Hindsight · review build · data not yet validated
There is increasing interest in massively parallel neural nets, genetic algorithms and other forms of “chaotic” or complexity theory computing, although most computer computations are still done using conventional sequential processing, albeit with some limited parallel processing.
Grader A: hit
Interest in neural nets was rising in 2009: deep belief networks (2006) and GPU-trained deep networks (2009, up to 70x faster) came before the 2012 AlexNet breakthrough. Most everyday computation stayed sequential; dual-core CPUs became common in PCs in the late 2000s and most applications were not rewritten for parallelism.
Test: Two-part descriptive test: growing research interest (met, 2006-2009 papers) and most computation still sequential with limited parallelism (met, dual-core PCs, mostly serial software).
Grader B: hit
In 2009 almost all computation was conventional sequential code with multicore limited parallelism, and interest in neural networks was rising (deep learning results in speech from 2009).
Test: Industry test: sequential/limited-parallel computing ~100% of workloads; neural-net research interest growing.