Does Vsync help?

[vanished]

Just now, Valkyrie Lenneth said:

actually, framerate + response time + input lag ALL TOGETHER cause the total delay

yes but they're separate things.  The most common mistakes I see are 1. people thinking response time means input lag, and 2. thinking response time gives them any accurate indication of motion blur

[JoostinOnline]

Just now, Ryan_Vickers said:

it's worth noting that framerates, response times, and input lag are all totally different things for those who may not know

Sorry, I didn't mean for it to be taken seriously.

 

But we've all got that one friend who considers winning a CS:GO match more important than being there for the birth of his first-born child.

[Valkyrie Lenneth]

1 minute ago, Ryan_Vickers said:

yes but they're separate things.  The most common mistakes I see are 1. people thinking response time means input lag, and 2. thinking response time gives them any accurate indication of motion blur

yeah and that is incorrect

 

it is

 

fps + response time + input lag

 

input lag however is determined on many factors ,, such as ... wireless connections, hardware polling, ram cas latencies ( because ram also has a delay ), cpu rendering time . and way more ( such as software delay... take windows aero for example ) ^^

 

 

[vanished]

Just now, Valkyrie Lenneth said:

yeah and that is incorrect

 

it is

 

fps + response time + input lag

 

input lag however is determined on many factors ,, such as ... wireless connections, hardware polling, ram cas latencies ( because ram also has a delay ), cpu rendering time . and way more ( such as software delay... take windows aero for example ) ^^

I know it's incorrect, that's why I called them mistakes  

 

One thing I will say though is I highly doubt anyone can notice the difference RAM CAS latency makes in gaming... that's not on a level that affects measurable delay, that's just overall performance at that point.  You're not going to feel an extra 10 nanoseconds  

[Valkyrie Lenneth]

1 minute ago, Ryan_Vickers said:

I know it's incorrect, that's why I called them mistakes  

 

One thing I will say though is I highly doubt anyone can notice the difference RAM CAS latency makes in gaming... that's not on a level that affects measurable delay, that's just overall performance at that point.  You're not going to feel an extra 10 nanoseconds  

the ram cas latencies are somthing that is overlooked by many ppl... im personally an extreme ram tweaker and those 10nanoseconds reduced in the ram to cpu speed can actually make like a input delay difference of like 5-10ms easily  ( the nanoseconds are basicly PER TASK, trust me there is millions  of tasks going on on that part ^^ )

 

the ram situation is extremely advanced to put in a short text and would require a post of atleast a whole page probably to explain ( way to much work for now )

 

but yes trust me in this case, ram tweaking can make a huge difference and i can actually give u some ram settings that u would feel immediatly ( making the entire system more responsive immediatly noticeable )

[vanished]

Just now, Valkyrie Lenneth said:

the ram cas latencies are somthing that is overlooked by many ppl... im personally an extreme ram tweaker and those 10nanoseconds reduced in the ram to cpu speed can actually make like a input delay difference of like 5-10ms easily  ( the nanoseconds are basicly PER TASK, trust me there is millions  of tasks going on on that part ^^ )

 

the ram situation is extremely advanced to put in a short text and would require a post of atleast a whole page probably to explain ( way to much work for now )

 

but yes trust me in this case, ram tweaking can make a huge difference and i can actually give u some ram settings that u would feel immediatly ( making the entire system more responsive immediatly noticeable )

I would be extremely curious to see any test that proves that any adjustment to RAM timings can make 5 - 10 ms difference in overall system input lag

[Valkyrie Lenneth]

3 minutes ago, Ryan_Vickers said:

I would be extremely curious to see any test that proves that any adjustment to RAM timings can make 5 - 10 ms difference in overall system input lag

 

give me all of ur ram settings in my whisper including secondary timings, i will give u settings ( mostly secondary timings ) to give u an huge responsiveness increase

 

however i will refuse to give u any first timing settings because if those are unstable it can be data corrupting ur system ( unless u are willing to tweak + stress test ur ram for atleast 12hours of memtest89 / 24hours of CUSTOM prime95 filling all of the ram ) ect

