3 ms·

It's good keeping in mind the ThinkPad X13s was released with a Snapdragon as a way to prepare for Snapdragon X. Performance was not good, and the CPU was a bi

by nextos 3y ago

It's good keeping in mind the ThinkPad X13s was released with a Snapdragon as a way to prepare for Snapdragon X. Performance was not good, and the CPU was a bit hot (but fanless). Linux support was imperfect, but for a first attempt it is not too bad. A M2-like machine running Linux, with good support, would be fantastic as an ultra-mobile laptop. However, AMD is catching up in terms of performance per Watt, and has already caught up in terms of raw performance. Will a Snapdragon X have a sufficiently wide market niche?

tgtweak 3y ago
I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. Performance per watt is only part of the picture, power-at-idle is also a substantial contributor to battery life and total consumption. Further to this, the fact that almost all subsystems are on-chip for an arm processor and don't require separate off-die (typically older process node/lower efficiency) chips for this means that the entire system efficiency can be considerably higher even if the per-instruction performance per watt is marginally lower. Anecdotally, the operating system also determines, considerably, the system efficiency. Case in point - I installed chromeos (brunch) on my HP ultrabook (originally a windows 10 machine, 8th gen i7 "U" processor) and the battery lasts 12-18 hours vs 6-8 with windows doing essentially the same light workload.
- vardump 3y ago
  > I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. CPU backend side x86 and ARM is almost identical. It's mainly just the front end where the things are different. Like instruction decoding.
- The_Colonel 3y ago
  > I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. This is often repeated, but very thinly substantiated. ARM originally wasn't engineered for low power usage, just "normal power usage" in PC-like computers. It found its niche in low power devices mainly because the larger chips got dominated by other ISAs and weren't much interested in low margin embedded space. It's very unclear if ARM as an ISA has any intrinsically meaningful advantage in power consumption. Modern chip design kind of converged and most of them work on similar principles like decoding the ISA instructions into microcode which is then executed. x86 decoder might be more complex than an ARM one, but the difference on a chip is small and likely meaningless in terms of power consumption.
  - fanf2 3y ago
    ARM 1 was designed to be a low-power CPU, because they wanted to be able to use a cheap plastic package. Famously, they forgot to wire the first sample to its power supply, but it was still able to run of the power leakage from its other pins. https://arstechnica.com/gadgets/2022/09/a-history-of-arm-part-1-building-the-first-chip/ https://arstechnica.com/gadgets/2022/09/a-history-of-arm-par...

by nextos 3y ago

tgtweak 3y ago

I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. Performance per watt is only part of the picture, power-at-idle is also a substantial contributor to battery life and total consumption. Further to this, the fact that almost all subsystems are on-chip for an arm processor and don't require separate off-die (typically older process node/lower efficiency) chips for this means that the entire system efficiency can be considerably higher even if the per-instruction performance per watt is marginally lower. Anecdotally, the operating system also determines, considerably, the system efficiency. Case in point - I installed chromeos (brunch) on my HP ultrabook (originally a windows 10 machine, 8th gen i7 "U" processor) and the battery lasts 12-18 hours vs 6-8 with windows doing essentially the same light workload.

vardump 3y ago
> I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. CPU backend side x86 and ARM is almost identical. It's mainly just the front end where the things are different. Like instruction decoding.
The_Colonel 3y ago
> I don't think x86 will ever get to the level of power efficiency that ARM has, simply due to the fact the arm cores and instructions are all built and engineered for low power usage. This is often repeated, but very thinly substantiated. ARM originally wasn't engineered for low power usage, just "normal power usage" in PC-like computers. It found its niche in low power devices mainly because the larger chips got dominated by other ISAs and weren't much interested in low margin embedded space. It's very unclear if ARM as an ISA has any intrinsically meaningful advantage in power consumption. Modern chip design kind of converged and most of them work on similar principles like decoding the ISA instructions into microcode which is then executed. x86 decoder might be more complex than an ARM one, but the difference on a chip is small and likely meaningless in terms of power consumption.
- fanf2 3y ago
  ARM 1 was designed to be a low-power CPU, because they wanted to be able to use a cheap plastic package. Famously, they forgot to wire the first sample to its power supply, but it was still able to run of the power leakage from its other pins. https://arstechnica.com/gadgets/2022/09/a-history-of-arm-part-1-building-the-first-chip/ https://arstechnica.com/gadgets/2022/09/a-history-of-arm-par...