View Single Post
Old 18-06-2019, 04:14 PM   #5139
Tsurugi_Matsuda
Member
 
Join Date: Feb 2010
Posts: 142
Except the laptop gpu variants are still downclocked to be much slower than their desktop counterparts to conserve temps.

Imo nvidia only changed the naming, but is still very conservative on giving juice to their laptop cards
Well that only is applicable for the 1060/70/80 cards, but it's the GP107-based ones (1050/1050Ti/1050 3GB) that the Laptop Editions are made more powerful than the Desktop ones (higher clock frequency) due to a smaller die (14nm Samsung for GP107 VS 16nm TSMC for GP104/106). RTX on the other hand cramps a lot more silicon due to the die shrink (12nm TSMC).
Also, compared to the Maxwell 2 and earlier generations, the gap between the Desktop and Laptop Editions of the Pascal and Turing GPUs remain only on the clock frequency sections. Other portions of the GPU such as VRAM size and speed, bus bitrate and GPU Core counts all remain identical between the two.
I am wondering if the next generation (Ampere) made on 7nm Samsung would be better than the 14nm types in which both Desktop and Laptop variants would carry exact same specs due to a lower thermal footprint. Learning from Intel's Ice Lake CPUs of allegedly getting computing performance expected of a 65w part using a 15w chip.

Last edited by Tsurugi_Matsuda; 18-06-2019 at 04:24 PM..
Tsurugi_Matsuda is offline   Reply With Quote