We recently learned that Microsoft is adding a "Low Latency Profile" feature that effectively boosts CPU clock speeds temporarily during app startup so that software doesn't feel sluggish. This is a part of Redmond's Windows K2 efforts which sees the company revamping various parts of Windows 11 for better reliability and performance.
Some enthusiasts have already figured out a way to enable this low latency profile, and the performance gains are quite significant in some cases. While that sounds great on paper, I believe it highlights a bigger problem with Windows 11.
The crux of the matter is that temporary CPU bursts are just a hack to mask poorly optimized software. If your OS needs to boost processor speeds just to feel usable, perhaps the problem is the software, not the compute power of the hardware running it.
This pratice simply encourages developers to keep writing bad code and incentivizing users to shell out more money to purchase the latest and greatest in CPU, just so that they can get those few extra clock cycles to make Windows 11 feel fast.
Perhaps the even bigger problem is that Microsoft's own first-party software is guilty of being poorly optimized. The examples that we have seen here show performance gains in the Start menu, File Explorer, and Outlook. Why should third-party developers be incentivized to write optimized software when the company building an entire OS can't be bothered to do so either?
It's worth highlighting that this is not a mindset change that has happened overnight at Redmond. Some of you may remember that last year, Microsoft began testing an implementation that would pre-load File Explorer into memory so that it launches faster when a user opens it. Rather than actually figuring out what went wrong with File Explorer performance between Windows 10 and Windows 11, Microsoft simply decided to once again use a hardware hack to mask the problem.
To be fair to Microsoft, the rollout of a Low Latency Profile does emphasize a few good aspects too. It indicates that the firm is aware of the performance issues in Windows 11 and is acknowledging them, scheduler-level enhancements like these are still technically impressive, and reducing the perceived latency is still a good move from an end-user perspective. Early testing also indicates that there is no notable degradation in battery life either, which is why Microsoft thinks that this is worth the effort.
However, it also creates an environment where developers don't care about memory leaks and unoptimized software because they know that Microsoft will just throw compute at the problem to brute-force better performance. It also means that older PCs will begin to feel obsolete sooner because they don't pack that extra compute power that can compensate for bad software. As a result, customers will tend to gravitate towards the latest internals, which will lighten their wallet considerably.
Overall, there's nothing wrong with the engineering of the Low Latency Profile solution itself, it's just that it deprioritizes performant and efficient code, which may lead to bigger problems when even extra CPU cycles are not even enough to compensate for poorly optimized code. The best-performing software isn't the one that temporarily maxes out your CPU, it's the one that never needed to. With its latest approach, Microsoft is simply treating the symptoms, not the disease.