I still express doubts, given that X86-based CPU's may still run quite hot because of the complexity of each x86 core.
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.
AMD Claims New Laptop Chip Is 30% Faster Than M1 Pro, Promises Up to 30 Hours of Battery Life
- Thread starter MacRumors
- Start date
- Sort by reaction score
Because AMD wouldn't win at present.
The main reason though is that they now have "AI-ML" built into their silicon now. Where as intel doesn't.
To that extent, they are even further behind Apple. Since we have had this for a few years now. Wasn't it like A12 or A13? One of those. It's like MMX and 3dNOW back in the day. They finally have something, and at least from what they show. It looks like it will compete well with 1 and 1/2 year old+ AS.
They didn't dive into how much Power Per Watt, and what clock rates at what wattage. Or how much power use from the battery to run full tilt. If the system will need a fan. Or what you will be able to actually do for 30+ hours.
AMD usually will show where they are lagging behind intel in performance. So they generally are not adverse to showing exactly where they are, even if they are behind. But, this was not one of their better presentations. They shouldn't have even mentioned AS unless they show the whole story. Just compete against intel as they should.
If they needed to include Apple as a reference. They should have showed M2 ML performance vs their CPU.
Not to mention, AS beat intel in the Blender render. SO..... yeah. That says a LOT right there.
The "it drives competition" crowd seems to believe "We have McDonalds at home" when it comes to accepting claims like 30 hour battery life and speed improvements (for specific tasks) that are unexpectedly introduced into a product line that historically only saw gradual year-over-year improvements and was notable for its less-than-optimal power efficiency.
People have been talking about the "imminent death" of x86 since the mid-1990s. Any day now, I guess.
Its the current high end laptop soc from Apple. True it's not a flagship like M1 Max but no way in hell it's low-end midrange, that title goes to M2.
I don't know if they are resting... I am going to guess there is a lot going on in the chip market that we don't know about but only guess. What have been hearing about for the past year and half re: manufacturing/shipping/production slow downs and chip makers etc not being able to meet demands or production numbers.
Also, what they say and what they do are two different things. Apple has already shipped a product. Let's see how AMD handles a shipped product.
Even if it is there, all practical uses are lost and no developer offer support, at least for apps that are actively developing. Before it was mandatory and I could find some nifty use out of it. Not anymore.
There are small chances Apple invests comparable amount of money to AMD in chip design. AMD's entire business is focused around chip design at the end of the day and they have way more IP and designs than Apple and are present in more markets. Also, your estimation regarding the number of AMD employees is quite outdated. AMD alone has more than 16.000 employees and an additional 5.000 from Xilinx so more than 20.000, it's impossible for Apple to have close to the number of people AMD has that work in chip design.
Also Apple doesn't sell chips they sell finished products. AMD's long term potential in chip volume is higher than Apple's and their growth is mostly limited by the production capacity TSMC is able to allocate for them. AMD right now is TSMC's biggest 7nm costumer and soon they will be TSMC's biggest 4/5nm costumer.
This is getting annoying. Because the 7940HS is a mid-range SKU and most likely will be used in laptops that are priced similarly or below the entry M1 Pro Macbook. When AMD directly compares their chips to a certain product from another company it's because that's their direct response for that chip, this is why AMD didn't compare at launch the 7900XTX with the already available RTX 4090 because the 7900XTX is meant to compete with the 4080 which wasn't available when AMD launched their GPU.
And don't worry AMD has APUs that can be considered competition for the M1 Max or M2 Max, the 12 and 16 core Ryzen 7045 series.
Last edited:
That's nonsense. X86-based CPU don't universally run hotter even if they are clocked much higher. Go and watch some CPU stress tests for the M1 and M2 chips. They easily reach 100C and sometimes go even above that.
Backwards compatibility will always be a structural limitation of x86 based CPU’s. For every step forward they take, they have to make sure old 16-bit commands can still be decoded alongside 32 and 64 bit. That complexity can only be reduced by breaking backwards compatibility, and neither AMD nor Intel wants to be the first to say, “My chip isn’t compatible with what you do today.”
