Razer Blade 2018 (GTX 1070) + egpu + RTX 2080ti + 240hz FHD

Hello everyone, as I got a new job with a pretty good salary, I decided to upgrade my existing rig; I currently own a Razer Blade 2018 (i7 8750H & Gtx 1070) and a FHD 144hz 1ms monitor.

I was considering getting a Razer Core v2 to plug my mouse and keyboard in and have it, together with a 2080ti, pass the signal to a new FHD 240hz screen.

Would that be enough to have something as close as possible to 240fps in Apex, Destiny 2, PUBG, Fortnite, Overwatch and so on?

I'm planning to to run them at lowest settings and, if required, underscaled textures just for the sake of competitive bonus frames per second.

I know that egpu would bottleneck, but 6C 12T are probably going to keep things a bit balanced, aren't they?

Does anyone have the chance to test this for me? I can only find 4k data.

Also, considering how cooler the notebook will run, I might as well force the CPU to stay on turbo 24/7.

Sorry for asking something people might have already talked about, but it has been hard for me to gather information on FHD 240hz egpu rigs.

Thank you in advance!

 

The TB3 interface will severely bottleneck the 2080ti to a point where it is no use. Dont waste your money on it. The 6core CPU doesnt matter, the performance can just not be called upon in games.

 

A 2080ti would eat Apex min settings at breakfast tho, so how many FPS should I expect through a TB3 is the real question.

Thanks for your view on the matter by the way

 

Its exactly at minimum settings where the bottleneck starts to appear.

A 2080TI barely performs betters than a laptop 1080 in most cases and in some others it performs even less.


In theory the bottleneck shouldnt be as large since it is basically a PCI express 3.0 4X slo.Which should give the performance mentioned here https://www.techpowerup.com/review/nvidia-geforce-rtx-2080-ti-pci-express-scaling/6.html

But the TB3 port is using PCH lanes and not direct CPU lanes which causes a huge performance drop and there is also the TB3 communication overhead. So in the end you pay a solid 1000~1200USD for an extremely measly upgrade compared to the internal 1070MQ that you have now.

 

Yeah TB3 e-GPU setups aint where it should be yet, it's for folks who have MX150s and 1050's in their laptops. You should just sell that laptop, and with the extra money, instead of a core v2/2080 ti, buy an actual rtx2080 rb15 2019 edition advanced model. the performance boost will be MUCH better. also the 9750h is a great CPU.

I am seriously hoping a new revision of TB3 comes out with PCIe 4.0 lanes that actually dedicates directly to the CPU lanes instead of this round about B.S. they have now. Totally worthless in more powerful laptops, or at the very least, not even close to worth it with the cost ratio.

 

This. Pushing 240 fps over Thunderbolt3 just isn't going to happen, even in CSGO at 1080p. When you crank graphics settings to max and go for the highest resolution possible, that's when the bottleneck starts to minimize. This is because you're now in a low enough FPS range that you can push all that data through the Thunderbolt line.

 

High Resolution gaming is what benefit you are going to get out of an EGPU and not huge FPS numbers. The issue is people like me don't exist that much anymore, I like the highest quality a game can offer (4K, Max Settings, 60FPS) although it would surely help you in this regard with an EGPU the issue is you are wanting to play games at low resolution with huge FPS numbers which drains the bandwidth that TB3 offers. Hope this gives a bit more clarity as to why you are seeing 4K testing rather than lower resolution.

Sadly you are wrong the EGPU setup he is talking about would be a drastic improvement over a 1070 Max Q, I know first hand as my girlfriend uses one with her MSI laptop and when paired with an RTX 2080 TI it gets double the FPS and graphical fidelity in FFXIV at 4k then with her 1070 Max Q.

 

The numbers dont lie, they are visible in the video. You are presenting an usecase which is the only usable use case for a TB3 egpu now (the alienware solution doesnt have this downgrade). Just as you stated as well, the bottleneck is the largest at high cpu loads.

Apart from that some rare games arent affected by the bottleneck but the large majority are. So sadly you are leaving that part out of the equation. Youtube is filled with similar videos as I have posted earlier and i have my fair share of experiences with egpus as well. It is not just with high framerate low res gaming. 1440P high settings is already enough to show the problems.

Wouldnt surprise me that you have fairly okay results as well because the laptop you are talking about (1070mq must be of the GS series) is thermally constrained so by offloading heat it starts to run better as well nerxt to that 1 odd game that you mention.In general use, a 2080ti just doesnt perform in TB3 and thats a fact.

A nice discussion surrounding the performance drops can be found here
http://forum.notebookreview.com/threads/clevo-egpu-results.825412/

 

You are missing a huge point you lose an extreme amount of performance when you also use the internal monitor. I have my girlfriend setup using an external monitor more like a desktop setup thus performance loss is in the 15-25% range as we are not dealing with internal GPU overhead. Although you have a desire to prove me wrong the facts are at 4K the performance of a 1070 Max Q can never attain results that are playable at max settings. Regardless this is not what the OP wanted, the OP wanted max FPS at lower resolution which we both agree doesn't work with TB3.

