The longevity of M2 MacBook batteries is often compared to Windows laptops, but this isn't always the case.
The longevity of M2 MacBook batteries is often compared to Windows laptops, but this isn't always the case.
Hello there. Let's talk about how much power laptops actually use today. I'm comparing the M2 chip to my old power-hungry 2000 and the i7-10875H in a Dell XPS 9700. People say the M2 is incredibly efficient—nothing else can match it. It’s roughly 40% better than the next best option (the AMD-6800U). But is that true? What really counts for efficiency? I’ve spent countless hours testing, mostly to extend my laptop’s battery life. I tweaked BIOS settings, adjusted voltages, tuned frequencies, worked with liquid metal, and tried different cooling methods. My goal was to see how much power the CPU actually needed and how Windows behaved under various settings.
Here are some key points that highlight why the M2 MacBook stands out:
- The system’s overall power use is much lower. On a typical YouTube stream at 4K, 60fps, you’d normally see around 7–8 watts. With the M2, it drops to under 5 watts—about 3 watts just for the CPU itself.
- Integrated components handle most of the work, so there are fewer losses from VRMs, RAM, SSDs, screens, or extra chips.
- The GPU is optimized and doesn’t waste energy processing huge numbers of pixels like older Intel chips do.
- Even with modern Windows, efficiency remains strong—MacOS uses roughly 20% less power than Windows 10 for similar tasks.
All these factors together mean MacBooks can save a lot of energy. For example, my high-end 10875H running at 4K still draws only about 14 watts. That’s a massive improvement over older systems.
It’s not just about the CPU anymore. The whole design—from power delivery to cooling and software—works together to minimize waste. That’s why battery life has improved so much, even without bigger batteries or more powerful hardware.
I’m not a big fan of Apple’s ecosystem, but I see the point in efficiency. Still, I’ll wait until my current model is no longer reliable before upgrading.
Apple actually reduces the voltage on their intel chips, which helps them consume less power. In Windows, this feature is removed. After applying a similar adjustment, performance improves noticeably within about 45 minutes. However, it still highlights the shortcomings of Windows, especially during heavy tasks. Absolutely. For office work, I don’t need my CPU to run at full speed like it does on battery mode. That 1.5GHz setting before was perfectly fine for me—it didn’t cause any fan noise. The rise in power usage might come from charging initially, then active cooling kicks in, along with more data processing and GPU performance. Some displays also brighten a bit. I notice laptops are starting to adopt better battery efficiency, likely due to Apple’s increased pressure. We’re finally seeing changes in hdd laptops, which could lead to more efficient overall systems. There are models now matching the battery life of a MacBook, and I hope they become more widespread.
Did you verify the initial battery level when you received your ROG Strix G17? It came with a 60Wh pack, but you had 66Wh installed. After nearly two years, there were over 150 charge cycles, and you lost about 3% overall—mostly because you didn’t charge it for a month before using it again. You usually kept it at 80% and charged it twice daily. With a full charge, you get around 5–6 hours of video playback, and about 4.5 hours at 80%. That’s how the numbers add up.
Windows struggles a lot with managing processes. Many tasks run quietly in the background, often keeping services active so that when you launch something again, things start faster. These background services usually don’t need much CPU power, but as more of them accumulate, they add up. Windows isn’t very energy-efficient compared to MacOS. I wish it were. Regarding the high power usage when connected—once plugged in, everything seems to draw as much power as you want. That shouldn’t happen. We’re making progress. The new Intel EVO standard is a positive step. AMD’s offerings in ultrabooks are impressive. I believe we’ll reach 12 hours of regular use soon, which would be ideal. Right now, 8 hours is still seen as the typical benchmark.
Many believe Apple has long provided a solid method for developers to assign priority levels to tasks within their applications. This approach has encouraged developers to structure their work this way, ensuring that most operations are properly categorized. It aids efficiency even on older Intel processors by allowing background processes with lower priority to be consolidated and executed when the CPU becomes available—often within milliseconds. This capability enables the operating system to manage resources intelligently, keeping interactive elements like the user interface responsive.
Apple has also refined its strategy by assigning background or low-priority tasks to specific cores that aren’t actively boosting performance. This ensures that when a high-priority task demands maximum CPU power, the system can allocate full speed for it. In Windows, while some APIs exist for labeling tasks, few applications utilize this feature effectively. The OS typically evaluates whether a task belongs to the foreground to prevent unintended UI freezes, such as pausing a UI element for just a few milliseconds.
Extreme measures are taken to optimize power consumption. Even the lowest-priority background jobs may take seconds to activate after scheduling. This labeling technique was initially developed for Intel CPUs and is now widely adopted on Apple’s M1/M2 chips, which feature multiple cores of varying sizes. Intel, on the other hand, primarily uses a foreground/background distinction based on application status rather than developer-defined labels, limiting the granularity compared to Apple’s system.