Exactly. This is an announced chip from AMD that won't be shipping (according to AMD) in OEM laptops until late March. Apple could very easily talk about their M2 Pro / Max / Ultra chips now, or they could just as easily say nothing as they usually do and drop new Macs with them in March too.
I get it. CES is the one time a year that every company not worth a trillion dollars gets to have a go at putting on a flashy keynote and they have to announce something. Ready or not.
Good to see Intel and AMD feeling the heat from Apple Silicon.
Oh, that is better data. Thank you.
TSMC n5 volume production starts -> 4/20; n5 shipping with M1 ASi -> 11/20, a 7 month difference
TSMC n3(b?) volume production starts -> 9/22; n3 shipping with M? ASi -> 4/23?, a 7 month difference?
TSMC n3(b?) volume production starts -> 9/22; n3 shipping with M? ASi -> 4/23?, a 7 month difference?
Name doesn't check out. Please explain how AMD's Bulldozer almost bankrupted the company using the same process as previous successful designs. Would like your input on Itanium as well. TSMC's 7nm process has been used in many designs across many companies. You have not noticed any discrepancy in performance and efficiency between them ? There's even a large discrepancy between Apple using the same Arm ISA, on the same process, compared to other companies.
AMD Bulldozer - See what fab they were on. TSMC & Global Foundries was grossly behind Intel at that time.
Itanium's primary issue was that what they grossly overestimated how quickly they could make process jumps, so they designed something with the intention of sub-180nm when they were on 200+nm. Thus, they made significant cuts to the complexity of the chip.
It's like if someone at dominos promises you they can give you a large, 4-topping pizza for $25 but then at the last minute says "Wait, now it's a medium pizza with 3 toppings".
Ultimately, it's not up to the designers when it comes to chip performance as long as the designer is willing to pay to stress the process as much as possible. Anyone can make the 4-topping choice given that they have the $25 and dominos is able to do it. It's child's play.
As a designer, your only problem is convincing your mom to give you that $25. It's pretty simple in the grand scheme of things.
So you're blaming Bulldozer's failure on the process ? How was Opteron so successful ? Itanium's failure was about process and nothing to do with a new ISA ? I notice you didn't address my comment about comparisons between Apple's 7nm chips against other on the same TSMC process. There's substantial differences in GPU's on identical processes as well. We can even go deeper and look at the substantial differences in IP PPA on identical process nodes such as memories, standard cells, SERDES, PLL from various vendors. Any thoughts on this ?
Last edited:
Let’s see how the chip fares when it gets released later this year. My guess is that the claims, like many from PC manufacturers, are grossly inflated.
Yeah, pretty obvious they created a brand new account just to post the same thing over and over in order to get folks riled and to respond. You won’t be missing anything in their posts if you “Ignore”
They’ll just have to start posting from their main account again LOLApples been about big leads. I could be very wrong, but I think/hope as one of TSMC top clients they’re betting/investing heavily on 3nm, while probably securing a massive chunk of the chips as an early investor. 4nm was an improvement on 5nmbut very, very slightly; it was pretty much a disappointment. I don’t blame Apple for waiting for a better chip.
Apple is falling behind. Apple needs 3nm fast. Also needs to double the standard ram and storage against PC competitors. 16GB and 1TB is now dirt cheap after the shortage ended.
The M1 Max has the same CPU cluster as the M1 Pro. That's why there's no point in AMD doing an M1 Max comparison.
Secondly, the Ryzen will be in sub $1500 16" laptops. Comparable M1 Max 16" starts around $4000 USD.
The M1 Max has the same CPU cluster as the M1 Pro. That's why there's no point in AMD doing an M1 Max comparison.
Secondly, the Ryzen will be in sub $1500 16" laptops. Comparable M1 Max 16" starts around $4000 USD.