Intel skepticism aside, I hope they can deliver on this. M-series Macs seem streets ahead in terms of battery life right now and it doesn’t feel great buying any other portable.
Apple makes low clocked, very wide SoCs, and are always the first customer of the most cutting edge silicon node. This is very expensive. And Apple can eat it with their outrageous prices.
Intel (and AMD) go more for “balance,” with smaller cheaper dies and higher peak clocks. Their OEMs also “cheap out” by bundling a bunch of bloatware that also drains the battery to pad margins. You can find PCs with big batteries and better stock configs, but these are more expensive.
AMD is only just now getting into the “premium” game with the upcoming Strix Halo chip (M2 Pro-ish spec wise). Intel isn’t there yet, but there are rumors they will as well.
Even if you remove all that crap, battery life is nowhere near the same vs the M-series chips. So while it may be a problem, it’s still not anywhere close to the reason battery life sucks.
It can be if you run linux and throttle the chips. Even my older G14 last a long time, as the AMD SoCs are great, it can run fanless throttled down, and it just has a straight up bigger battery than razor thin Macs.
But again, it’s just not configured this way in most laptops, which sacrifice battery for everything else because, well, OEMs are idiots.
I don’t, I just run stock. I run an E495 and get something like 3-5 hours battery life, depending on what I’m doing, and after a few years of ownership, I still get around 3 hours battery life.
Current gen MacBooks have massive batteries. The MacBook Pro 14 inch is 70-73Wh, same as your G14, and the 16 inch MBP is 100Wh, the legal limit to take on an airplane. Even the 13inch air, apple’s thinnest and smallest, is still 52Wh.
What specific driver and linux tools do you use to throttle your CPU?
Also throttling often produces the opposite result in terms of extended battery life as it likely takes more time in the higher states to do the same amount of work whereas running at a faster clock speed, the work is completed faster and the CPU returns to a lower less energy using state quicker and resides there more of the time.
I would be interested to hear your results. Have you done any tests comparing a throttled versus throttled system with the tools you are using?
On my G14, I just uses the ROG utility to disable turbo and make some kernel tweaks. I’ve used ryzenadj before, but its been awhile. And yes I measured battery drain in the terminal (but again its been awhile).
Also throttling often produces the opposite result in terms of extended battery life as it likely takes more time in the higher states to do the same amount of work whereas running at a faster clock speed, the work is completed faster and the CPU returns to a lower less energy using state quicker and resides there more of the time.
“Race to sleep” is true to some extent, but after a certain point the extra voltage one needs for higher clocks dramatically outweighs the benefit of the CPU sleeping longer. Modern CPUs turbo to ridiculously inefficient frequencies by default before they thermally throttle themselves.
I thought so too, but if Apple is getting more than 2x the battery life vs competitors while having a more dense screen, then I suppose it’s not as significant as I had thought.
There were some benchmarks that showed Ryzen getting very close and in some cases beating with Zen 4 based Z1 Extreme already. They just aren’t in laptops.
I think in terms of actually doing stuff AMD is close in terms of power draw (W/performance) but it’s the little things like going to sleep and while completely idle that the entire MacBook draws so little power that needs to catch up – and that’s not entirely on the processor.
Intel skepticism aside, I hope they can deliver on this. M-series Macs seem streets ahead in terms of battery life right now and it doesn’t feel great buying any other portable.
Honestly, a lot of that is budget.
Apple makes low clocked, very wide SoCs, and are always the first customer of the most cutting edge silicon node. This is very expensive. And Apple can eat it with their outrageous prices.
Intel (and AMD) go more for “balance,” with smaller cheaper dies and higher peak clocks. Their OEMs also “cheap out” by bundling a bunch of bloatware that also drains the battery to pad margins. You can find PCs with big batteries and better stock configs, but these are more expensive.
AMD is only just now getting into the “premium” game with the upcoming Strix Halo chip (M2 Pro-ish spec wise). Intel isn’t there yet, but there are rumors they will as well.
Even if you remove all that crap, battery life is nowhere near the same vs the M-series chips. So while it may be a problem, it’s still not anywhere close to the reason battery life sucks.
It can be if you run linux and throttle the chips. Even my older G14 last a long time, as the AMD SoCs are great, it can run fanless throttled down, and it just has a straight up bigger battery than razor thin Macs.
But again, it’s just not configured this way in most laptops, which sacrifice battery for everything else because, well, OEMs are idiots.
I don’t, I just run stock. I run an E495 and get something like 3-5 hours battery life, depending on what I’m doing, and after a few years of ownership, I still get around 3 hours battery life.
Current gen MacBooks have massive batteries. The MacBook Pro 14 inch is 70-73Wh, same as your G14, and the 16 inch MBP is 100Wh, the legal limit to take on an airplane. Even the 13inch air, apple’s thinnest and smallest, is still 52Wh.
What specific driver and linux tools do you use to throttle your CPU?
Also throttling often produces the opposite result in terms of extended battery life as it likely takes more time in the higher states to do the same amount of work whereas running at a faster clock speed, the work is completed faster and the CPU returns to a lower less energy using state quicker and resides there more of the time.
I would be interested to hear your results. Have you done any tests comparing a throttled versus throttled system with the tools you are using?
On my G14, I just uses the ROG utility to disable turbo and make some kernel tweaks. I’ve used ryzenadj before, but its been awhile. And yes I measured battery drain in the terminal (but again its been awhile).
“Race to sleep” is true to some extent, but after a certain point the extra voltage one needs for higher clocks dramatically outweighs the benefit of the CPU sleeping longer. Modern CPUs turbo to ridiculously inefficient frequencies by default before they thermally throttle themselves.
Isn’t the screen eating most of the power in laptops? I just have an old T490 that I don’t use very much so I might be not that well informed.
I thought so too, but if Apple is getting more than 2x the battery life vs competitors while having a more dense screen, then I suppose it’s not as significant as I had thought.
Denser screen shouldn’t use more power though? Lighter, brighter, larger yes, denser IMO not so much.
Denser takes a more powerful gpu and higher data rates to drive more pixels.
Oh yeah, the gpu would have to work more.
There were some benchmarks that showed Ryzen getting very close and in some cases beating with Zen 4 based Z1 Extreme already. They just aren’t in laptops.
I think in terms of actually doing stuff AMD is close in terms of power draw (W/performance) but it’s the little things like going to sleep and while completely idle that the entire MacBook draws so little power that needs to catch up – and that’s not entirely on the processor.
Qualcomm is making some m1 level chips (for computers)