So...
TL;DR:
Intel and nVidia sent out marketing droids to lie and they got some facts wrong.
Unfortunately no marketing droids died in the process.
Nvidia just fired the first salvo in what promises to be a classic and long-lived benchmark death match vs Intel. In a webpage titled "Correcting Intel's Deep Learning Benchmark Mistakes," Nvidia claimed that Intel was using outdated GPU benchmark results and non-current hardware comparisons to show off its new Knights Landing …
In this context it's neural network operations.
It's all the craze because, you know, Google and ads and Big Data and consumer targeting.
GOFAI stuff like "number of SLD resolution calls per second" or even newfangled quasi-logic stuff it ain't.
That's my view and I'm sticking to it.
Although running a SAT solver on a Xeon Phi sounds feasible ... hmmmm ...
Many - or let's say all HPC & Super Computing insiders & programmers already know the truth!
Big INTEL fears that the future earnings guidance (important for shareholders) will be crushed by IBM (OpenPOWER Foundation) POWER 8 & 9 CPUs for HPC Server with Nvidia powerfull HPC NVLink Interconnect GPU products (Pascal in 2016 / Volta in 2017) and so - the 'New' Intel Knights Landing Xeon Phi - is in fact already OUTDATED!
So Intel try to 'fake' accurate benchmarks with stupid PR tricks - boomerang hit anyone = facepalm Intel marketing - 5 sec. applause - then Round 1 TOK !