I (still) have 4 cores on one die that are sitting in disabled mode - wish i could get them working.
not all cores are worth bringing back.. if they flap on and off all the time your HW errors fly up. .96 seems to balance a lower WU but much lower HW errors for better avg.
would be nice to have release notes right??
I am guessing that it may 'give up' on some cores which is good for problem ones. but would be nice to control that ourselves. It would make sense that they have the enable_cores feature now to go along with this design instead of earlier firmware that always tried to bring everything back
Should be able to manually control which cores are disabled in .96 by reverse engineering the mask from the asic_test file. It looks like it works similarly to a netmask from ip networking. Convert the mask to binary, and line up the bits with an array of 192 bits representing each core...do an XOR between the two..values that end up as a 1 are active and 0 are disabled. Or so goes the theory, anyway.