 

 

 

also dont forget to give ur exact type number of ur ram sticks

 

[Valkyrie Lenneth]

4 minutes ago, Ryan_Vickers said:

I would be extremely curious to see any test that proves that any adjustment to RAM timings can make 5 - 10 ms difference in overall system input lag

the mistake that everyone makes about ram, is they think that it affects ram ( including linus made this mistake )

 

all tighter ram and quicker ram does is system responsiveness ( it does not add any performance in rendering almost, maybe takes of 1 or 2seconds off a 30sec file render )

 

short said

 

ram is all about system responsiveness, it will not really affect FPS unless the system responsiveness would become a bottleneck in certain particle heavy situations ( causing microstutters )  microstutters are usually due to a system responsiveness situation ( not in all cases ) and can be solved by tweaking the memory

[vanished]

4 minutes ago, Valkyrie Lenneth said:

 

give me all of ur ram settings in my whisper including secondary timings, i will give u settings ( mostly secondary timings ) to give u an huge responsiveness increase

 

however i will refuse to give u any first timing settings because if those are unstable it can be data corrupting ur system ( unless u are willing to tweak + stress test ur ram for atleast 12hours of memtest89 / 24hours of CUSTOM prime95 filling all of the ram ) ect

 

 

 

also dont forget to give ur exact type number of ur ram sticks

 

 

1 minute ago, Valkyrie Lenneth said:

the mistake that everyone makes about ram, is they think that it affects ram ( including linus made this mistake )

 

all tighter ram and quicker ram does is system responsiveness ( it does not add any performance in rendering almost, maybe takes of 1 or 2seconds off a 30sec file render )

 

short said

 

ram is all about system responsiveness, it will not really affect FPS unless the system responsiveness would become a bottleneck in certain particle heavy situations ( causing microstutters )  microstutters are usually due to a system responsiveness situation ( not in all cases ) and can be solved by tweaking the memory

 

At this point I'd rather just read about it before I go messing around myself.  But you know who would be very ready and willing to play with RAM timings to see what could happen?  @MageTank.  Sounds to me like if there's anyone who can verify or disprove that changing RAM timings can get you a 5 - 10 ms input lag reduction generally across all system functions, it will be you two.

[Valkyrie Lenneth]

3 minutes ago, Ryan_Vickers said:

 

 

At this point I'd rather just read about it before I go messing around myself.  But you know who would be very ready and willing to play with RAM timings to see what could happen?  @MageTank.  Sounds to me like if there's anyone who can verify or disprove that changing RAM timings can get you a 5 - 10 ms input lag reduction generally across all system functions, it will be you two.

i have checked hes profile he is using overclocked ram in mhz, i doubt he tweaked hes secondary timings for input lag, because ram cancels eachother out as u increase mhz ( it is better to keep it at stock speeds tweaking the latencies , the higher bandwidth wont do much for u unless u are running servers ( but the latencies will help a consumer )

 

however if u have a 3200 cas 14 stick and u clock it to 3800 keeping the cas 14 and u manage to keep it stable ( probably HUGE VOLT ) then the 3800 cas 14 would be way more responsive, but then again u would get more benefits from tweaking the timings on stock speeds ( because then u could reduce it by probably 2-3 cas if the sticks can do that

 

^ is just a judgement from the profile i checked ^^

 

[ivan134]

9 hours ago, JoostinOnline said:

Sorry, I didn't mean for it to be taken seriously.

 

But we've all got that one friend who considers winning a CS:GO match more important than being there for the birth of his first-born child.

Went from talking out of your ass since you don't play competitive games to stupid stereotypes. Never seen that before. 

[JoostinOnline]

10 hours ago, ivan134 said:

Went from talking out of your ass since you don't play competitive games to stupid stereotypes. Never seen that before. 

I used to.  I got sick of it because the (very vocal) minority ruined it for me.  I tend to do way worse at video games when I'm under pressure.  I'm sorry if I offended you.

