Which of the following best describes the reason why no one is deploying a Dynamo-like system (the 1999 HP paper, not the 2007 Amazon paper) to accelerate native code / binaries at scale?
hpl.hp.com/techreports/19
Conversation
Replying to
I don't know. But compiler people seem to have a history of making remarkable claims that don't quite pass the smell test, which might be relevant here. For example, Proebsting's law:
Obviously can't be true, as there is an upper limit to how fast a given program can be made to run on a given CPU, when approaching that limit advances must slow down.
1
1
My immediate objection is that it tacitly assumes that compilers exist in a vacuum. Why give credit to (say) automatic vectorization optimizations instead of the underlying newly available SIMD instructions? I doubt that there is a clear answer.