Modern mobile processors are extremely powerful, and way more power efficient than desktop chips. Many Linux distros have ARM variants.
And (AFAIK) it’s mostly because they are tuned for that, whereas laptop processors outside Apple have some overzealous boosting configs/clocks these days.
You are not wrong though. Developing and supporting them (as a company) would be very labor intense, so no one really wants to make that jump beyond Qualcomm/windows.
I think that is the biggest issue with ARM. x86 is so wide spread and backwards compatible that it’s just more cost effective to eat the efficiency losses. If Arm ever gets as big as x86 I can see it dominating all but maybe the Gaming PC space. Proton compatibility layer is already compatible with it however so it might even take off there one day as well.
It just needs a platform desktop users want to buy. There isn’t one (as Qualcomm and Altera aren’t very appealing TBH), and there probably won’t be since Nvidia/AMD/Intel are so far ahead in graphics.
Once that happens though (say AMD starts releasing ARM CCDs?), the switch will not be as hard as you’d think.
I agree, maybe my wording was confusing but yeah I see the biggest problem for ARM is market penetration. Functionally it’s there already but since so few use it no one is making anything for it and since no one is making anything for it no one uses it.
I will say, I remember a little ARM/x86 assembly, and I think ARM SVE2 and some other bits about the ISA are the coolest thing since sliced bread. Backwards compatibility, going forward from ArmV9, is so much better, if that makes any sense, and the same compile binary should “scale” to huge and tiny cores really well (whereas a lot of x86 assumes you’re running something ancient, and AVX is a total mess).
Modern mobile processors are extremely powerful, and way more power efficient than desktop chips. Many Linux distros have ARM variants.
And (AFAIK) it’s mostly because they are tuned for that, whereas laptop processors outside Apple have some overzealous boosting configs/clocks these days.
You are not wrong though. Developing and supporting them (as a company) would be very labor intense, so no one really wants to make that jump beyond Qualcomm/windows.
I think that is the biggest issue with ARM. x86 is so wide spread and backwards compatible that it’s just more cost effective to eat the efficiency losses. If Arm ever gets as big as x86 I can see it dominating all but maybe the Gaming PC space. Proton compatibility layer is already compatible with it however so it might even take off there one day as well.
It just needs a platform desktop users want to buy. There isn’t one (as Qualcomm and Altera aren’t very appealing TBH), and there probably won’t be since Nvidia/AMD/Intel are so far ahead in graphics.
Once that happens though (say AMD starts releasing ARM CCDs?), the switch will not be as hard as you’d think.
I agree, maybe my wording was confusing but yeah I see the biggest problem for ARM is market penetration. Functionally it’s there already but since so few use it no one is making anything for it and since no one is making anything for it no one uses it.
Yeah exactly!
I will say, I remember a little ARM/x86 assembly, and I think ARM SVE2 and some other bits about the ISA are the coolest thing since sliced bread. Backwards compatibility, going forward from ArmV9, is so much better, if that makes any sense, and the same compile binary should “scale” to huge and tiny cores really well (whereas a lot of x86 assumes you’re running something ancient, and AVX is a total mess).