Blur Buster's G-SYNC 101 Series Discussion

Talk about NVIDIA G-SYNC, a variable refresh rate (VRR) technology. G-SYNC eliminates stutters, tearing, and reduces input lag. List of G-SYNC Monitors.
User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by Chief Blur Buster » 16 Aug 2017, 14:10

Indeed, it's complicated stuff.

Whenever "VSYNC OFF" is mentioned without GSYNC, it's simply ordinary "VSYNC OFF" -- the kind you use on any monitor. That's what we did for our 1000fps and 2000fps+ lag tests.

Takeaways for ordinary "VSYNC OFF" in CS:GO gaming were:
-- Increasing refresh rate lowers lag at the same frame rate
-- Increasing refresh rate lowers lag variability (tighter min/avg/max)
-- Increasing frame rate lowers lag at the same refresh rate
-- Increasing frame rate lowers lag variability (tighter min/avg/max)
-- Doing all 4 of the above is the best-of-all-worlds for eSports players. Higher Hz and figher fps

Lag variability can be important to CS:GO players. Random lag variance means less consistent aiming, more shooting/misses before hitting tiny far away targets, etc. Based on this specific test, I'm not surprised (once you hit the frame rate stratosphere), why eSports players usually prefer VSYNC OFF over GSYNC/FreeSync.

If you only play CS:GO and have a very old GTX 660, it's still worthwhile upgrading to a GTX 1080i or GTX Titan or Radeon RX Vega if you're playing for eSports championship money -- even if you only play CS:GO and already get "only" 300fps on your brand new 240Hz monitor.

That said, if you're playing on modern game engines that can't get frame rates exceeding refresh rates, I prefer playing frame-capped GSYNC or FreeSync. Even a 240Hz GSYNC monitor also reduces lag of low frame rates. ~60fps@240Hz has lower average lag than ~60fps@60Hz. (That doesn't even include the reduced lag variability).

And now with the ultra-low-lag VSYNC ON techniques, even VSYNC ON is a contender again for lag-critical applications (in certain cases) -- that became important again, thanks to VR, and has spinoff applications to fixed-Hz desktop gaming.

I say! Right tool for the right job;
The right mode (ULMB? VSYNC ON? VSYNC OFF? GSYNC? Etc) for the right game.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
RealNC
Site Admin
Posts: 3741
Joined: 24 Dec 2013, 18:32
Contact:

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by RealNC » 16 Aug 2017, 14:34

It should be noted here that a very fast GPU is going to be idle in CS:GO. If you want more FPS, you need the CPU with the highest IPC (highest single-core peformance) you can get. CPU overclocking is what gets you FPS in CS:GO. Upgrading the GPU might give you nothing, actually, unless you want to play at high resolutions (1440p or 4K) and/or with graphics settings maxed out. However, most pros use low details to reduce the "details noise" on the screen and make enemy movement easier to spot.

I estimate that something like a GTX 1060 is all you need for 500FPS on CS:GO if you play on 1080p. People who can't get 500FPS are almost always bottlenecked by their CPU.

Currently, the highest FPS numbers can only be achieved with Intel i7 CPUs, with hyperthreading DISABLED, and overclocked to the maximum speed possible.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by Chief Blur Buster » 16 Aug 2017, 15:09

And CS:GO also in single threaded mode, for better lag consistency (min/avg/max). Multicore mode can be great for many things, but, CS:GO...it's begging for good single thread performance to get the best data.
[Edit: That's for lag consistency if you're running at lower frame rates; apologies]

For stratosphere frame rates, you do need CS:GO multicore, though! Stratospheric frame rates do indeed compensate.

We've got a good surge of eSports readers now (480Hz went viral in Google News this week) so we've gotta increasingly watch what we say when "2ms lag doesn't matter" or "500fps is enough" -- any extra millisecond of advantage clearly matters for a rapidly increasing pie-share of our readers mind you, and various lag advantages do shows in our tests.

-- One minor note: While many tweaks are true, overclocking is no longer needed anymore to frequently go into 1000fps territory in CS:GO. The Intel i7-7740X can easily give the oomph at its default until there's many players on the server --

Bring on the 256 tick servers!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
RealNC
Site Admin
Posts: 3741
Joined: 24 Dec 2013, 18:32
Contact:

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by RealNC » 16 Aug 2017, 15:34

Chief Blur Buster wrote:And CS:GO also in single threaded mode, for better lag consistency (min/avg/max). Multicore mode can be great for many things, but, CS:GO...it's begging for good single thread performance to get the best data.
This will get you stuck to under 200FPS though, mind you. There is currently no CPU on this planet that can reach 500FPS without multicore rendering enabled. Even 240FPS is probably not possible, let alone 480FPS.

When you see "500FPS" or similar numbers in CS:GO, it is implied that multicore rendering is enabled.

Edit: Btw, an empty server does not count. Starting a local game with no bots is not indicative of what FPS you'll get if you actually start a 5v5 online match.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by Chief Blur Buster » 16 Aug 2017, 16:25

Thanks, updated my post with the caveat!
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

casinca
Posts: 9
Joined: 01 Sep 2014, 10:02

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by casinca » 16 Aug 2017, 17:26

Thanks guys for the clarification ;) , these tests are indeed so valuable for gaming optimization; that's all i was looking for :)

It may seems insignificant in term of ms but here and there, if you're accumulating ms of input lag from various sources, at then end, when you and your opponent both click at the same time, this is where it can make a difference.

