02-01-2016 08:05 AM - last edited on 03-06-2024 02:57 AM by ROGBot
02-01-2016 09:19 AM
02-01-2016 10:31 AM
Gps3dx wrote:
Can you please provide more details ? can you screen-capture it as video ?
does the music comes from a "line-in" source ?
are you alt+tab from a game ( graphics heavy duty app ) to the desktop ? or is it a pure desktop work alt-tabing ( office and web browsing ( non 3D ))
far fetch idea: read here
02-01-2016 12:11 PM
Gps3dx wrote:
Can you please provide more details ? can you screen-capture it as video ?
does the music comes from a "line-in" source ?
are you alt+tab from a game ( graphics heavy duty app ) to the desktop ? or is it a pure desktop work alt-tabing ( office and web browsing ( non 3D ))
far fetch idea: read here
02-02-2016 07:26 AM
02-02-2016 08:20 AM
iozho wrote:
Nvidia still has to fix the high DPC latency.
02-02-2016 08:50 AM
Gps3dx wrote:
well.. I know that drivers & apps ( nvidia specificly ) can rise your DPC latency - which is a "bug" state of these drivers/apps.
nontheless, the Windows itself is also part of the blam. ( i.e with win8 and above, on same HW, you won't necessary be getting this DCP latency problem )
I know the from win8 above such DCP latency issues from the HW ( I.E ACPI timers ) were solve by the OS itself.
but with win7.... things not so great.
please see the following link: HERE to diagnose simply the issue...
I know today that any such "simple" DPC diagnostic tools does NOT provides the best real/correct info ( at least for win8 and above )- as only windows diagnostic tools ( some sdk ) is the nearest to the "truth".
As a quick fix - please try to use the following: TIMERTOOL to lower your DPC Latency.
As an interesting anecdote, the following provided:
I know that using such tools on my retired Asus G55VW with win8 x64 did not help to get better FPS, as it is a problem with how win8 and above are built ( using this tool actually lowered FPS in "PCMark Vantage" benchmark app ).
I know this, cause I tested it myself with some scientifically technique called "factorial design" - The idea in short is that you perform about 1/10 tests then what you really should do - but from these small quantity but well planned repetitions, you can get results that statistically speaking are precise 99.5% like doing in addition the hole other 9/10 tests.
The DPC latency "fixer" app I used back then was timer-resolution.
I wanted to check the weight effect of DPC Latency fixer app (timer-resolution) on FPS.
As the report was not written fully in English I provide only some of the results:
The test planning:
The results:
The the results:
you can see the the coefficients before "timerResolution" is much smaller then resolution it-self.
Thus don't think that "fixing" DPC latency fixer apps will help you gain a lot of FPS, if at all 🙂