We need to see more products to understand what Apples plans are. It’s still not clear whether the M1 lineup is representative for their long term strategy or whether it was a one time test of horizontal scaling technology. I still think that if Apple is serious about desktop their next step will be vertical scaling and possibly new generation of high-perf cores. It’s also notable that they have been steadily improving their E-core technology every year.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
CPU Wars 2022 - Apple Silicon vs. Intel/AMD
- Thread starter maflynn
- Start date
- Sort by reaction score
We need to see more products to understand what Apples plans are. It’s still not clear whether the M1 lineup is representative for their long term strategy or whether it was a one time test of horizontal scaling technology. I still think that if Apple is serious about desktop their next step will be vertical scaling and possibly new generation of high-perf cores. It’s also notable that they have been steadily improving their E-core technology every year.
To quote statistician Richard McElreath, “all [statistical] models are wrong, some [statistical] models are useful”.
Obviously, it’s impossible to create a benchmark that accurately predicts performance on every existing workflow into a simple comparable number: different tasks have different demands (cache use, vectorization, heavy RAM use, etc.), so a general benchmark (like Geekbench or SPEC or even Cinebench) is going to have limited predictive power for how fast a specific task will be between systems.
However, limited predictive power can still be very useful! Even if Geekbench performance only correlates 70% with performance on my specific EEG preprocessing pipeline, that’s still very valuable information given that Ars Technica or even niche YouTube reviewers probably aren’t going to test any of the scientific/statistical software where my computer does most of its heavy lifting.
As long as they can tell us *something* about general performance, general benchmarks are going to be useful, even if they’re often “wrong”.
It's unclear that this represents "strategy". Current Intel seems to be in a constant scramble to try to ship (some version of) what Marketing promised two years ago.
Meaning that what will ship for any given new product (laptop?, desktop?, HEDT?) seems primarily determined by what they're actually able to put together in non-trivial numbers.
- If the yields and performance are low, sell it as some specialty part to some country that sells very few and claim victory.
- If the yields are low and energy is bad, but performance is high, start with an HEDT part priced so high you don't sell many.
- If the gating item is your CPU tiles (rather than TSMC's GPU or IO tiles) sell the version that includes lots of GPU and IO per CPU (probably mid-range desktop). etc etc.
This is not "strategy", it's simply doing whatever you can to keep up with promises you made two years ago. (Meaning, uh, perhaps if you **** and made fewer promises way in advance, you could ACTUALLY implement strategy...)
Well if we take a look at their updates to Meteor Lake in the Linux Kernel 6.1, we can definitely see that they are giving a hell of a focus On the mobility devices.
Same can be said about AMD on their P core and cache rescheduling being done.
This is the great unknown, seeing what M2 pro/max/ultra will look like and/or seeing true pro level desktop computers. The studio is really the only desktop on arm that is "pro"
People are aching to see what an ARM Mac Pro will look like
I kinda want to give them a couple more years before I really fully make up my mind. Pat Gelsinger has only been back for about a year and a half. New CPU designs take a few years to go from design to fab/market, so the stuff that they are releasing today is really stuff that they began engineering in ~2019 or so. They are more or less still releasing stuff that was probably already in the pipeline before Gelsinger took over.
My hope is that this is all just a strategy to buy time, because I definitely hope to see them develop higher IPC cores. They are achieving jaw-dropping performance with insanely high clock speeds on Raptor Lake (and those performance numbers are truly impressive), but with that kind of power consumption, there is no way these top-of-the-line parts will ever be practical for laptops and mobile users. You can't put a 250W TDP CPU into a thinkpad (sorry, marketing department. I don't care what tdpDown is on paper. 😂)
Last edited:
They just released the higher IPC core in 2021. I would't expect a new P-core from Intel until at least 2025-2026. Will it be able to deliver better power efficiency? Intel would probably need to start completely from scratch to achieve that.
Honestly I'd rather have the M* Pro chips be focused on delivering more efficiency and having full performance focus on the Max chips.
We could get some really good battery improvements on the 14" Pro and 16" Pro that way.
We could get some really good battery improvements on the 14" Pro and 16" Pro that way.
Intel laying off 11% of their workforce. Odd that this is after getting a ton of money from Congress. I wouldn't be surprised if AMD has something in the works too.
www.marketwatch.com
Intel reportedly plans to lay off thousands of workers, with details potentially emerging alongside quarterly earnings
Intel reports quarterly results on Oct. 27. Its last big layoff round, comprising 12,000 job cuts, was announced in tandem with first-quarter earnings in 2016.
www.marketwatch.com
Intel laying off 11% of their workforce. Odd that this is after getting a ton of money from Congress. I wouldn't be surprised if AMD has something in the works too.
Intel reportedly plans to lay off thousands of workers, with details potentially emerging alongside quarterly earnings
Intel reports quarterly results on Oct. 27. Its last big layoff round, comprising 12,000 job cuts, was announced in tandem with first-quarter earnings in 2016.
Fire marketing sales folks and hire more fab technicians . There is no real conflict there with the CHIPs act. Suppose to be trying to create more manufacturing jobs. Not composing more spin. If creating humongous bucketloads of spin was real manufacturing the USA would be king of the world. ( i.e., that non stop political ads on TV right now and for next 4 weeks. )
Just me speaking without knowing any details, but i suspect the layoffs are due to softening market and more global recession type fears then anything specifically wrong at Intel. Other sectors are already laying people off, so while this is disheartening (I feel bad for those who are impacted), its not the last we hear about layoffs