Comparing M.2 vs TB3 performance – 2015 15″ Dell Precision 7510 (Q M1000M) [6th,4C,H] + GTX 1080 Ti @ 32Gbps-M.2 (ADT-Link R43SG) + Win10 1803

Featured Builds 34 Comments

There is no denying that a TB3 eGPU can dramatically improve application processing speed as can be found by many examples on this forum. Gaming on the side then being an extra bonus.

Question then, if your primary use case is gaming is a TB3 eGPU worth it? In this build we compare performance between a M.2 versus TB3 eGPU attached to a single system (identical specs, OS, system build) to help answer that question. The M.2 eGPU being a maximum performance reference before Thunderbolt 3 transport layer encoding is applied.



System specs (model inc screen size, CPU, iGPU, dGPU, operating system user which eGPU is being used)

  • 2015 15″ Dell Precision 7510
  • Intel i7-6820HQ (max 3.6Ghz 4-core with unlocked multis)
  • 16GB RAM (2x8GB, dual channel)
  • Intel HD530 iGPU + Nvidia Quadro M1000M dGPU
  • 2.5″ SATA drive bay for primary boot drive, though OS build used was booted off a WinToGo USB NVME SSD
  • M.2 (NVME SSD) slot is free for M.2 eGPU use
  • Windows 10 Professional 1803


eGPU hardware (eGPU enclosure, video card, any third-party TB3 cable, any custom mods)

  • ASUS Nvidia GTX 1080 Ti Turbo, flashed with EVGA FTW3 vbios. Has 6P + 8P PCIe power.
  • ADT-Link R43SG (50cm) M.2 eGPU adapter
  • 2 x 220W Dell DA-2 AC adapters 1st powers 75W slot + 75W 6P with cable (1). 2nd powers 150W 8P using cable (2)
    – cable (1) a 8P to 6P+8P power splitter cable supplied with the ADT-Link R43SG
    – cable (2) a 6P to 8P PCIe extender with power-on split-pin/paper-clip as shown here and here
  • M.2 protect board – protects M.2 slot being damaged from repeated insertions


To use the ADT-Link R43SG as a TB3 eGPU add the following, making it equivalent to a ADT-Link R43SG-TB3 kit:


Hardware pictures (note: require 5+ posts for “attach files” button to appear. Or can link from say )

underside cabling

GPU power cabling

Left: Connecting ADT-Link R43SG as a TB3 eGPU using JEYI Leidian NVME-to-TB3 enclosure, Right: It’s TB3 identification


Installation steps (what did you do to get it all going?)

The M.2 and TB3 eGPUs are functionally equivalent in use aspects apart from the latter supporting hotplugging and of course performance discussed in the Benchmarks section below.

Nvidia Optimus accelerated internal LCD mode engages for both. Neither interface saw any system instability.

If considering Hackintoshing, a M.2 eGPU is detected and appears on the PCIe BUS and with correct drivers can be made to work. Eg: this GTX 1080 Ti has macOS 10.13.6 or earlier Nvidia web drivers. A TB3 eGPU isn’t so easy. TB3 on a Hackintosh requires SSDT/DSDT mods to emulate the Mac’s native TB3 interface as has been done on Dell XPS 95xx systems. While TB3 peripherals work there I’ve yet to see working TB3 eGPU.


Benchmarks (all at FHD 1920×1080 )

Dell Precision 7510M2-extTB3-extdifference M2-intTB3-intdifference
Forza Horizon 410570-33.3%9064-28.9%
Firestrike (gpu)2819621752-22.9%2425418239-24.8%
Far Cry 59389-4.3%8678-9.3%
AIDA64 write (MB/s)28862160-25.2%
hwinfo64 PCIe portport 9port 5


The first two results show a > -22% difference. Why?

Even though this site refers to TB3 as 32Gbps-TB3 (it’s electrical link), Intel does disclose the TB3<->TB3 link is 22Gbps. So then let’s gather those same benchmarks with greatest difference and run them on a 16Gbps-M2 interface. We’d expect then “22Gbps” TB3 to outperform 16Gbps-M2.


Dell Precision 751032Gbps-M232Gbps-TB3difference 16Gbps-M232Gbps-TB3difference
Forza Horizon 410570-33.3%7470-5.4%
Firestrike (gpu)2819621752-22.9%2703421752-19.5%
AIDA64 write (MB/s)28862160-25.2%14612160+47.8%
hwinfo64 PCIe portport 9port 5port9port 5


Performance Analysis

Forza 4 has a significant 33.3% decrease in FPS, beyond what we’d expect given TB3’s 25.2% decrease in AIDA64 bandwidth (H2D). Oddly, even a 16Gbps-M2 interface with supposedly less bandwidth outperforms TB3. TB3 should be just an encode/decoding PCIe transport pair so why are we seeing such a performance decrease over TB3?

We find clues as to why by using the bandwidthTest.exe tool included DaVinci Resolve, running commandline below while with the eGPU connected on a 32Gbps-M2, 32Gbps-TB3 and 16Gbps-M2 interface.