Many are discussing CPUs, but YouTube personalities focus mainly on performance numbers. Apple’s GPU performance is actually better when measured by perf/W than Intel, AMD, or Nvidia. This is because Apple opts for a tile-based deferred rendering system, which reduces memory traffic significantly in most tasks. Modern GPUs consume most of their power from memory, cache, and subsystems. Using a TBDR approach for well-optimized applications can cut memory bandwidth needs by up to 10 to 100 times. This leads to huge savings on memory controllers and improves cache efficiency since data isn’t constantly flooded in from fragmented render outputs. AMD, Intel, and Nvidia are unlikely to adopt TBDR fully, as current games aren’t built to exploit this method, leaving much of the silicon underutilized. However, some TBDR-inspired work may emerge for future multi-die GPUs from AMD or Nvidia, aiming to minimize die-to-die communication.
The system's background management and priority handling are still unclear to me, but I understand the issue. Windows struggles to decide which applications run in the background versus foreground. When multiple programs launch, it tends to prioritize only essential services and native apps, leaving everything else free to operate. My friend using an M1 MacBook Pro 13 has the machine running continuously without shutdowns, maintaining optimal performance with full RAM utilization and SSD storage. His battery performance remains unaffected, even under constant usage. This impresses me greatly. My own experience shows that keeping all applications open can halve my battery life, regardless of workload. MacOS seems to handle this better than Windows, though I hope future updates improve scheduling efficiency. Some manufacturers are already customizing BIOS to limit CPU activity on performance cores, which would benefit Windows more in the long run. Dell has made significant progress with their XPS models this year, using tuned BIOS that aggressively manages core usage and transitions between e-cores and p-cores seamlessly. This results in impressive battery life—up to 13 hours compared to just 8 last year—and a smooth experience. For most tasks, it stays under 10 watts with most cores in C-state, but can reach 130W when heavily loaded. It’s a solid design, especially for high-performance workloads. If scheduling improvements aren’t coming soon, we should aim for processors that conserve power more effectively.
But how smart your scheduler will be depends on whether developers assign priorities to tasks. Otherwise, you might face: 1) UI issues if low-priority tasks are mistakenly marked high, or 2) excessive battery drain. There are established solutions that have existed for years, and Microsoft provides APIs for this. The problem lies in Microsoft’s difficulty in getting teams to implement such features. Most smart scheduling approaches I’ve seen rely on opening an app triggers it as high priority, which is a poor way to judge importance. Another challenge Microsoft faces is managing windows and threads. On modern operating systems like macOS, Linux, or BSD, creating many threads comes with minimal cost, allowing thousands of threads—sometimes over 100 in apps like Finder. Historically on Windows, spawning threads was expensive, often capped at around 64 per process. This meant most apps ran with fewer than five threads, and developers would limit threads to match CPU cores. Even now, the OS encourages using virtual threads within a single core to improve performance, since creating many threads can be costly due to thread creation overhead.
I've been seriously thinking about moving away from Windows for work and study tasks. I've been considering MacOS since before the M1 chip, maybe even since 2016. It wasn't a good choice back then because the battery improvements weren't significant. Now, with better battery life, I'm leaning toward starting with Apple devices first—maybe an M2 or MX laptop—and eventually switching to Windows. The cost will be around four thousand dollars, but honestly, I'm tired of poor laptop batteries and all the issues Windows has on mobile. It feels like a big decision right now. The only real factor is money. In my country, Apple is much cheaper than buying locally in the US, so it makes sense to pick up one there and bring it back home. It's really frustrating. Thank you for breaking everything down so clearly for me. I thought Windows was better for this, but it turns out it isn't.
Modern Windows offers many similar capabilities to macOS, yet Microsoft’s main challenge has always been their hesitation to actively support new technologies or push developers toward them. Apple tends to phase out old APIs while keeping existing apps functional for years, but developers needing to adopt newer ones must discontinue the older interface within a year of its deprecation. This encourages gradual adoption rather than sudden change. In contrast, Microsoft has rarely forced such transitions, allowing developers more freedom. This approach lets teams use widely adopted open-source tools that rely on APIs from the past, making it easier to maintain compatibility. However, it also creates difficulties in motivating developers to move forward and poses significant challenges for Microsoft, which must carefully manage legacy systems over decades, as even minor fixes can impact critical applications. Rumors suggest Microsoft may be shifting toward lightweight virtual machines for older software, potentially leading to a more streamlined core OS in the future. We’ll see how this unfolds.