2017 14-inch Lenovo Yoga 920 + TitanV@32Gbps-TB3 (Sonnet Breakaway 550) + Win10 [Nova Nerd]^  


Active Member
Joined:3 months  ago
Posts: 5
December 12, 2017 3:49 am  

Because I can:




Out of the box with no tweaking, iGPU off to squeeze every bit of power. Again I shamelessly pitch the yoga 920 as a plug n play laptop with zero thunderbolt implementation caveats. In a perfect world msi would have keep updating the x16 laptop dock and other OEMs given their shots at it but oh well.

Its going into my “gaming workstation” in the end but I just had to run some tests for science! (Mining tests next…for reasons) Of course its bottle-necked but still one hell of a gaming laptop.

Edited: 2 months  ago

3RYL, Robert Evans, goalque and 2 people liked
New Member
Joined:2 months  ago
Posts: 1
December 12, 2017 6:19 pm  

This is insane, can you test some machine learning, deep learning benchmarks ? How much is the performance bottle-necked comparing to Desktop setup?

Honorable Member Moderator
Joined:1 year  ago
Posts: 600
December 12, 2017 7:28 pm  

This is hilarious. Like sticking a V8 into a VW Bug.  :mrgreen: 

Thanks for the post: For Science!

Posted by: Ivan Georgiev

This is insane, can you test some machine learning, deep learning benchmarks ? How much is the performance bottle-necked comparing to Desktop setup?

At least for the 3DMark benchmarks, you can answer this yourself. Go into the 3DMark result database, and select a 8600K or any other CPU alongside the Titan V, then compare the results with the ones Nova Nerd linked to.

For example, here is one of the top Titan V results for Time Spy: i9-7980XE+Titan V.

Nova Nerd gets 66.8 FPS in GPU Test 1, 57.39 FPS in Test 2. The linked system gets 94.56 FPS in GPU Test 1, 84.77 FPS in GPU Test 2. So the bottleneck is ~30%.

If you look at the Fire Strike Ultra, the top result for a Titan V is an AMD Threadripper 1950X with a Titan V.

Nova Nerd gets 41.03 FPS in GPU Test 1, 23.19 FPS in GPU Test 2. The AMD desktop gets 53.71 and 29.3 FPS, respectively. The bottleneck here is around 20-25%. The high resolution draws them closer, but it is still clear the Titan V is  overkill for a 8550U (as pretty much expected). Something to keep in mind, though. The top result is really the highest-end of results and is typically the result of a highly overclocked and fine tuned system, benchmarked for records/bragging rights. A more “fair” comparison is probably 5% lower or so, giving us a bottleneck that is closer to 15-20% at high resolutions. Considering what we are comparing here (a 15W mobile CPU, and a 150W+ desktop monster), the 8550U is pretty damn impressive.


Edited: 2 months  ago

"Dire Wolf II" - HP ZBook 15 G4: Core i7-7820HQ, 32GB, M1200 dGPU, R9 Fury@32Gbps-TB3 (Mantiz Venus) eGPU, HP Z27q (5K) + Dell P2715Q (4K)
"Timber Wolf" - HP Spectre x360: Core i7-8550U, 16GB, GTX1050Ti@16Gbps-TB3>TB2 (AKiTiO Thunder2), Dell U3011
"Stormcrow" - Lenovo T430s: Core i7-3520M, 8GB, GTX750Ti@10Gbps-TB1 (AKiTiO Thunder2) eGPU, Sony 4K TV
"Phoenix Hawk" - Intel NUC 33217CK: Core i3-3217U, 8GB, GTX670@4Gbps-mPCIe2 (PCE164P-N03) + Linux Mint 18.2 (64-bit)

"Always listen to experts. They'll tell you what can't be done, and why. Then do it."- Robert A. Heinlein, "Time Enough for Love."

theitsage liked
New Member
Joined:2 months  ago
Posts: 3
January 2, 2018 4:44 am  

When you unplug the eGPU, do you need to restart?

Eminent Member
Joined:2 months  ago
Posts: 29
January 3, 2018 2:25 am  

3DMark scores with 1080 in this thread


Active Member
Joined:3 months  ago
Posts: 7
January 4, 2018 10:26 am  

Does your Breakaway Box always charge the battery on your Yoga 920? On mine the power delivery most of the times just maintains the actual charge level  of the battery without charging the battery. But on a few occations it actually charged my battery to 100% (from 80%). Very strange… 



Please Login or Register