which perform better between GEFORCE 940mx/1080ti and JETSON TX2?

Accelerated Computing CUDA CUDA Programming and Performance
1

a same algrithm run in different gpu,there is the tests i had done:

GEFORCE 1080ti[3854 cores,1582MHZ]:time slaps 200ms;
GEFORCE 940MX[384cores,1006MHZ]:time slaps 2890ms;

2890 / 200 = 38541582 / （3841006） may be sure.

Then,I need to run this algrithm in JETSON TX2[256 cores ,cannot find its frequency],the time[LET IT BE Tj] slaps will satisfy this :

Tj / 200 = 384 / 256

Q:the way i get the time slaps in JETSON TX2 is right? If not,how should i get it ?

PLus:if Tj is surely more than 200ms,it will be useless for me to run this algrithm in JETSON TX2.

2

I wouldn’t expect a TX2 to be as fast as a 1080Ti, running a CUDA kernel.