eGPU Performance Loss - PCI Express vs. Thunderbolt
 
Notifications
Clear all

eGPU Performance Loss - PCI Express vs. Thunderbolt  

 of  10
  RSS

enjoy
(@enjoy)
Reputable Member
Joined: 4 years ago
 
Posted by: vava726

 

An other interesting thing to explore is the performance drop is influenced by architecture (Pascal vs Maxwell).

Is worth to buy an old 980 or a 1060 ? The amount of dram vram is a variable that increases the drop ?

   

I think right now that EVGA GTX 1060 6GB SC is the best GPU (performance and price) for eGPU if you want FULL HD gaming like me. Also all benchmarks shows that with EVGA GTX 1060 6GB SC you can use Internal Display without problems if you have at least Thunderbolt 2 as we have 20% performance drop that is typical for eGPU for External Display. I don't think less VRAM will help you in gaming it will help you only to have less performance drop as you will have less FPS....

MacBook Pro (Retina, 15-inch, Later 2013) 3.2GHz Quad Core Intel i7-4750HQ / 8 GB 1600 MHz DDR3 / 256GB SSD + 1TB
mini eGPUPCI Express vs. ThunderboltMac CAN gameGaming Laptops vs. MacBook Pro with eGPU

 
2012 13" MacBook Air [3rd,2C,U] + GTX 1060 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.12 [build link]  


ReplyQuote
konceptz
(@konceptz)
Active Member
Joined: 3 years ago
 

Regarding the 1080ti tests.  

 

I sent you you a message but if you inspect the heaven benchmark you will see that the desktop implementation should not be considered reference. 

 

There is an over clock on the gpu core and the memory. 

2016 15" MacBook Pro (RP460) [6th,4C,H] + GTX 1080 Ti @ 32Gbps-TB3 (AKiTiO Node) + Win10 [build link]  

ReplyQuote
enjoy
(@enjoy)
Reputable Member
Joined: 4 years ago
 

I add some EXTRA info. It is duel of Hackintosh with 1080TI vs. MacBook Pro 2016 15" with eGPU with 1080TI

Hackintosh with 1080TI (PCIe slot 126Gbps) /Intel Core i7 6700K @4.60 GHz CPU, 16GB DDR4 DRAM/

MacBook Pro (Retina, 15-inch, Late 2016) Thunderbolt3 (32Gbps) External Display:

MacBook Pro (Retina, 15-inch, Later 2013) 3.2GHz Quad Core Intel i7-4750HQ / 8 GB 1600 MHz DDR3 / 256GB SSD + 1TB
mini eGPUPCI Express vs. ThunderboltMac CAN gameGaming Laptops vs. MacBook Pro with eGPU

 
2012 13" MacBook Air [3rd,2C,U] + GTX 1060 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.12 [build link]  


ReplyQuote
ed_co
(@ed_co)
Reputable Member
Joined: 3 years ago
 
Posted by: enjoy

 

I add some EXTRA info. It is duel of Hackintosh with 1080TI vs. MacBook Pro 2016 15" with eGPU with 1080TI

Hackintosh with 1080TI (PCIe slot 126Gbps) /Intel Core i7 6700K @4.60 GHz CPU, 16GB DDR4 DRAM/

MacBook Pro (Retina, 15-inch, Late 2016) Thunderbolt3 (32Gbps) External Display:

   

Huge gap man, more than expected... the % difference is really high in external (35% approx and not 20%). I am a little bit dissapointed. There is anything that can be improve in order to stretch the gap?

2017 15" MacBook Pro (RP560) [7th,4C,H] + GTX 1080 Ti @ 32Gbps-TB3 (Mantiz Venus) + macOS 10.13 & Win10 [build link]  

ReplyQuote
enjoy
(@enjoy)
Reputable Member
Joined: 4 years ago
 
Posted by: ed_co
 
Huge gap man, more than expected... the % difference is really high in external (35% approx and not 20%). I am a little bit dissapointed. There is anything that can be improve in order to stretch the gap?

The Hackintosh CPU is overclocked to 4.60GHz and this made a huge difference for sure...

IMO after all benchmarks and tests - GTX 1060, RX480, RX580 and GTX 1070 are the best GPUs on the market for eGPU, because its just not worth to spend big money for 1080 and 1080Ti and to have huge performance drop....

All benchmarks here are for FULL HD 1080p resolution, you must know that if you up the resolution to 1440p or 4k even on the External Display 1080 & 1080Ti will UP the performance drop with big % .... 

MacBook Pro (Retina, 15-inch, Later 2013) 3.2GHz Quad Core Intel i7-4750HQ / 8 GB 1600 MHz DDR3 / 256GB SSD + 1TB
mini eGPUPCI Express vs. ThunderboltMac CAN gameGaming Laptops vs. MacBook Pro with eGPU

 
2012 13" MacBook Air [3rd,2C,U] + GTX 1060 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.12 [build link]  


ReplyQuote
gtosi
(@gtosi)
Eminent Member
Joined: 3 years ago
 

 Maybe we should not compare the same graphics card in different machines. A quick search shows that for the same desktop graphics card it is normal to see up to 20-30% differences in Unigine performance depending on the rest of the computer (especially CPU). (e.g.  https://www.techpowerup.com/forums/threads/unigine-heaven-4-0-benchmark-scores.198888/)  

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
ed_co
(@ed_co)
Reputable Member
Joined: 3 years ago
 
Posted by: enjoy

 

Posted by: ed_co
 
Huge gap man, more than expected... the % difference is really high in external (35% approx and not 20%). I am a little bit dissapointed. There is anything that can be improve in order to stretch the gap?

