It may be the software problem as well. On Windows I regularly need to find which new app started to eat battery like crazy. Usually it ends up being something third-party related to hardware, like Alienware app constantly making WMI requests (high CPU usage of svchost.exe hosting a WMI provider, disabling Alienware service helped), Intel Killer Wi-Fi software doing something when I did not even know it was installed on my PC (disabling all related services helped), Dell apps doing something, MSI apps doing something... you get the idea.
It seems like a class of problems which you simply can't have on macOS because of closed ecosystem.
Without all this stuff my Intel 155H works pretty decently, although I'm sure it is far away from M-series in terms of performance.
The Mac ecosystem isn’t as closed as you’re alluding to. You can easily download unsigned binaries and run them. Furthermore, if you’re looking for a battery hog, look no further than Microsoft Defender, Jamf Protect, and Elasticbeat. All 3 of those are forcibly installed on my work laptop and guzzle up CPU and battery.
> You can easily download unsigned binaries and run them
It's definitely becoming less easy over time. First you had to click approve in a dialog box, then you had to right-click -> open -> approve, now you have to attempt (and fail) to run the app -> then go into System Settings -> Security -> Approve.
I wanted to install a 3rd party kernel extension recently, and I had to reboot into the safety partition, and disable some portion of system integrity protection.
I don't think we're all that far from MacOS being as locked-down as iOS on the software installation front...
Yep, they will lock all that down. It's been coming for years. Tech companies have learned to do their anti-consumer work slowly and progressively over time instead of dropping it all at once. The whole frog in boiling water thing...
Microsoft is working towards this too. They wish so bad that they were Apple.
> You can easily download unsigned binaries and run them
Of course, but I assume you don't really need to install third-party apps to control hardware. In my case Alienware and Dell bloat came from me setting up an Alienware monitor. MSI bloat came from setting up MSI GPU. Intel Killer stuff just got automatically installed by Windows Update, it seems.
> Microsoft Defender
This one I immediately disable after Windows installation so no problems here :)
On work we get CrowdStrike Falcon, it seems pretty tame for now. Guess it depends on IT-controlled scan settings though.
Re: Microsoft Defender, I’m actually talking about defender on macOS. It is a multi platform product. I hear infosec is pretty happy with it. Me? It uses 100% CPU even when I’m doing nothing. I’m not happy.
Try some of the steps on this page [1]. In particular, enabling real-time protection stats and then adding exclusions for the processes causing the most file scans can help.
What's mad is that you would have thought that Microsoft would use the Surface devices to show hardware manufacturers what could be done if you put some effort in, but I've heard so many horror stories from Surface owners about driver issues
Windows doesn't do it any favors, for sure. Running Linux with with every tweak under the sun for better life still leaves a large gap between x86 laptops and MacBooks, however, and while there's probably some low hanging optimization to be taken advantage of there I think the real problem is that x86 CPUs just can't idle as low as M-series can, which is exacerbated by the CPU not being able to finish up its work and reach idle as quickly.
I wonder if Windows and Linux just can't yet work on heterogeneous CPUs as well as macOS does. Intel chose an interesting direction here, going straight from one to three kinds of cores in one chip. I almost never see LPE cores being used on Windows, and on Linux you have obscure soft like Intel LPMD which I tried but was not able to notice any battery life improvements.
I'm a bit out of my depths here, but I believe a significant contributing factor is how early Apple made multi-CPU Macs available, with the earliest being the summer 2000 revision of PowerMac G4 tower (dual 500Mhz PPC G4s), pre-dating the release of OS X. They made it easier for devs to take advantage of those cores in OS X, because this yielded performance boosts that were difficult to match in the x86 world, which was still heavily single-CPU.
Because the OS and apps running on it were already taking advantage of multithreading, making them efficiency core friendly was easy since devs only had to mark already-encapsulated tasks as eligible for running on efficiency cores, so adoption was quick and deep.
Meanwhile on Windows there are still piles of programs that have yet to enter the Core 2 Duo era, let alone advance any further.
Would this have been with MacOS 7’s Multiprocessing Services? I managed to play with an SMP Mac clone (DayStar Genesis MP), but all I really could do in the end is use some plugins for Photoshop.
For reference, I get 29 tokens/s with the same model using 12 threads on AMD 9950X3D. Guess it is 2x faster because AVX-512 is 2x faster on Zen 5, roughly speaking. Somewhat unexpectedly, increasing number of threads decreases performance, 16 threads already perform slightly worse and with 32 threads I only get 26.5 tokens/s.
