I think it’s interesting to see the guy in the article saying he’ll never buy another launch CPU from Intel again. I would expect that once Intel puts this behind them, they’ll be upping the QC process to catch these type of things. And hopefully AMD learns enough they can test for it too so it never happens to them.
The hilarious thing is AMD probably won’t need to learn their lesson: their chips are using about half the watts for the same performance. This is firmly an Intel failure and them jucing the silicon past it’s ability to cope to win benchmarks.
And their Zen5 cores look to be even more efficient than Zen4, so I’m sure some engineers at AMD are laughing about this whole thing.
One thing I find very amusing about this is that AMD used to have a reputation for pulling too much power and running hot for years (before zen and bulldozer, when they had otherwise competetive CPUs). And now intel has been struggling with this for years - while AMD increases performance and power efficiency with each generation.
Yeah, the Prescott P4s and the dual-core P4s were both extremely good at being space heaters while also being extremely bad at being a CPU.
The whole P4 era was a suckfest for Intel, and they didn’t really shake the slow, hot, power hungry thing until Sandy Bridge (so basically 2000 to 2011 ish). And before anyone mentions them, Core/Core 2 were MUCH more performant and certainly better performance-per-watt than the X2/X4 Athlons, but still not particularly efficient in comparison to Sandy Bridge and later.
I’d also argue Bulldozer wasn’t necessarily a bad uArch, but it just had a couple of poor design decisions that made sense historically but did not really work in practice, and had fucking awful timing going up against Sandy Bridge, which was fucking excellent.
I think it’s interesting to see the guy in the article saying he’ll never buy another launch CPU from Intel again. I would expect that once Intel puts this behind them, they’ll be upping the QC process to catch these type of things. And hopefully AMD learns enough they can test for it too so it never happens to them.
The hilarious thing is AMD probably won’t need to learn their lesson: their chips are using about half the watts for the same performance. This is firmly an Intel failure and them jucing the silicon past it’s ability to cope to win benchmarks.
And their Zen5 cores look to be even more efficient than Zen4, so I’m sure some engineers at AMD are laughing about this whole thing.
One thing I find very amusing about this is that AMD used to have a reputation for pulling too much power and running hot for years (before zen and bulldozer, when they had otherwise competetive CPUs). And now intel has been struggling with this for years - while AMD increases performance and power efficiency with each generation.
Intel had that rep before AMD back in the P4c days
Yeah, the Prescott P4s and the dual-core P4s were both extremely good at being space heaters while also being extremely bad at being a CPU.
The whole P4 era was a suckfest for Intel, and they didn’t really shake the slow, hot, power hungry thing until Sandy Bridge (so basically 2000 to 2011 ish). And before anyone mentions them, Core/Core 2 were MUCH more performant and certainly better performance-per-watt than the X2/X4 Athlons, but still not particularly efficient in comparison to Sandy Bridge and later.
I’d also argue Bulldozer wasn’t necessarily a bad uArch, but it just had a couple of poor design decisions that made sense historically but did not really work in practice, and had fucking awful timing going up against Sandy Bridge, which was fucking excellent.