On 24/09/2019 05.02, David T-G wrote:
Carlos, et al --
...and then Carlos E. R. said... % % Once I timed writing very large file (ISO image) from hard disk to an ... % slowed up. The hypothesis we had is that the system cache memory was % used up. I did a few tries, and this is what worked best (faster):
That's interesting. On my laptop I tend to run into the same SERIOUS slowdown whenever pulling data, either via the built-in SD card slot or the USB readers.
% % dd if=ISO of=DEV bs=16M oflag=direct status=progress
Oho! That is *very* interesting.
% % This allows to write several sticks at the same time without big impact. % % Not your case, but the idea is that you can try things :-)
Indeed :-) Now I wonder if it might actually be more productive to image the cards to files on the SSD and then mount *those* to xfer at leisure. I'm definitely going to have to benchmark that!
I believe that writing big enough files to SSD may hit a limitation on the SSD of continuous writing. In this case, I think that reading the files out of the card will be faster than imaging it first.
% % You are reading from card to hard disk. If it is USB3 to rotating rust, % the limitation will be the write speed of the HD and you will hit it % easily, perhaps doing 150 MB/S. If the destination is m.2, who knows.
Yep. If I do go with spinning secondary, it will be straight striping (no fault protection) across multiple devices (the more heads the better for all practical purposes).
Thanks again & HANN
:-D
Welcome :-) -- Cheers / Saludos, Carlos E. R. (from 15.0 x86_64 at Telcontar)