Shouldn't they be the same if we are speaking about same precision? For example, [0] shows M4 Max 17 TFLOPS FP32 vs MAX+ 395 29.7 TPLOFS FP32 - not sure what exact operation was measured but at least it should be the same operation. Hard to make definitive statements without access to both machines.
M4 Max doesn't even disclose TFLOPS so no clue where that website got the numbers from.
TFLOPS can't be measured the same between generations. For example, Nvidia often quotes sparsity TFLOPS which doubles the dense TFLOPS previously reported. I think AMD probably does the same for consumer GPUs.
Another example is Radeon RX Vega 64 which had 12.7 TFLOPS FP32. Yet, Radeon RX 5700 XT with just 9.8 TFLOPS FP32 absolutely destroyed it in gaming.
Something is wrong with power governor then. I have an opposite experience, was able to tune Linux on a Core Ultra 155H laptop so it works longer than Windows one. Needed to use kernel 6.11+ and TLP [0] with pretty aggressive energy saving settings. Also played a bit with Intel LPMD [1] but did not notice much improvement.
I also own a 155H laptop using Linux Mint! Would you share your settings with TLP and LPMD? I am not getting not much longer battery life than Windows 11 on it after some tinkering, so seeing somebody else's setup may help a lot. Thanks!
Won't say I got much longer battery life, and even what I got may be as well explained as "TLP made energy profile management almost as good as on Windows, and then Windows's tendency to get a bunch of junk processes seeping on your battery tipped the scales to favor Linux". Also I ended up switching back to Windows because of never-ending hardware issues with Linux, installing it on 155H back in February 2024 was especially rough but even 6 months later I randomly got Bluetooth not working anymore after Ubuntu update.
TLP: don't remember details now, as I recall scaling governor does not do anything on modern CPUs when energy perf policy is used. CPU_MAX_PERF_ON_BAT=30 seems to be crucial for battery savings, sacrificing performance (not too much for everyday use really) for joules in battery. CPU_HWP_DYN_BOOST_ON_BAT=0 further prohibits using turbo on battery, just in case.
LPMD: again, did not use it much in the end so not sure what even is written in this config. May need additional care to run alongside TLP.
Also, I used these boot parameters. For performance, I think, beneficial one are *mitigations, nohz_full, rcu*
Weren't we comparing CPUs though? Those Blender benchmarks are for GPUs.
Yes, but I was asked about Blender GPU.
Blender CPU tasks are highly parallel. AMD's Ryzen Max 395 has great MT performance. It's generally 5-20% slower in CPU MT than the M4 Max depending on the application.
It reproduces for me in FF 142 on Windows. When I first went to https://cloudflare-quic.com/ it said HTTP/3, but after a few hard refreshes it says HTTP/2 and hasn't gone back to 3
I restarted FF and am now seeing something similar. Hard refreshing alternates between 2 and 3, and soft refreshes quickly get back to 3 most of the time
No, I have also observed it in Firefox (via Flatpak) on Fedora Linux 42. When I filed the original GH issue that webcompat-bot turned into this Bugzilla item (https://github.com/webcompat/web-bugs/issues/168913), my full report didn't make it into Bugzilla.
AMX is indeed a very strong feature for AI. I've compared Ryzen 9950X with w7-2495X using single-thread inference of some fp32/bf16 neural networks, and while Zen 5 is clearly better than Zen 4, Xeon is still a lot faster even considering that its frequency is almost 1GHz less.
Now, if we say "Zen5 is the leading consumer CPU for AI" then no objections can be made, consumer Intel models do not even support AVX-512.
Also, note that for inference they compare with Xeon 8592+ which is the top Emerald Rapids model. Not sure if comparison with Granite Rapids would have been more appropriate but they surely dodged the AMX bullet by testing FP32 precision instead of BF16.
It may be the software problem as well. On Windows I regularly need to find which new app started to eat battery like crazy. Usually it ends up being something third-party related to hardware, like Alienware app constantly making WMI requests (high CPU usage of svchost.exe hosting a WMI provider, disabling Alienware service helped), Intel Killer Wi-Fi software doing something when I did not even know it was installed on my PC (disabling all related services helped), Dell apps doing something, MSI apps doing something... you get the idea.
It seems like a class of problems which you simply can't have on macOS because of closed ecosystem.
Without all this stuff my Intel 155H works pretty decently, although I'm sure it is far away from M-series in terms of performance.
reply