4 ms·
Exactly how is it much better? It's not the first chip with a 15-25W TDP. When compared to other chips with such a TDP, it's incrementally faster in single core
by sudosysgen 6y ago
Exactly how is it much better? It's not the first chip with a 15-25W TDP. When compared to other chips with such a TDP, it's incrementally faster in single core and it trades blows in multicore. It does this at a process node ahead.
For very light loads, it does have a slightly better power performance than existing low power PC processors, but not by that much. A Ryzen LPP core is somewhere between the big and the small cores of the M1 in power performance in such cases.
It's not the first PC chip to be an SOC, or to have unified memory, or to have many coprocessors. Literally all of those were done simultaneously, in chips of comparable performance, at the same power.
It's also been a decade that we know that instruction sets don't matter. All modern PC processors use a different instruction set and translate from x86. To switch them to ARM would be a matter of adding a few extensions to expose all the functionality, and change the instruction translation stage a bit.
- coldtea 6y ago>Exactly how is it much better? It's not the first chip with a 15-25W TDP. When compared to other chips with such a TDP, it's incrementally faster in single core and it trades blows in multicore. It does this at a process node ahead. Being a process node ahead is already "much better", and IS a feat. Where are the process node ahead chips from the competitors? On top of that, it's Apple's own SoC design, it has its own GPU design and co-processors, tons of work into making speed optimizations all around (from the buses to memory handling), and came with a new OS release, a port to the new architecture, a bridge layer for iOS apps, and a perfectly capable x86 translation layer. Rosetta2 alone would be difficult to pull all for most of the industry (Microsoft failed at the same, and their ARM laptops are completely subpar to the M1 machines as well in many levels, despite costing the same or more). >It's not the first PC chip to be an SOC, or to have unified memory, or to have many coprocessors. Which mainstream PC chip in actual laptops/desktops people use has unified memory and doesn't use sockets/DIMMS? >It's also been a decade that we know that instruction sets don't matter. All modern PC processors use a different instruction set and translate from x86. That's a moot point, since they still carry all the x86 baggage in microcode, since tons of bad decisions leading to things like Spectra. ARM doesn't.
- sjwright 6y agoIn terms of GPU cores it seems evident that Nvidia is clearly number one—disregarding any thermal limits—and AMD is clearly number two. It seems Apple is a solid contender for third place now. I don't think Intel or the smaller embedded players (Mali, PowerVR etc) have anything quite so good. Am I missing something?
- sliken 6y agoDunno, do you want a GTX 3070 for $499 or a AMD 6800 for $580? The benchmarks come out at about the same per $ (AMD wins by 10-15%). Sure you could get the 3080 ... or the 6800XT. Pick your price point and AMD and Nvidia will have something close. Sure AMD has nothing to compete at the $1500 price point, but the reviews I've seen claim that the $1500 nvidia is just a big waste not close to justifying the price. PS5, Xbox X/S, and Apple have been preferring AMD lately. Even Tesla seems excited by the RNDA2 cards.
- sjwright 6y agoYes, Nvidia and AMD are #1 and #2. I was contemplating who might be #3 right now.
- sudosysgen 6y ago>Being a process node ahead is already "much better", and IS a feat. Where are the process node ahead chips from the competitors? Not if you still don't beat them cleanly. The end result is what matters. Also, being able to pay more than other people isn't a technical feat. >On top of that, it's Apple's own SoC design, it has its own GPU design and co-processors, tons of work into making speed optimizations all around (from the buses to memory handling), and came with a new OS release, a port to the new architecture, a bridge layer for iOS apps, and a perfectly capable x86 translation layer. Half of that is software, which I didn't mention, and the rest is par for the course. The buses and interconnects in a Zen chip are far superior, which you notice as better I/O. >Which mainstream PC chip in actual laptops/desktops people use has unified memory and doesn't use sockets/DIMMS? You didn't mention Sockets/DIMMS, but literally any AMD APU has unified memory, and use memory DIMMs. Seems like the best of both worlds to me. Just as good memory performance, and it's upgradeable. >That's a moot point, since they still carry all the x86 baggage in microcode, since tons of bad decisions leading to things like Spectra. ARM doesn't. AMD is not affected by Spectre. ARM also uses speculative execution.
- 2muchcoffeeman 6y ago* Exactly how is it much better? It's not the first chip with a 15-25W TDP. When compared to other chips with such a TDP, it's incrementally faster in single core and it trades blows in multicore. It does this at a process node ahead.* Ok. Where can I buy a laptop with similar performance, battery life and cooling then? Same arguments came out with the first iPhone. “It’s just a big screen! Other phones have a big screen!”
- sudosysgen 6y agoSimilar performance and cooling? Buy a Ryzen 4800U laptop. Similar battery life? A good part of it is down to the software, sadly. For the trade-off of 8 instead of 12 hours of battery life, you'll have a repairable and open computer. Seems competitive. You also get upgradeable RAM, a GPU that is actually useful because it has drivers for a useful API, and much more I/O.
- threeseed 6y ago1) M1 is significantly faster than 4800U at single-core. 2) M1 is up to 20 hours of battery life. 3) Not the software since the existing Intel based Macs share the same OS.
- sudosysgen 6y agoM1 is somewhat slower in multicore, and has a fifth of the I/O performance. They are comparable. Importantly, you're comparing the M1 in a Mac Mini, to the 4800U in a laptop. Comparing the M1 in the Air, for example, the 4800U is quite a bit faster in multicore and the single core performance advantage shrinks. As for 2 and 3, you're comparing to AMD, not to Intel. AMD CPUs have significantly better power efficiency than Intel. If you want to compare battery life without taking into account software, compare power consumption. The Zen chip uses slightly more power than the M1 on a laptop.
- coldtea 6y ago>As for 2 and 3, you're comparing to AMD, not to Intel. AMD CPUs have significantly better power efficiency than Intel. Still nowhere near as the same battery life has been seen in any AMD laptop. Not even in ones with less perf. Seems that you're grasping at straws at this moment :-)
- willyt 6y agoRead the anandtech article again. 15-25w is the power consumption of the whole Mac mini measured at the wall plug. I.e it includes power supply, ssd, gpu comparable to RX560, external buses etc etc. If you want an x86 chip that beats its performance by a significant margin you have to pay more for just the chip than for the whole Mac mini.
- sudosysgen 6y agoThe 4800U also includes a GPU, and 10W is very generous for SSD and RAM. Since both are SoCs, external buses and so are done in the chip.