Quote from: Neenyah on June 22, 2023, 11:20:21Hm, they don't consume twice the power (vs AMD) 🤨

AMD is better, sure, but not in idle where they are absolute disaster and they can't come anywhere near Intel's idle efficiency. In full load they are not twice better/more efficient than Intel.

WaitWhat should have said 'up to twice' the power to be more accurate. But sentiment seems to be the same.

As for idle power, yes. Currently, the multi-chiplet 7X45 series consumes more at idle than Intel's and AMD's monolithic SOC/APU's. It seems AMD didn't do any optimizations on the I/O die for mobile compared to desktop. But these chips are in a niche market (DTR) and are 'halo' products that (probably) sell 1/50th to 1/100th of mainstream H series. Hence AMD chose faster time to market & lower cost route versus a new I/O die. But in the end, is 4 hours of Wi-Fi run time that much different than 5-6 of an Intel HX series? There will be times that it will but on average, a user is till going to have to plug in and recharge during a business day.

It makes no sense to compare laptop chips at consumption levels above 50W - this is nonsense.

Quote from: WaitWhat on June 22, 2023, 07:17:56Intel offers higher performance than both but consumes twice the power so I wouldn't count that as a win/fair comparison.

Hm, they don't consume twice the power (vs AMD) 🤨
 
Jarrod did quite an extensive video about it 👉 Best Laptop CPU 2023? Ryzen 9 7945HX vs Intel i9-13980HX ( youtu.be/mdWAfPfYTnU )

AMD is better, sure, but not in idle where they are absolute disaster and they can't come anywhere near Intel's idle efficiency. In full load they are not twice better/more efficient than Intel.

Don't know what the others are talking about but M2 Max used around 70W in CB R23 so M2 Ultr should at least double that to 150W making it quite comparable to AMD's chips and it doesn't look good: AMS' 12 core mobile chip (7845HX) scores roughly the same and their desktop chip (7950X) scores 33% more while not losing in efficiency. Intel offers higher performance than both but consumes twice the power so I wouldn't count that as a win/fair comparison.
The Apple M series only seemed great because Intel stagnated on 14nm & 4 cores for so long that Apple had plenty of time to catch up but now that speed of x86 progress has normalized again Apple is falling behind quickly. This will only get worse because AMD has larger gen-over-gen gains than Apple and is already in the lead. The only advantage that Apple has is that they are TSMCs biggest customer and therefore get the best nodes earlier. Their last architectural change with IPC improvement came with A13 since then it's just clock speed increases because of better TSMC nodes.

Quote from: ArsLoginName on June 22, 2023, 01:49:24It also shows that Apple's 24 core Ultra (60 W, 5 nm) is only 3.5% faster than AMD's last generation 16 core 5950x (105 W, 16 core). When adjusted to TSMC's 'N5 technology provides about 20% faster speed than N7 technology or about 40% power reduction (directly from TSMC's website), what is so shocking? More cores on a newer node being about as fast as 33% less 2.5 year old cores on an older node?

Why stop at comparing manufacturing technology. Normalise it for wattage and clock speed for a better comparison.

It also shows that Apple's 24 core Ultra (60 W, 5 nm) is only 3.5% faster than AMD's last generation 16 core 5950x (105 W, 16 core). When adjusted to TSMC's 'N5 technology provides about 20% faster speed than N7 technology or about 40% power reduction (directly from TSMC's website), what is so shocking? More cores on a newer node being about as fast as 33% less 2.5 year old cores on an older node?