As TFA says, they are running the algorithm multiple times and they check that the results match, to guard against transient errors caused by radiation.
The permanent errors caused by radiation must be identified by periodic self tests. When the permanent damage is in a redundant structure, e.g. as mentioned in TFA when they find some memory bits that are permanently damaged, they must avoid using what is damaged.
Eventually radiation will destroy something that is essential, but until then the Snapdragon CPU should be usable.
wongarsu•23m ago
Yeah, that's kind of awesome, isn't it?
Flying a helicopter on Mars was inspiring and useful for scouting, etc. But maybe the best thing coming out of it is undeniable proof that off-the-shelf hardware without radiation hardening is perfectly viable on Mars if you can just reboot it fast enough
__patchbit__•47m ago
Tesla AI chips already fielded in vehicle fleets are a better option were it not for the institutional SLS rocket styled bureaucracy's bias.
gavinsyancey•25m ago
How do you propose to install them in the rover, which is already on Mars? The helicopter base station CPU is already in the rover, since it was included to communicate with the helicopter. And it's no longer needed for that purpose, since the helicopter crashed and broke a propeller.
schlauerfox•17m ago
Not that you are casting shade, but to clarify in defense of a former professor of mine who worked on the FPGA avionics at Nasa Jet Propulsion Laboratory, the helicopter was designed for 5 test flights, and completed 72 before the rotor damage. https://science.nasa.gov/mission/mars-2020-perseverance/inge...
mapt•15m ago
This system sounds like one of the many pieces of science equipment whose costs are >98% in one-off R&D engineering & mission ops, and <2% in marginal cost of construction/launch.
Imagine a hundred of these exploring Mars semi-autonomously, maybe with LoRa mesh networking, for not a whole lot more money than it cost to send one.
1970-01-01•1m ago
So 100% of the Snapdragons on Mars are no longer sitting idle and are tasked doing useful work. Why can't we do the same for Earth?
tibbydudeza•1h ago
adrian_b•58m ago
The permanent errors caused by radiation must be identified by periodic self tests. When the permanent damage is in a redundant structure, e.g. as mentioned in TFA when they find some memory bits that are permanently damaged, they must avoid using what is damaged.
Eventually radiation will destroy something that is essential, but until then the Snapdragon CPU should be usable.
wongarsu•23m ago
Flying a helicopter on Mars was inspiring and useful for scouting, etc. But maybe the best thing coming out of it is undeniable proof that off-the-shelf hardware without radiation hardening is perfectly viable on Mars if you can just reboot it fast enough