3 ms·
Is it me or is a 6 year old product on 14nm against whatever isn't exactly the most level of comparisons?
by 6SixTy 1y ago
Is it me or is a 6 year old product on 14nm against whatever isn't exactly the most level of comparisons?
- galaxy_gas 1y agoIts been across the board for me. Have ryzen 6xxx-8xxx mini PCs that are 45-60W and my 5 year old M1 Mini is comparable at a fraction of power draw
- protimewaster 1y agoI would think a low TDP Ryzen chip shouldn't be that much different. Were those the 15W TDP Ryzens?
- galaxy_gas 1y ago30W and 54Ws. iGPU performance is also better on the Mac compared to the non-65W Ryzens. The good thing is that the Ryzen boards can be upgraded on RAM and Storage but that seems to be changing with soldered DDR5 on many mini PCs now.
- conception 1y agoMac Mini is i believe the most efficient consumer computer made ever. It uses the same energy most computers Ethernet ports use.
- TheDong 1y ago> Mac Mini is i believe the most efficient consumer computer made ever My raspberry pi is sitting at 3 watts, and runs linux software just fine. A Mac Mini uses multiple times that, and can barely run any software I care about (i.e. linux). They're both consumer computers, and yes there's a massive difference in their capabilities, but "most efficient ever" is a really strong claim.
- ZaoLahma 1y agoThe raspberry pi is great. Absolutely love them. Been running them as dev boards (robotics) and servers (at home) for 10+ years. Last time I truly, genuinely, felt excited about Christmas was when my SO gave one to me back in 2014 or so. But I wouldn't compare the experience to a Mac Mini, and I wouldn't call a raspberry pi a "consumer computer". It solidly falls into the amateur dev board category. Heck, it doesn't even have a power button.
- alpaca128 1y agoThe claim was about efficiency, not just power draw. In your comparison you talk about software compatibility, but the Mac can run Linux and many open source tools & apps found on it. So the Pi is only ahead if you arbitrarily limit the comparison to a specific distro that's not Asahi or an application like sxiv, and in addition are the kind of user who wants that software and doesn't care about Mac exclusive programs. Which is valid, but has nothing to do with a metric like power efficiency.
- lostlogin 1y agoI’m running Ubuntu on the Mac, in UTM. It works well. Quite why networking from VM is more reliable than networking from the host is a question I’d like to ask Apple. SMB does not reliably start at boot. Imagine if you could turn on a Mac mini, have docker containers start and then have them access network shares reliably. The Ubuntu VM on the same host can. That’s ridiculous.
- lostlogin 1y ago> A Mac Mini uses multiple times that, and can barely run any software I care about (i.e. linux). You’re both right, and wrong. The Mac uses 3-4W most the time. The reason mine is using more is that it has an Ubuntu VM running Pihole (so it is running Linux). This doubles/triples the power usage. If I wasn’t such a chicken, I’d run it on the host in docker, but I’ve learned the pain of having Pihole on the same host as other services. Dead dns is not not good. I really wish I could run Ubuntu on the metal. Something as simple as having docker containers run on boot is unreliable, and the god awful hack I have made to get it to work is so disgusting it’s shameful.
- lostlogin 1y agoThe comparison wasn’t meant to be fair - power usage is why I upgraded. Both are on 10gbe, and I suspect the card in the Nuc was a bit of a power draw.