I know better than to click on an article like this, but I fell for it. Now, I guess I swallowed the bait, because I'm also posting.
4. Core i5-2500K: That time Intel won so hard that AMD gave up for half a decade
Any article about this period that doesn't
also talk about
Intel's manufacturing advantage is bordering on journalistic malpractice. It's easy to forget, but Intel had literally a couple
years lead on the manufacturing tech of anyone else in the chip game, and Sandybridge probably came right at the peak of that. Intel's lead was so bad that I was strongly of the opinion that Intel should've been forced to spin off its fabs, which looks like it could finally happen for other reasons.
No doubt AMD had other things going wrong with it, but even a comparable design from them wouldn't have been competitive due to Intel's superior node. I think AMD knew this, and it's one of the factors that motivated some bad decisions on their part, like the FPU-sharing of Bulldozer and hoping they might be able to compete with GPU-compute (i.e. the whole HSA debacle - look it up).
The
only way AMD CPUs managed to regain relevance is
both that Intel's 7 nm fab node hit years worth of delay, allowing TSMC to catch up & even pass them,
and AMD
finally waking up and adopting a properly modern architecture (many thanks to Mike Clark and Jim Keller, for that). Here's what Jim said about the latter point:
"Mike Clark was the architect of Zen, and I made this list of things we wanted to do [for Zen]. I said to Mike that if we did this it would be great, so why don’t we do it? He said that AMD could do it. My response was to ask why aren't we doing it - he said that everybody else says it would be impossible. I took care of that part. It wasn’t just me, there were lots of people involved in Zen, but it was also about getting people out of the way that were blocking it."
Source: https://www.anandtech.com/show/1670...storrent-ceo-ljubisa-bajic-and-cto-jim-keller
Even though it was executed successfully, Zen could not match Skylake's IPC. The first gen was only relevant due to having double the core count. It took further generations of refinement and
a properly competitive TSMC node for AMD to briefly pass Intel on single-thread performance.