MSI Afterburner missing interrupts

[Majestic]

So I noticed this when I was making my overhead analysis, and also retroactively to my 480 test results. Whilst completing the same tests, the 3.6ghz graphs are all shorter eventhough I replicated the tests near perfectly.

Especially Rise of the Tomb Raider stood out. The pollingrate I used was set to 100ms to get the most accurate graphs, but somehow when the CPU is hitting 100% on all threads, the pollingrate of MSI Afterburner is put on the back ...burner ?

 

Spoiler

 

 

 

Is MSI Afterburner simply missing the interrupts because the games have higher priority? Or am I missing something. I can't find anyone else with a similar issue, which I find odd that nobody else noticed this.

[zMeul]

that's why no one with a whiff of decency in GPU testing don't use on machine data recording tools, where possible

[Majestic]

4 minutes ago, zMeul said:

that's why no one with a whiff of decency in GPU testing don't use on machine data recording tools, where possible

Yeah I guess, but I don't have the equipment for proper testing

[zMeul]

1 minute ago, Majestic said:

Yeah I guess, but I don't have the equipment for proper testing

most people at home don't

[Majestic]

2 minutes ago, zMeul said:

most people at home don't

Sadly the people who do, make other mistakes in testing methodology. Or simply refuse to test stuff like Driver overhead for fear of upsetting fanbases.

I'd really like to see a proper, objective test, being done on driver overhead and to explain the shift I saw myself in frame delivery (Nvidia -> AMD -> Back to Nvidia). But alas, not even PCPer, Gamers Nexus or Digital Foundry is willing to touch this touchy subject.