RTX 2080 Super 24% Performance Penalty over TB3
Evidence: NoteBookCheck


Laughable GTX 1070 Max Q Performance a full 35.59% worse than above however it is running in spec not as you said "overheating".
Evidence: NoteBookCheck


As I said above your claim that a 1070 Max Q would perform the same as an RTX 2080 TI let alone albeit a weaker RTX 2080 Super is false.

Also this is worst case scenario as this is Fire Strike and it is not in 4K and we are goign to see the worst performance hit in this benchmark our scores only improve in Fire Strike Extreme due to lower bandwidth requirements!

 

If you know my history on thi sforum you know that I have run an external eGPU setup for years. I am well known with the limitation. Firestrike is not an indication of performance fo eGPU's because it simpply doesnt bottleneck and flatters the performance results.

Again I posted a video with real life results in situations not focused on getting the highest FPS possible on low settings.

I dont rhave my eGPU setup anymore or else I would have made recordings specially for you to see. BUt I ditched it because poor performance.

 

Not having it anymore would explain why you are not seeing current performance numbers that I’m seeing regardless a 35% uplift over an integrated graphics card is astounding and as I said I’ve got real world numbers in games she plays as well. I understand it’s a bottleneck but to say it’s the same performance as a max q of an earlier generation is silly :\.

Your statement on fire strike is also off base it’s a good determination and shows a 25% difference off what it should be getting which is huge but it also shows that it’s better than a 1070 Max Q which it is unquestionably.

Also If you’ve seen my other posts that RTX 2080 Super is doing worse in the Fire Strike test then my heavily modded RTX 2080 Max Q in my 2019 Razer Blade. But in games the RTX 2080 Super is getting higher FPS numbers in 4K. Which means the bandwidth is limiting the score in fire strike not the FPS results at high resolution due to lower bandwidth requirements!

Sent from my iPhone using Tapatalk

 

How is it silly, it is shown. In normal circumstances a 2080ti is a solid twice as quick. But thanks to bandwidth restrictions, PCH latency, protocol overhead you already had a 35% drop in even non bandwidth restricting titles including Firestrike. In most real life titles this drop is far higher. You claim that firestrike bottlenecks, i beg to differ and so do reviewers. The Firestrike scores are always closest to the scores compared to the same cards in desktops.

But dont take my word for it. As you can see below on the highest settings you already see the TB3 solution with a 2080ti only 18 frames in front of the 1070MQ, on lower settings this difference disappears almost as expected and said by both of us. But those 17 fps more while the potential even in the AGA is much higher shows the large drop that I described. This is a DX12 title so it is making far less calls compared to DX11 which flatters the TB3 a bit. I cannot find any benchmarks that show games such as Assassins Creed origins and Odyssey , Division 2 and other recent titles. But during my own testing, the internal 1070 of my Alienware was constantly having the same framerate on 1440P high settings as the 2080ti Aorus TB3 box. This is why I returned it and to keep my money in my pocket until I will buy a desktop again. Some games such as CS GO do not show a drop though, which in that case would show the potential of a TB3 dock if it wouldnt have those limits in more modern games.

https://imgur.com/elzjxa1

Unfortunately reviews do not test the heavier games with the TB3 setups, users report TB3 2080ti scores of around 18000 graphics score, that is in the ballpark of a 1070 non max Q laptop GPU.

 

TB3 Egpu setting isn’t great on high FPS gaming, it’s due max PCIe3x4 lanes bandwidth. The gap with desktop is less on high resolution so for 4K gaming I would say 2080ti great pairing with Blade 15, but for 1440 or 1080p maybe the price for performance isn’t great at least from my experience. Other things to note there’s also some titles that don’t run well with TB3 egpu like a lot latest title from Ubisoft, there’s a lot frame drops on some areas, I believe it’s due drm layers, some steam titles are also included, i noticed when pubg first launch it was really laggy but now it runs great. Most laptops with dgpu involves pch controller on data bandwidth that comes from gpu, this reduces also a bit performance on graphics side unlike ultrabook TB3 configuration that goes straight to cpu without involving pch controller, MacBook 15 and Alienware TB3 are apparently rare laptops that its TB3 has not involving pch and goes straight to cpu like ultrabook TB3.
Gaming temperature improvement however is great on TB3 egpu setting. I see up to 20 degrees cpu thermal improvement, and i don’t mind set cpu on its max tdp and OC max the gpu in the enclosure on this setting unlike playing with Blade dgpu.
Temperature difference playing around hour AC Oddysey.
Corev2 2080 OC 160/1000 cpu gaming mode (45 watt)


Blade 15dgpu no OC cpu balance mode (25 watt)


If you need performance comparison number 1070dgpu vs 2080ti TB3 egpu.
https://www.3dmark.com/compare/fs/16524814/fs/16153079

 

Its closer to a 2.0 x4 slot with overhead considered