User avatar
jorimt
Posts: 2481
Joined: 04 Nov 2016, 10:44
Location: USA

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by jorimt » 16 Aug 2017, 19:01

Not a problem @casinca,

And yes, if you're looking for the fastest frame delivery with the least amount of input latency, V-SYNC OFF + high refresh rate + high framerates is still the way to go.
(jorimt: /jor-uhm-tee/)
Author: Blur Busters "G-SYNC 101" Series

Displays: ASUS PG27AQN, LG 48CX VR: Beyond, Quest 3, Reverb G2, Index OS: Windows 11 Pro Case: Fractal Design Torrent PSU: Seasonic PRIME TX-1000 MB: ASUS Z790 Hero CPU: Intel i9-13900k w/Noctua NH-U12A GPU: GIGABYTE RTX 4090 GAMING OC RAM: 32GB G.SKILL Trident Z5 DDR5 6400MHz CL32 SSDs: 2TB WD_BLACK SN850 (OS), 4TB WD_BLACK SN850X (Games) Keyboards: Wooting 60HE, Logitech G915 TKL Mice: Razer Viper Mini SE, Razer Viper 8kHz Sound: Creative Sound Blaster Katana V2 (speakers/amp/DAC), AFUL Performer 8 (IEMs)

Vleeswolf
Posts: 37
Joined: 25 Aug 2017, 15:59

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by Vleeswolf » 25 Aug 2017, 16:03

Hi jorimt, thanks for the great series and discussion. Did you ever measure the input lag of VSYNC (NVCP) ON + GSYNC ON, NV driver frame limiter at upper range - 3 FPS, with Max Prerendered Frames set to 1? Does that setup still give more input lag than VSYNC (NVCP) ON + GSYNC ON, RTSS frame limit at upper range - 3 FPS?

User avatar
jorimt
Posts: 2481
Joined: 04 Nov 2016, 10:44
Location: USA

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by jorimt » 25 Aug 2017, 23:47

@Vleeswolf

Disclaimer: MPRF is a deceivingly complex setting with variable effects, and even I'm still wrapping my head around all of its intricacies (a reason why I didn't bother digging into it in my predominately sync-themed article), so take my analysis with a grain of salt, but as far as I understand it now...

MPRF only has an effect below the framerate limiter; if your framerate is sustained above the set framerate limit, then MPRF isn't doing anything.

As long as the framerate can be sustained above the set framerate limit, the framerate limiter is letting the engine know that it wants it to render frames in a certain amount of time, every time, which ultimately reduces input latency, because the engine can accurately predict and output the next frame with the specified frametime.

With "free range" framerates however, the engine has to calculate render time on the fly, and to prevent frametime spikes (e.g. the system not outputting anything for one or multiple refresh cycles at a time), it pre-renders a few frames ahead of time to compensate for whatever framerate the system is able/not able to output at any given time, delivering one or more of those pre-rendered frames when it becomes necessary for the system to catch up and render new frames; MPRF controls how many frames are pre-rendered in this process.

The general rule is that there is more input lag at say, 135 FPS @144Hz with no FPS limit compared to 135 FPS @144Hz with an FPS limit, so if you're sustaining 141 FPS @144Hz 95% of the time, but you drop below it 5% of the time, that's when the MPRF setting would kick in.

As I stated in my article, setting it to "1" may improve things if your system can take it, or it may put more strain on the CPU and cause more frametime spikes. That, and lower MPRF values may have no effect at all, as some games already have an equivalent value of "1" set internally.

Thus, if you have a framerate limit set and the framerate is always sustained above it, MPRF doesn't matter, if not, it depends, and you'll just have experiment per game.

As for the Nvidia limiter vs RTSS, I haven't been able to make the former match or better the latter (input latency-wise) in any configuration thus far.
(jorimt: /jor-uhm-tee/)
Author: Blur Busters "G-SYNC 101" Series

Displays: ASUS PG27AQN, LG 48CX VR: Beyond, Quest 3, Reverb G2, Index OS: Windows 11 Pro Case: Fractal Design Torrent PSU: Seasonic PRIME TX-1000 MB: ASUS Z790 Hero CPU: Intel i9-13900k w/Noctua NH-U12A GPU: GIGABYTE RTX 4090 GAMING OC RAM: 32GB G.SKILL Trident Z5 DDR5 6400MHz CL32 SSDs: 2TB WD_BLACK SN850 (OS), 4TB WD_BLACK SN850X (Games) Keyboards: Wooting 60HE, Logitech G915 TKL Mice: Razer Viper Mini SE, Razer Viper 8kHz Sound: Creative Sound Blaster Katana V2 (speakers/amp/DAC), AFUL Performer 8 (IEMs)

Vleeswolf
Posts: 37
Joined: 25 Aug 2017, 15:59

Re: Blur Buster's G-SYNC 101 Series Discussion

Post by Vleeswolf » 26 Aug 2017, 02:21

I always understood MPRF as a queue of frames to be displayed, which is added to if the game or driver hits and tries to exceed some speed limit, in particular VSYNC. The queued up frames can be used to compensate for frame time spikes. However, frames in the queue are displayed in order, hence lag increases as the queue grows in length. If you run at half-rate vsync, in particular, the effect of the MPRF setting is pretty noticeable on some games, with 1 giving the lowest lag.

My hunch was that the NV driver frame limit would also allow the pre-render queue to grow, making the MPRF setting important.

Post Reply