• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

ShadowPlay GPU usage

Gregster

Gregster

Gregster

Caporegime
Joined
24 Sep 2008
Posts
38,284
Location
Essex innit!
So since the 331.65 drivers, I have used ShadowPlay or at least had it on in the background and found it a useful tool. The thing I have noticed though is BF4 bugs. The game never felt smooth (unlike the beta) and in fact was the worst game I had ever played for smoothness. I had read several people were having smoothness issues, so I paid it no mind and put it down to the game being broken/poorly optomised/yada yada yada....

So here is the thing. I was recording my fps with fraps and they seemed questionably low for what I was getting in the beta and luckily I had a couple of bench runs saved from back then and comparing was about a 10% difference. So I ran the fraps bench on Siege of Shaghai 32 man server without ShadowPlay being on.

Here is the results:

With ShadowPlay ready to record but not recording.
Frames, Time (ms), Min, Max, Avg
68375, 803311, 53, 150, 85.116

With ShadowPlay off from the Geforce Experience
Frames, Time (ms), Min, Max, Avg
42800, 457239, 56, 179, 93.605

As you can see, ShadowPlay just running in the background and not being used is giving a 9.4% hit to performance, which is quite a bit but I would accept that but the stutter experienced in BF4 was quite bad (quite a game killer).

I will do some more tests today to see the average impact on performance and if it brings that stutter to other games.
 
I haven't noticed this.
It is very hard to bench the difference on a multiplayer game where anything can happen.
Try running the tomb raider benchmark with it off, then with it on & see the difference. That benchmark is controlled so should give a clearer indication of any performance hit.

From my experience, I have not noticed any hit in performance, but I haven't tried in a benchmark scenario, only looking at fraps & it showing the same fps I am used to
 
Ran some tests with a couple of games.

Tomb Raider with ShadowPlay ready to record (but not recording)
31 minimum 47 average 67 max
and without ShadowPlay
46 minimum 58 average 77 max
23% performance hit on the average

Batman Arkham City with ShadowPlay ready (but not recording)
35 minimum 75 average 126 max
And without Shadowplay
44 mimimum 100 average 180 max
33% performance hit on the average

It seems there is a fair hit on performance and none of the game bench runs were smooth with it there.
 
I'll check out TR later and see if I can replicate the results, I've had SP on manual mode since it came with the driver update but I haven't noticed any difference but who knows.

I'll post back when I'm on my comp :)
 
No such thing as "Ready and not recording" (If you have Shadow Time Enabled)

ShadowTime recording is EXACTLY the same as Manual recording. Except ShadowTime only keeps the last X minutes of gameplay...it's STILL recording.

Either Shadowplay is recording or it's not - There is no in-between.
 
Nice one. I did ask in the BF4 thread if someone wouldn't mind confirming my findings, just to see if it was me or a normal thing.

It is a non issue, as it takes seconds to turn off Shadowplay but it would be great to have it there and record when needed. BF4 will bring some funny moments that would be great to record and add them all together for a nice montage :)

If anyone else would like to have a test of this as well, it would be highly appreciated :)

Sin_Chase said:
No such thing as "Ready and not recording" (If you have Shadow Time Enabled)

ShadowTime recording is EXACTLY the same as Manual recording. Except ShadowTime only keeps the last X minutes of gameplay...it's STILL recording.

Either Shadowplay is recording or it's not - There is no in-between.
Click to expand...

I get that but I wasn't expecting that kind of hit. 33% is a big chunk and even with high fps games, they feel stuttery. The conference said around a 2% hit on performance and clearly what I am getting is slightly more....
 
Last edited:
Depends on your rig.

I see no difference with it being enabled, certainly not 33%. I record to a high performance RAID-0 Stripe however.

BF4 Beta with Shadowplay was a stutery mess, retail is perfect however.
 
Sin_Chase said:
Depends on your rig.

I see no difference with it being enabled, certainly not 33%. I record to a high performance RAID-0 Stripe however.

BF4 Beta with Shadowplay was a stutery mess, retail is perfect however.
Click to expand...

I record to an SSD.

BF4 Beta didn't have the option to use ShadowPlay, as it wasn't available till the 28th (I got a leaked driver early but still wasn't before the Beta had finished).
 
Sin_Chase said:
Ditto.

Strange, not done a scientific test (Will do once my 780Ti arrives) but I notice zero difference recording or not on my 680.
Click to expand...

I put it down to being BF4 and the many reported issues that users were having and didn't think much of it. I am very picky when it comes to smoothness and notice the silliest of things but don't fuss over them as I accept them. It was only the benching that got me to try it with and without and this is where I noticed smoothness differences.

The more people that try this and let the community know, we could possibly come up with an answer or Nvidia may well do. I fully understand that Beta can have bugs/faults, so I am certainly not faulting this excellent program and look forward to using it.

LtMatt said:
Good work Gregos. You can't beat user documented findings. This reminded me of my AMD ULPS thread.
Click to expand...

Maybe just me being picky Matt, you know what I was like on the Tomb Raider bench with smoothness and I can be anal at times (leave it). Hopefully others will confirm the same as me, as I would hate to be the only one and then being driven to insanity trying to find a fix :D
 
Gregster said:
Maybe just me being picky Matt, you know what I was like on the Tomb Raider bench with smoothness and I can be anal at times (leave it). Hopefully others will confirm the same as me, as I would hate to be the only one and then being driven to insanity trying to find a fix :D
Click to expand...

Nothing wrong with a bit of anal lad. When the rivers run red, take the dirt track instead. ;)
 
I will try this at about 9ish tonight and see if I can replicate that because so far my experience has been basically a 0 performance hit.
 
Well I would test, but Fraps + Tomb Raider don't seem to play nice, with Fraps running the game launches with just a black screen and if I run it after launching the game it can crash. Anyone know a fix?
 
Deceptor said:
Well I would test, but Fraps + Tomb Raider don't seem to play nice, with Fraps running the game launches with just a black screen and if I run it after launching the game it can crash. Anyone know a fix?
Click to expand...

Just run the inbuilt bench :D No need for Fraps.
 
You must log in or register to reply here.
Back
Top Bottom