... anyone who's built their own computer or even reads websites like tom's hardware or benchmarking sites knows this.
Tom's hardware was heavily suspected of falsifying benchmarks in exchange for money from Intel and Microsoft years ago (Google it you lazy pricks!). Tom's and other sites often provide BS benchmarks. Meaning, they spec a new Intel against an old AMD and claim Intel is way better. In other cases, they load up a system with 256GB of memory and give the system they want to talk down 4GB. While these cases are technically not false, the merit and the motive behind such benchmarks are worth chastising people that publish this type of information. Often finding the spec's of the real test gear is buried and extremely difficult to find.
Sorry, but I trust companies proving benchmarks as far as I trust my turds will fly. If you work in IT, setting up your own bench marks for these types of applications should be trivial. Most vendors will give you gear to demo. I don't trust companies that rely on vendors for gear and revenue to provide real data, there has been way too much shit in the pool. If I don't personally test or know the person that did, it did not happen.
A desktop system with 256GB of RAM? this was years ago you say?
While I am aware certain memory configurations may have been used to influence benchmarks, these typically do not have more than a 10% performance impact.
Intel played a dirty game, AMD sued, got paid, case settled and closed.
If you wish to shed more information and enlighten us regarding discrepancies in testing I'm sure you will get many readers. I'd suggest collaborating with a competitor of Tom's Hardware.
Oh, the memory amounts you mention make no sense, FYI.