 

If you have a 60hz monitor I definitely think V-Sync is worth it though.  As nice as minimal lag is, tearing is unbearable.  At least for me.

[xAcid9]

22 hours ago, Vincent123 said:

I have PUBG and I turned vsync on and it runs a lot smoother. Is it ok for me to have it on or should I shut it off. Tell me what you think. 

I enable Vsync almost on all games nowadays because like you mentioned, it's smoother when you're not using G-Sync/Free-Sync monitor and save power if your GPU able to push much higher FPS.

Try this RTSS trick to make certain games response better even with Vsync enabled. Credit to RealNC for bringing this up.

http://www.blurbusters.com/howto-low-lag-vsync-on/

[xAcid9]

21 hours ago, Valkyrie Lenneth said:

i have checked hes profile he is using overclocked ram in mhz, i doubt he tweaked hes secondary timings for input lag, because ram cancels eachother out as u increase mhz ( it is better to keep it at stock speeds tweaking the latencies , the higher bandwidth wont do much for u unless u are running servers ( but the latencies will help a consumer )

Nah, he tweaked his ram timing from primary to tertiary.

 

[MageTank]

23 hours ago, Valkyrie Lenneth said:

i have checked hes profile he is using overclocked ram in mhz, i doubt he tweaked hes secondary timings for input lag, because ram cancels eachother out as u increase mhz ( it is better to keep it at stock speeds tweaking the latencies , the higher bandwidth wont do much for u unless u are running servers ( but the latencies will help a consumer )

 

however if u have a 3200 cas 14 stick and u clock it to 3800 keeping the cas 14 and u manage to keep it stable ( probably HUGE VOLT ) then the 3800 cas 14 would be way more responsive, but then again u would get more benefits from tweaking the timings on stock speeds ( because then u could reduce it by probably 2-3 cas if the sticks can do that

 

^ is just a judgement from the profile i checked ^^

 

I have tweaked every timing imaginable. As for volts, 1.39vDIMM, 1.14 VCCIO, 1.15 VCCSA. Far from excessive in my eyes, but I'll gladly test any theories you may have regarding memory's impact on input lag. If you would like, I can reduce clock speeds to JEDEC and work strictly with timings, but I am certain I already know the outcome. 

 

I would like to remind you that CAS latency is not the be all, end all in regards to latency. On modern Intel platforms, that would be RTL or Round Trip Latency, and it's the culmination of everything (clock speeds, primaries, secondaries, tertiary, etc). Balance is key to achieving the best absolute RTL/IO-L values, and believe me when I say I have found the best balance for my kit, lol. 

 

As for input lag, it's possible for memory speed to have an impact due to the nature of how CPU's become I/O bound, and how memory speed impacts their performance in that scenario. The higher your minimum framerates in those situations, the less input lag you will have. As for the raw data, I would need to run tests when I get home. I am currently on vacation this week, so it will be some time before I can actually test this. 

[Valkyrie Lenneth]

32 minutes ago, MageTank said:

I have tweaked every timing imaginable. As for volts, 1.39vDIMM, 1.14 VCCIO, 1.15 VCCSA. Far from excessive in my eyes, but I'll gladly test any theories you may have regarding memory's impact on input lag. If you would like, I can reduce clock speeds to JEDEC and work strictly with timings, but I am certain I already know the outcome. 

 

I would like to remind you that CAS latency is not the be all, end all in regards to latency. On modern Intel platforms, that would be RTL or Round Trip Latency, and it's the culmination of everything (clock speeds, primaries, secondaries, tertiary, etc). Balance is key to achieving the best absolute RTL/IO-L values, and believe me when I say I have found the best balance for my kit, lol. 

