti for 1080p (FullHD) and 2160p (4k):

 

  GTX1080Ti FullHD GTX1080Ti  4k
Unigine Heaven
Heaven MBP2018 external

Valley GTX1080Ti 2160p

FPS drops from 122 FPS to under 60 FPS, which is a loss of over 50%. This was to be expected.

@oliverb, it seems you have used two different settings for 1080p and 4k. Anyway, I have used settings same as yours.

 

  GTX2080Ti FullHD GTX2080Ti  4k
Unigine Heaven
Heaven.1080p

Heaven4k

So from these results, FPS drop is almost 35%. 

Comparing US both, you have much better CPU & Mine have a better GPU.
2.1 1080p results have 20% improvement.
2.2 4k results have 57% improvement over 1080 Ti.

That makes me think TB3 not bottlenecking 1080Ti, 2080 TI and the 4k monitor is a good combination for eGPU setup., Pls share your thoughts too.

 

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 

No, I have used the same settings. You have used different ones: "Quality Ultra" on FullHD, and "Quality High" on 4k. You should use the same settings for FullHd and 4k.

Edit: D*mn*d, I have used different setting, too. My fault. I wanted to compare another benchmark, but didn't find it ./

Posted by: Jithin Jose

2.2 4k results have 57% improvement over 1080 Ti.

All the conclusions are unprecise, because you we confused "Quality High" with "Quality Ultra". And me too.. Lol

This post was modified 5 months ago

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
(@jithin_jose)
Eminent Member
Joined: 6 months ago
 

Okay, But it seems I made a good decision by buying 4k 60fps and not 1440p 144hz   😀

This post was modified 5 months ago

Dell XPS 13, 9370, RTX 2080 Ti


ReplyQuote
OliverB
(@oliverb)
Noble Member
Joined: 9 months ago
 

I had the RTX2080Ti, too and let's look for the Heaven Beanchmark ExtremeHD, that is again another benchmark, but very well comparable with the GTX1080Ti

Benchmarks coming:

2018 15" MBP & 2015 13" MBP connected to RTX2080Ti GTX1080Ti GTX1080 Vega56 RX580 R9-290 GTX680


ReplyQuote
 of  2