I see Arrow Lake as the stop-gap CPU that Intel launched while they are working on Panther Lake. Maybe they should have just launched 8 P core desktop variants of Meteor Lake instead, as ML was initially designed for mobile and desktop use also. The whole thing with the tile architecture was (is) supposed to make that a lot more straight…
I see Arrow Lake as the stop-gap CPU that Intel launched while they are working on Panther Lake. Maybe they should have just launched 8 P core desktop variants of Meteor Lake instead, as ML was initially designed for mobile and desktop use also. The whole thing with the tile architecture was (is) supposed to make that a lot more straightforward. Several tests of mini-PCs with the 185H Laptop SoC (6P cores) such as the ASUS NUC model with it showed that this ML model for mobile was able to go pretty fast, if it has better cooling and is allowed to go to a PL1 of 85 or 100 W.
Meteor Lake's Redwood Cove/Crestmont cores were pretty minor upgrades over Raptor Lake though. Arrow Lake is much better in that both core designs got huge updates. Unfortunately my MTL laptop can't go to 85 or 100W, so I can't speak to whether it could have gotten close to ARL.
I agree with that, in principle. IMHO especially the improvements from Gracemont to Skymont are quite substantial. However, the proof of a pudding is in the eating, or here it is in the performance. And Arrow Lake is a disappointment in that regard. It is a more efficient CPU, but otherwise about level with Raptor Lake. And, as all of Arrow Lake's tiles are fabbed at TSMC (and there at full cost to boot), Intel won't make much if any money from that CPU generation. It might have been economically more beneficial for Intel to launch Meteor Lake also into as desktop variants. But, all that is now only theoretical. In the meantime, Raptor Lake continues to be available both in desktop and mobile versions. I wonder what the sales of those are compared with Arrow Lake.
Would you care to explain why Intel won’t make money by having its chips manufactured by TSMC? How is AMD then able to remain profitable? Last year Intel reported that manufacturing Meteor Lake on Intel4 led to huge losses. Moving to TSMC was probably the best choice they had.
I see Arrow Lake as the stop-gap CPU that Intel launched while they are working on Panther Lake. Maybe they should have just launched 8 P core desktop variants of Meteor Lake instead, as ML was initially designed for mobile and desktop use also. The whole thing with the tile architecture was (is) supposed to make that a lot more straightforward. Several tests of mini-PCs with the 185H Laptop SoC (6P cores) such as the ASUS NUC model with it showed that this ML model for mobile was able to go pretty fast, if it has better cooling and is allowed to go to a PL1 of 85 or 100 W.
Meteor Lake's Redwood Cove/Crestmont cores were pretty minor upgrades over Raptor Lake though. Arrow Lake is much better in that both core designs got huge updates. Unfortunately my MTL laptop can't go to 85 or 100W, so I can't speak to whether it could have gotten close to ARL.
I agree with that, in principle. IMHO especially the improvements from Gracemont to Skymont are quite substantial. However, the proof of a pudding is in the eating, or here it is in the performance. And Arrow Lake is a disappointment in that regard. It is a more efficient CPU, but otherwise about level with Raptor Lake. And, as all of Arrow Lake's tiles are fabbed at TSMC (and there at full cost to boot), Intel won't make much if any money from that CPU generation. It might have been economically more beneficial for Intel to launch Meteor Lake also into as desktop variants. But, all that is now only theoretical. In the meantime, Raptor Lake continues to be available both in desktop and mobile versions. I wonder what the sales of those are compared with Arrow Lake.
Would you care to explain why Intel won’t make money by having its chips manufactured by TSMC? How is AMD then able to remain profitable? Last year Intel reported that manufacturing Meteor Lake on Intel4 led to huge losses. Moving to TSMC was probably the best choice they had.