 

As for input lag, it's possible for memory speed to have an impact due to the nature of how CPU's become I/O bound, and how memory speed impacts their performance in that scenario. The higher your minimum framerates in those situations, the less input lag you will have. As for the raw data, I would need to run tests when I get home. I am currently on vacation this week, so it will be some time before I can actually test this. 

from my testing timings such as refresh rate(1.5 to even 3x the stock),four act win timing(as low as minimum chipset allowance aka 16 for my particular system),   had the biggest impact on the input latency

 

yes it matters on all the timings put together in the end but these are the ones that have the biggest changes in the input latencies

 

since the cas latencies themself are within the ram alone and reduce general memory latency it does affect some input latency, but not quite as much as these 2 settings ( u already should know why ) in ur testing this is just a reference what to look for ( in other words do not forget about these )

 

also it was just an judgement on the profile, since i saw just the standard timings without any secondary or third timings added ( i guess its coz most ppl wouldnt understand anyway but yeah )     thx for clearing that one up

 

pst in reply to ur voltages... ive run up to 2.1v through ddr3 stickies just so u get the point of how extreme i am going

 

actually those huge increments in voltage are insane for ram tweaking... sadly cannot be used 24/7 though

 

as for >I can reduce clock speeds to JEDEC and work strictly with timings, but I am certain I already know the outcome. <

i can confirm it is not about the specific mhz that will make it better... it depends on mhz / cas ratio

 

example 1333mhz cas 6 kit would destroy a 1600mhz cas 9kit , in latencies ofc  ( depending on how high the frequencies are i estimated 1cas equals around 166-233mhz, and if u go at extreme high frequencies it is between 233-266mhz+- )

 

also why i mentioned it is better to keep the ram at stock mhz, because u usually can squeeze way tighter timings out of ur kit rather then u can mhz ( lets say at stock u can reduce ur cas by 3 total, but increasing mhz u would have to remain ur current cas... and the mhz increment isnt quite up to the cas/mhz ratio of that u would have with stock stickies

 

i am personally not interested in ddr4 due to there extreme high cas latencies ... about latency alone ddr3 still wins by quite a margin ( look at 2400mhz cas 9 for example ), that is because i prefer the lowest latencies possible rather then bandwidth ( because the bandwidth is quite useless for me :v )

 

[MageTank]

1 minute ago, Valkyrie Lenneth said:

from my testing timings such as refresh rate(1.5 to even 3x the stock),four act win timing(as low as minimum chipset allowance aka 16 for my particular system),   had the biggest impact on the input latency

 

yes it matters on all the timings put together in the end but these are the ones that have the biggest changes in the input latencies

 

since the cas latencies themself are within the ram alone and reduce general memory latency it does affect some input latency, but not quite as much as these 2 settings ( u already should know why ) in ur testing this is just a reference what to look for ( in other words do not forget about these )

 

also it was just an judgement on the profile, since i saw just the standard timings without any secondary or third timings added ( i guess its coz most ppl wouldnt understand anyway but yeah )     thx for clearing that one up

I do not buy your previous statements of 5-10ms difference in input delay coming from these timings alone. Ram does not work that way, this coming from a fellow "extreme ram tweaker", I could not imagine the difference being that dramatic. The timings in and of themselves do not dictate input latency. Altering them, you would mostly be altering framerates in regards to situations in which your CPU is I/O bound, which isn't all the time. You would have to not only force an I/O bound situation, but you would also need an accurate means to measure it. If you can provide even the slightest bit of evidence to your claims, I would be more inclined to believe you.

 

I am also curious to what random settings you told the other guy in private. There exists no magical numbers that alter input lag, as all IC's and memory controllers treat individual timings different. To suggest that you know universal timings that are not only stable, but will also yield a noticeable result when changed, sounds silly to me.

 

If you can provide your testing methodology, along with results for which I can compare my own to, then I will gladly conduct the test. If you cannot provide this, I am lead to believe you are seeing a placebo effect, and are attributing it to memory. Make no mistake, I am a firm believer in memory's impact on gaming performance, but I do not believe in misrepresenting it's value in that regard. 

