As far I’m aware hpet isn’t deprecated neither does anything bad to the system. Works as way to proper measure the system timers in more accurate way, hardware vendors improved their overall timers but several softwares and vendors suggest to keep them enabled for better accurate data of the current system state.
Software who rely on those timers can generate invalid data by using ticks that doesn’t represent accurate data creating a misconception that either benefits or hurts the performance.
Some old cases impacted previously but not in the way that most folks consider. Any “real” gain it’s just by a placebo effect of using non-accurate timers.
When a second doesn’t work like a second, several softwares can behave oddly on how they “calc” the proper time elapsed, making the mileage vary on the data but not on the process itself.
If your game had stutter before, wasn’t because hpet but something else messing with the performance itself that lost it’s track of time because of the lack of accuracy.
Maybe a service, update or some kind of unwanted software running on the machine. It’s not the job of hpet to improve or hurt your performance, just a reliable way to give accurate data when softwares query for it. The pc have several timers for diverse purposes, disabling one of them would feed certain softwares with an innacurate data.
High precision event timer are what explicit says, gives high accuracy on that query for more accurate measurement.
It’s like using Pi as 3,14 vs 3.141592653589793238…
If you doing several calculations and doing measurements the results can vary and can cause jitter on measurements.
If you lose precision you create an error margin that can mess with the whole calculation or just show a visual feedback of irregular data. Most folks claims about fps increase but those folks forget that they don’t know what timer they used on the other measurement that a second on that second timer could be different than a second on the hpet.
If you had stutters before, was because something else. The lack of better precision timers can be masking it. Either by improperly reporting to drivers which could be a driver issue or setup issue of it or just some background task/process that had abnormal behavior in your system.
If your measurement it’s solely on “feeling” or “fps showed” are pretty much a placebo or some background task with different behavior.
If you properly used a independent gear to proper gauge the value outside your system, you could check the difference between non-hpet, hpet and the hardware independent level timers and you would notice that the hardware precision and hpet would be somewhat similar and non-hpet would provide weird values based on the application measurement.
Shadow of mordor it’s a classic example of that. Generating abnormal values when hpet gets disabled, but overall game experience doesn’t change, in fact gets worst even with more fps because the jitter of input lag goes all over the place.
Also input lag by example varies based on variance of those timers, if you can’t calc them in a accurate manner, mostly you’re deceived by the values that are generated and sometimes you can actually get worst results even with better values if that value was used for something really important.
This source links some useful data to proper gauge, most of them within margin of error and mostly showing none to marginal benefits at best with the price of what appears to miss calculation of cpu clock also.
https://community.amd.com/t5/general-discussions/disable-hpet-in-windows-11-on-a-laptop-with-ryzen-5900hx-rx6800m/m-p/496400
A MVP has a post for this thread:
https://www.reddit.com/r/Amd/comments/eoutjw/a_reminder_disabling_hpet_is_snake_oil/
That relates to this:
https://www.reddit.com/r/Amd/comments/cgi1td/3700x_analysis_pt_deux_notes_about_hpet_and_smt/euie3m2/?context=8&depth=9
The intensive/complete test:
https://www.reddit.com/r/Amd/comments/cfdog5/realworld_3700x_analysis_and_other_ramblings/
Folks still disagrees on some topics, still. The delta shows almost no variance within the error margin. Abou the mouse/keyboard issues could be related to the gear, driver, software or specific vendor.
As far I’m aware there’s no isssue whatsoever with or without hpet on or of on overwatch both tested on g4560, fx-8350 and ryzen 7 5800x.
Also on those gpus: gtx 1080 zotac amp edition and rx 6900xt lc sapphire (OEM)
Using both logitech m90 mouse but also microsoft wireless mouse from combo 5500.
The game was loaded from adata falcon cacheless nvme ssd, without any OS on it, only the game.