Hacker News Reader

Comment by weare138 | Hacker News Reader

Intel didn't beat AMD back in the day because they had a better processor. Intel relied on anti-competitive practices to lock AMD out of the desktop and server markets. They were successfully sued by AMD but the damage was done. Then Intel sat on it's laurels and quit innovating, slowly doling out incremental improvements through their 'tick-tock' release cycle. Intel dug their own hole and I have no sympathy for Intel while AMD buries them in it.

View on HN

You appear to be conflating Intel's 1990s antitrust woes and the success of the Core microarchitecture which launched in 2006.

Despite Intel's well-documented anti-competitive practices, AMD had decisively gained the upper hand in performance by the time of K8 (2003). Intel's microarchitecture at the time, NetBurst, was uncompetitive and relied almost entirely on inertia, and yes, anti-competitive behaviour to retain market share.

Core leap-frogged that and Intel remained ahead for the better part of a decade afterwards. To say that Intel sat on its laurels between the time of the first round of antitrust litigation in the early 90s is quite inaccurate.

Do you remember the bus license thing where they killed off nvidia?

I don't agree. Anti-competitive practices from Intel were orthogonal to the success of Conroe.

It was a better processor: https://www.anandtech.com/show/1963

I agree that Intel did engage in a lot of horrible anti-competitive things, but that has nothing to do with how incredible Conroe was. I find it kind of funny how this whole AMD vs Intel competition gets people's emotions running as if its their team and turf.

AMD created x86-64, Intel had just caught up by that point.

intel did a bit more than just "catch up"; take a look at the benchmarks in the linked article. they followed it up with solid improvements through sandy/ivy bridge. then they treaded water for eight years or so. even still, after the launch of conroe, amd didn't have a competitive high-end part until zen 2 (excepting a few high core count oddities along the way).

the userspace fixes required to make those intel implementation secure did a bit more than question those "improvements".

Reply on news.ycombinator.com