[Valkyrie Lenneth]

17 minutes ago, MageTank said:

I do not buy your previous statements of 5-10ms difference in input delay coming from these timings alone. Ram does not work that way, this coming from a fellow "extreme ram tweaker", I could not imagine the difference being that dramatic. The timings in and of themselves do not dictate input latency. Altering them, you would mostly be altering framerates in regards to situations in which your CPU is I/O bound, which isn't all the time. You would have to not only force an I/O bound situation, but you would also need an accurate means to measure it. If you can provide even the slightest bit of evidence to your claims, I would be more inclined to believe you.

 

I am also curious to what random settings you told the other guy in private. There exists no magical numbers that alter input lag, as all IC's and memory controllers treat individual timings different. To suggest that you know universal timings that are not only stable, but will also yield a noticeable result when changed, sounds silly to me.

 

If you can provide your testing methodology, along with results for which I can compare my own to, then I will gladly conduct the test. If you cannot provide this, I am lead to believe you are seeing a placebo effect, and are attributing it to memory. Make no mistake, I am a firm believer in memory's impact on gaming performance, but I do not believe in misrepresenting it's value in that regard. 

To suggest that you know universal timings that are not only stable, but will also yield a noticeable result when changed, sounds silly to me. 

 

^ there are no universal timings that are stable , neither do i send any timings unless the user is willing to stability test ( i would simply assist in tweaking their timings ) , ive helped many ppl with ram tweaking however i havent helped anyone at all lately for quite a while ( most ppl just dont wanna put the effort into it hehe )

 

however trust me, u will notice that the refresh rate and tfaw(four act win) will have the most impact even in your testing ( input lantency alone )

 

i am planning to make an extended tutorial about input latencies / and how ram timings affect them in the future

 

keep in mind timings such as tfaw affect the time to the chipset hence why they affect input latencies more then other timings in general ( sadly tfaw has its limits on how tight it can go and that limitation is limited by the chipset, even if the ram can handle it the chipset will not be able to go tighter :c )

 

pst, extreme ram tweakers together usually make a great team(and end up in only agreements usually)

 

[MageTank]

4 minutes ago, Valkyrie Lenneth said:

To suggest that you know universal timings that are not only stable, but will also yield a noticeable result when changed, sounds silly to me. 

 

^ there are no universal timings that are stable , neither do i send any timings unless the user is willing to stability test ( i would simply assist in tweaking their timings )

 

however trust me, u will notice that the refresh rate and tfaw(four act win) will have the most impact even in your testing ( input lantency alone )

 

i am planning to make an extended tutorial about input latencies / and how ram timings affect them in the future

 

keep in mind timings such as tfaw affect the time to the chipset hence why they affect input latencies more then other timings in general ( sadly tfaw has its limits on how tight it can go and that limitation is limited by the chipset, even if the ram can handle it the chipset will not be able to go tighter :c )

 

 

Again, I am going to need some proof from you. It makes zero sense why tFAW helps input latency. tFAW is the window in which your ram can get 4 active strobes (rows), which is why it is typically 4x tRRD (ras to ras active delay). tFAW only matters when you are beating on the rows. This is not to say you won't see an improvement in memory latency benchmarks, because you can, but there exists no evidence or reason as to why it will aid in input performance. You keep telling me to test it, and that I will see, but you provide zero evidence (or even a reason) for me to test against. So I will ask you again.  Do you have ANY proof for this claim? 

[done12many2]

...don't mind me.  Just setting up my lawn chair.

[Liam-Wabbit]

There's many reason for micro stutters, bad drivers being one.
I had to find a old 2012 driver for my 440 GTX graphic card for Lara Croft, guardian of light game. this had really bad freezing problem on the card. Euro Truck Simulator 2 I had to go into Task manager and disable some of the CPU cores in affinity settings off the game,
where as most games that run unreal engine seem to be fine. 
The only thing you can do is monitor your system when it happens to see if there any cause. 

Often there a CPU spike and then a big frame drop. some games do it. 

some games work fine and it's not linked to any hardware brand ! Personally I think it down to drivers