C:\Program Files\Blackmagic Design\DaVinci Resolve\bandwidthTest.exe --htod --mode=shmoo --csv > out.csv

We then review the gathered bandwidth versus block size information where we see:


  1. TB3 reduces the 32Gbps PCIe bandwidth anywhere from 25% to 63.5%.
  2. The greatest reduction occurs on small block sizes. 2kb seeing the greatest 63.5% reduction.
  3. TB3 is outperformed by 16Gbps-M2 up to 8kb block size, only matching performance at 16kb block size and outperforming it thereafter.
  4. It appears that Forza 4 is bandwidth bound and utilizes small block transfer sizes.
  5. The min 25% reduction occurs at a block sizes of 200kb and greater -> TB3 is performance optimized for large block sizes.


Comments (eg: how has the eGPU improved your workflow or gaming)

Is a TB3 eGPU worth if for gaming? We’ve shown the additional TB3 transport layer decreases x4 3.0 32Gbps bandwidth anywhere from 25% to as much as 63.5% with bandwidth bound apps/games registering this performance reduction. . The reference 32Gbps-M2 interface itself is 4 times less bandwidth than an Intel desktop.

Is the problem simply the TB3 controller being clocked too slow? Maybe. We do see that TB3 is optimized for large block sizes as used for data transfer on SSDs. Intel certainly have plenty of room to improve this small block transfer performance in TB4.

Q: How to maximize gaming performance on a notebook?

From this build and performance analysis we can suggest:

  • seek a notebook with a decent dGPU
  • seek a candidate system offering a factory direct PCIe eGPU interface like a Alienware Graphics Amplifier port
  • cobble together a eGPU using the NVME SSD’s M.2 slot, also a direct PCIe eGPU interface, like shown in this build
  • obtain a TB3 eGPU now knowing it’s performance limitations on bandwidth-bound games/apps


What are the pros and cons of a M.2 eGPU over Thunderbolt 3 (TB3)?

M.2 eGPU pros over TB3:

1. gets full x4 3.0 bandwidth – TB3 has 25%-64% less H2D/write bandwidth than M.2

2. noticably lower cost for eGPU adapter hardware

3. being able to utilize the eGPU in a macOS Hackintoshed system

M.2 eGPU cons over TB3:

1. require a candidate notebook with dual-storage (M.2 + M.2 or M.2 + SATA) to be viable (use M.2 slot for the eGPU and other storage for boot OS)

2. requires underside eGPU cabling to the internal M.2 port

3. requires DIY tweaking/testing to confirm compatibility


Related Articles



Share this Post

Notify of
Newest Most Voted
Inline Feedbacks
View all comments

Isn’t Forza 4 known for being a game that runs poorly on egpus? Or have I made that up?



Thank you for this build. This is exceptionally insightful (and significant) and finally we have some more understanding of ‘where’ performance is lost. Added to my permanent reading list.


@nando4, Thank you very much for your research! I guess I am subject to the same perfomance degradation with my i7-9750h coming from a i7-7700h.

I documented the issue here:

You will find my system specs here: builds/2019-15-lenovo-yoga-c940-gtx1650-q-max-i7-9750h-6c-12t-rtx-2070-super-32gbps-tb3-razer-core-x-chroma-win-10-pro/

Have you allready contacted Intel in regard to the suspected BIOs issue?!

Best regards!




Posted by: @nando4
3. Unoptimized BIOS on non-Mac “H” 8th/9th gen CPU systems leading to a further 15% performance degradation. M.2 eGPUs are also affected. A system BIOS type patch is required as a fix. 

Hi nando4!
Do you know if 10th gen “h” processors are doing better in this regard?!


@nando4 Is there a patch that can be applied for unoptimized BIOS on non-Mac “H” 8th/9th gen CPU systems? I myself own a dell xps 9560 and havent known about this bios issue. How did you find out? 


Posted by: @nando4 Posted by: @Donwey @nando4 Is there a patch that can be applied for unoptimized BIOS .. I myself own a dell xps 9560 and havent known about this bios issue. Dell XPS 9560 is 7th gen, eg: i7-7700HQ so is not applicable. If you have a 56Wh-battery XPS 9560 with slot for 2.5″ SSD, then can get a tremendous eGPU boost by using the available NVME SSD M.2 slot to host the eGPU using an ADT-Link R43SG. Just like the opening post build. Thank you, i have already ordered ADT-Link R43SG. I will create a new build… Read more »



Thank you for elaborating. I was aware of (2) being an issue previously. Regarding (3), was there a recent system you came across/or on this forum where BIOS updates made the difference? That’s another (important) variable it seems. Do we have before/after BIOS update results which I may have missed? 15% performance difference is significant. 