The Hackintosh CPU is overclocked to 4.60GHz and this made a huge difference for sure...

IMO after all benchmarks and tests - GTX 1060, RX480, RX580 and GTX 1070 are the best GPUs on the market for eGPU, because its just not worth to spend big money for 1080 and 1080Ti and to have huge performance drop....

All benchmarks here are for FULL HD 1080p resolution, you must know that if you up the resolution to 1440p or 4k even on the External Display 1080 & 1080Ti will UP the performance drop with big % .... 

   

I see. Nice argument. Should be great to analize it as well for 1440p and 4k resolutions, in order to see, what is the best price/performance and is worth it...

2017 15" MacBook Pro (RP560) [7th,4C,H] + GTX 1080 Ti @ 32Gbps-TB3 (Mantiz Venus) + macOS 10.13 & Win10 [build link]  

ReplyQuote
enjoy
(@enjoy)
Reputable Member
Joined: 4 years ago
 
Posted by: ed_co
 
I see. Nice argument. Should be great to analize it as well for 1440p and 4k resolutions, in order to see, what is the best price/performance and is worth it...

We have some benchmark with Razer Blade (TB3) with dGPU 1060 feat. eGPU with GTX 1080 (you can see that on Internal Display GTX 1080 in eGPU shows the same results as the dGPU 1060 inside Razer Blade... on External Display GTX 1080 shows not bad results at all but also shows us 20% performance drop vs. if the GTX 1080 is in Desktop PC :

Rise of Tomb Rider 2k and 4k in eGPU:

 

Rise of Tomb Rider 4k in Desktop PC:

 

Rise of Tomb Rider 1440p (3k) in Desktop PC | Rise of Tomb Rider 1080 (2k) in Desktop PC:

 

MacBook Pro (Retina, 15-inch, Later 2013) 3.2GHz Quad Core Intel i7-4750HQ / 8 GB 1600 MHz DDR3 / 256GB SSD + 1TB
mini eGPUPCI Express vs. ThunderboltMac CAN gameGaming Laptops vs. MacBook Pro with eGPU

 
2012 13" MacBook Air [3rd,2C,U] + GTX 1060 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.12 [build link]  


ReplyQuote
P-Mac
(@p-mac)
Trusted Member
Joined: 3 years ago
 
Posted by: enjoy
 
you must know that if you up the resolution to 1440p or 4k even on the External Display 1080 & 1080Ti will UP the performance drop with big % .... 

It's actually the opposite...the higher the resolution and the more stressed the card is, the smaller the performance gap becomes compared to the same card installed in a desktop. The bottleneck starts to shift from PCIe bandwidth back to raw GPU performance.

In your Tomb Raider benchmarks above, at 1440p, the desktop's 1080 performs 24% faster than the eGPU one, but at 4K, the gap narrows to 22%. 

I use a TITAN Xp (2017) in my Thunder2 along with a 1080p 144Hz monitor, and I've noticed similar quirks; the Thunderbolt connection becomes a significant bottleneck when trying to achieve 144fps average (by slightly lowering graphics settings), and is made apparent by the GPU utilization being below 99% and the framerate not reaching 144, despite not having any reason to be lagging (CPU isn't throttling, etc.)

For a few games, I've targeted 60fps and simply raised image quality as high as it could go, even invoking DSR in games that don't have a native supersampling setting. I have a lot of my games rendering to 3K, sometimes 4K resolutions to then be downsampled to the 1080p monitor...not only does this look crispy as hell, it reduces the apparent performance gap when compared to the card in a desktop. 

Another case in point, in Unigine Superposition (1080P extreme preset...a very compute-heavy benchmark), my TXp eGPU performs within 7-10% of desktop TXp results. 

Food for thought. 

MAG341CQ 34" Ultrawide // Logitech G900 & HyperX Alloy FPS Pro (MX Blues) // Scarlett 2i4 + Yamaha HS7

 
2016 15" MacBook Pro (RP460) [6th,4C,H] + RX Vega 64 LC @ 32Gbps-TB3 (AKiTiO Node) + macOS 10.14 & Win10 [build link]  


itsage liked
ReplyQuote
nkh004
(@nkh004)
Eminent Member
Joined: 4 years ago
 

This is probably a dumb questions.

Can you utilize both TB ports at the same time? TB2 box has two ports, as does my iMac

2011 27" iMac (HD6970M) [2nd,4C,Q] + GTX 1060 @ 10Gbps-TB1 (AKiTiO Thunder2) + macOS 10.12.4 & Win10 [build link]  

ReplyQuote
 of  10