Hi and welcome to the forums!
While we try to keep Etcher as fast as possible, flashing/verifying speed also depends on the type of connection you’re using and the target actual read/write speed. With USB 2 you should never go past 25MB/s for example and also flashing more than one target can reduce the flashing speed based on how the USB connection was made.
The numbers you wrote here look reasonable to me, I get ~18MB/s on OSX with a USB hub and only flashing one target, while it jumps up to ~24MB/s on verification.
If I flash two targets at the same time, I get ~16MB/s flashing and always 24MB/s verifying as I’m using a USB 2 and a USB 3.1 targets, but the hub is connected via USB C so I’m not reaching the main bottleneck there but rather on the USB 2 target.
I hope this makes sense and clears your doubts!
The test was done on the same equipment back to back by just changing the OS on the hardware. The port is USB 3.0.
It would seem that different OS is getting difference flash / verify speeds. The Linux performance was quite disappointing for flash, but it was better for verify.
I’m interested to get closer to higher speeds, especially for the verify process.
Any idea how to improve it on my existing set-up? I can do the testing.
Hi,
On Linux, you can try comparing the results you get on Etcher with those you get using dd to start checking for bottlenecks. Hardware plays an important role, so the USB 3 port needs a USB 3 target to work at full-speed as mentioned.
You can also try modifying the source code (available here ) to use a different block-size.
Other than that, we can’t really suggest much unfortunately, a lot comes from the hardware used so maybe try with different combos of sdcards + readers, usb sticks + usb hubs, etc. that vary a lot.