One method to reduce performance loss of TB3 eGPU vs M.2 is to make the eGPU do more work. Use a higher resolution to slow down the frame rate which will require less data to be sent over TB3. There is a sweet spot of resolution vs framerate that differs between TB3 and M.2 because of this latency difference. This will not be as useful for games that send more data when higher resolutions are used. The USB4 spec describes a latency/bandwidth tradeoff for PCIe (chapter 11, page 422) and USB3 (chapter 9, page 330) traffic which may be applicable… Read more »


Posted by: @nando4 Performance Analysis Forza 4 has a significant 33.3% decrease in FPS, beyond what we’d expect given TB3’s 25.2% decrease in AIDA64 bandwidth (H2D). Oddly, even a 16Gbps-M2 interface with supposedly less bandwidth outperforms TB3. TB3 should be just an encode/decoding PCIe transport pair so why are we seeing such a performance decrease over TB3? We find clues as to why by using the bandwidthTest.exe tool included DaVinci Resolve, running commandline below while with the eGPU connected on a 32Gbps-M2, 32Gbps-TB3 and 16Gbps-M2 interface. C:Program FilesBlackmagic DesignDaVinci ResolvebandwidthTest.exe --htod --mode=shmoo --csv > out.csv We then review the gathered… Read more »


Hi nando4. I have the R43SG with a reference RX5700xt on the spare M.2 port of my HP Zbook studio G3. However, after unplugging the M.2 cable to take my laptop elsewhere, I plugged it back in today but the card isn’t being detected. I had this same problem last week but unplugging the battery and plugging it back in worked. In another instance, I just held down the power button to shut the laptop down and when I booted it back up, the card was working. In my current situation, I have tried the methods I just listed but… Read more »


@nando4, Protecting the M.2 slot and contact pins is a great idea. This would make the ADT-Link R43SG more dockable. At the moment it’s not very convenient to switch from stationary to mobile use.



I can’t see how they can make the connector at the same level because the post at 80mm on the laptop motherboard would interfere. You could only connect adapters 60mm or less. The R4xx adaptors can be reduced manually to 60mm, but 80mm SSDs cannot be reduced. So if you don’t care about connecting SSDs to the protector adapter, and don’t mind reducing your 80mm adapters to 60mm, then ADT could make the protector adapter 20mm long. Adhesive might be required to keep the adapter in place as you connect M.2 devices to it. ADT might have to create their own M.2… Read more »


I downloaded the whole DaVinci Resolve v16.0 1.5GB almost and found that bandwidthtest.exe could be used seperately and it is only 200KB.
So can you post the bandwidthtest tool in this post? Or could it cause some legal issues?

Joaquin Lobos

@nando4, How did you manage to deal with the cable? Did you leave the laptop without its cover?


Hi i want to try this on my new Dell Precission 7710 with 6820hq i7, 32GB, ssd sata 3, quadro M3000M.
I have to do the same steps?
I want to use M.2 M PCI-E Version PCIe PCI-E V9.5 EXP GDC Laptop Docking Station,
It will be good?
and as GPU 980Ti
I will buy a 1080Ti later but as i remember form past 2 years 10th and 20th generation doesn’t work so good with eGPU on laptops. Is still the same or now the drivers and “error 43” better?

Thank you.

Alvin Dmitri

Hey @nando4 , I’m relatively new to building computers here so pardon my ignorance. I’m currently trying to set up an eGPU with Nvidia Quadro RTX4000 for engineering purposes, coupled with a laptop that has a basic geforce card for light gaming. I am quite against the idea of an underside cabling to the internal M.2 port as that ruins the warranty of the laptop, so I’m going to try to opt for the thunderbolt option instead. I’d go with the former if I’m unable to find a laptop with thunderbolt and 32gb RAM under 1100USD. I am quite unsure… Read more »


Hello guys, first of all I would like to say that’s great to see a comunity involved on eGPU with so many ppl helping each other. You all rock! I have an notebook Dell G3 with an i7 9750h processor and a dGPU GTX1660ti max-q. Its a nice dGPU and run almost all games in high/ultra in my notebook screen (it’s a pretty decent graphic card for 1080p).  The thing is that I’m also using this notebook for VR gaming and that’s way gpu demanding. So I was wondering about a possibility to upgrade my performance with an eGPU. So… Read more »


Posted by: @nando4
@phila-delphia , consider running Forza 4 on your H system + eGPU, the compare against running it on a high-end U system + eGPU and compare FPS results to get your answer.

Thank you! I do not own a 10th gen cpu. I just was wondering if the problem you decribed about the unoptimized 8th/9th gen CPU bios still remains with the actual cpus
Best regards

@Nando if i tried an mini PCIe adapter in ALL IN ONE Lenovo C540 with 3rd gen intel cpu i7 3770s will it work ???

@Nando Hi i hope u answer my question can i use an HP probook 650G2 with i7 6820HQ but the problem this laptop come with just 1:M.2 port THEN CAN I USE SSD 2.5″ CADDY BAY IN PLACE OF OPTICAL DRIVE TO BYPASS THIS PROBLEM ?

Aaron Miller

Awesome info guys, exactly what I was looking for. Thanks a ton!

@lwf,  thanks for the reply. I don’t think it’s been bottlenecked by the CPU. 10875H is a 8 core 16 thread CPU, very powerful. And as I said the external 3070 is performing worse than the on board 2060. If it’s the CPU at least the performance of the two GPUs should be the same.