2024 BrainChip Discussion, page-6761

I was personally surprised to read that Loihi 1 had these issues.

Possibly explains why it was so hard for researchers to actually convince Intel to let them have an actual chip to experiment with while Brainchip was scattering AKD1000 like confetti:

From the NASA Brainstack research program:

“A summary of past tests, run times, and application execution success results below are shown below in Table 2. The table shows a breakdown of the applications executed and which applications executed successfully. One of the key observations from this testing is the Loihi-1 processor does not perform well when limited to a shorter run time. The reason for this result was determined to be that due to the experimental nature of the Kapoho Bay and

Loihi-1 silicon design, the processor does not operate properly at colder temperatures and needs several minutes to self-heat before it begins to execute successfully. TES-13’s nominal interior temperature hovers around 0°C in its 500km, 45° circular orbit, and the Kapoho Bay was determined to not properly operate until reaching 15°C.

Additionally, experiment runs exceeding 220 minutes have resulted in overheating conditions in the processor, resulting in automatic self-shutdown.

However, these behaviors were a known operational artifact of the Loihi- 1, revision ‘A’ fabrication”

My opinion only DYOR

Fact Finder