Because you’re saying they’re only good for simple tasks, yet they’ve been running as powerful of software as x86 for years now, between Apple’s M chips, the various server based chips, and even Windows (although its still lackluster in many regards there, but I think there’s native Adobe ARM on Windows apps now). There’s been high core count server ARM chips for years already.
From 4+ years ago:
Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
www.anandtech.com
Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
www.anandtech.com
And I believe multiple major server/cloud companies (Amazon, I forget who all else) have been making their own ARM server chips as well. Yet you’re wondering if ARM will ever be used for that? Its already happening.
I didn’t present market figures, and I honestly couldn’t care less, was just pointing out that you’re likely wrong on both of your main points, that AMD is dominating HEDT (they are certainly dominating Intel there in most if not all performance metrics, but Intel is barely even competing in that market these days, both are just pushing their server chips to that market anyways, but Intel is still probably outselling AMD there simply because of inertia from their overall dominant market position and most of the HEDT market being through OEMs like Dell that offer mostly Intel based workstations).
Wouldn’t it be until ARM chips running Windows match or beat x86? Since that was the entire supposition is that ARM chips are inferior to x86. If you’re deadset on only Windows mattering there, for whatever reason you’re doing that.
There’s a reason I said MacBook Pros being used as workstations, stuff like the Max and Ultra based chips, since it would obviously be silly to compare all MacBook sales. Could add in Mac Studio (and Mac Pro since they do have M chips now) as well if you really wanted to compare. Frankly I don’t care at all about that, was merely pointing out that the claim that ARM chips are only used for simple tasks, and will never be used for workstation or server is just bafflingly ignorant of where things have been for years.
Also, Nvidia is now sellling an AI workstation that uses their ARM based chips. I haven’t see how their CPUs do and the GPU is doing the heavy lifting there anyway. Which is also the case with a lot of workstations, making these comparisons even further muddled, but also further representing that x86 vs ARM is not an inherent issue for most pro workloads these days.