ShadowPlay - GVR - Mirillis Action - Fraps - Show your performance hit

Gregster · 5 Aug 2014 at 13:59

pgi947 said:
Ok so after scouring forums, it turns out shadowplay and sli has a bug causing the performance hit, comment from nvidia as recently as the end of July says it will be fixed 'soon' in an upcoming driver release. Though the bug has been known since November 2013 so I wouldn't hold your breath as it doesn't seem high up on nvidias list of priorities.

Good stuff. At least they know about it and it isn't a big issue. I ran it at 1600P to see if it helped with maybe the resolution being a bottleneck but didn't. Curious to know if any of the other recording software apps do the same.

LtMatt · 5 Aug 2014 at 14:08

Did some testing on Tomb Raider at 1080P highest possible settings, 290 crossfire.

Mirillis Action
MIN:36.9
AVG:72.1
MAX:106

GameDVR
MIN:50
AVG:69
MAX:106

Without Recording
MIN:58
AVG:84
MAX:104

21% performance penalty using crossfire in DVR and 16% using Mirillis Action. Interesting results.

Gregster · 5 Aug 2014 at 14:20

That was Crossfire? Seems very low scores Matt.

LtMatt · 5 Aug 2014 at 14:43

That was using x4 SSAA and Ultra Shadows.

pgi947 · 5 Aug 2014 at 15:09

So it looks as if multicard recording From either camp has its issues with performance hit, I'll try fraps and dxtory if I get chance as iirc you can set either of those to render from the cpu rather then gpu, see if that hits multicard too.

LtMatt · 5 Aug 2014 at 15:13

pgi947 said:
So it looks as if multicard recording From either camp has its issues with performance hit, I'll try fraps and dxtory if I get chance as iirc you can set either of those to render from the cpu rather then gpu, see if that hits multicard too.

Yep seems that way tips. I'm quite disappointed by that. I will relay some info on the Raptr forums about it and see what's what.

Gregster · 5 Aug 2014 at 15:43

LtMatt said:
That was using x4 SSAA and Ultra Shadows.

Ahhh fair. I was on 2xSSAA and normal shadows. Switching to 4XSSAA and ultra Shadows gave me 75 fps average when recording.

LtMatt · 5 Aug 2014 at 15:47

LtMatt said:
Did some testing on Tomb Raider at 1080P highest possible settings, 290 crossfire.

Mirillis Action
MIN:36.9
AVG:72.1
MAX:106

GameDVR
MIN:50
AVG:69
MAX:106

Without Recording
MIN:58
AVG:84
MAX:104

21% performance penalty using crossfire in DVR and 16% using Mirillis Action. Interesting results.

GameDVR

Mirillis AVI

No recording

I had real problems getting Mirillis Action to work with Tomb Raider. So I'm not totally convinced with the results from using that. I don't have any other games installed to test at the moment though. There was also some noticeable stutter while using Mirillis, which can be seen in the vid. There was none using DVR, so that's a plus point. Typically Mirillis performs very well in other games though.

Gregster · 5 Aug 2014 at 16:14

They are both very bad for stutter in truth.

LtMatt · 5 Aug 2014 at 16:27

Gregster said:
They are both very bad for stutter in truth.

Heh, i just watched ours side by side and they looked identical but mine didn't have the flickering hair.

Gregster · 5 Aug 2014 at 16:50

LtMatt said:
Heh, i just watched ours side by side and they looked identical but mine didn't have the flickering hair.

Yer, that's what I meant, both of ours have stutter. The hair is a strange one and does that in the bench more than it doesn't but doesn't do it in the game. I am not knocking YouTube as it is free but when you watch the raw file, it has none of the stutter and the colour is far more vibrant.

LtMatt · 5 Aug 2014 at 16:55

Gregster said:
Yer, that's what I meant, both of ours have stutter. The hair is a strange one and does that in the bench more than it doesn't but doesn't do it in the game. I am not knocking YouTube as it is free but when you watch the raw file, it has none of the stutter and the colour is far more vibrant.

Yer i know.

I think it's just 30fps tbh. 60fps will always look a lot smoother.

That hair flickering happens to me sometimes too, just got lucky those times lol.

tommybhoy · 5 Aug 2014 at 17:16

LtMatt said:
That hair flickering happens to me sometimes too

We don't need reminding.

shankly1985 · 5 Aug 2014 at 17:20

Tomb Raider 1080p - Full settings with 2xssaa
Single 290 @977/1250
i7 2600k @stock
1600mhz 16GB Ram
Am not sure why my Max was higher when was recording, something makes me think RadeonPro. RP crashed when I tried to record.

DVR Off - Min 40.7 - Max - 66.0 - AVG - 53.3

TombRaider.exe_2014-08-03-02-00-19-860.png

DVR On - Min - 40.0 - Max - 70.0 - AVG - 54.9

Tip!
If you have trouble recording to another hard drive, make sure the folder you writing to has full write control for your windows user account.
---------------------------------------------------------------------------------------------------------------------------------------------------------------

Thief 1080p - Full settings
Single 290 @977/1250
i7 2600k @stock
1600mhz 16GB Ram

DVR - Off - Min 45.9 - Max 76.9 - AVG 60.2

Shipping-ThiefGame.exe_2014-08-03-01-48-29-033.png

DVR - On - Min 42.1 - Max 72.1 - AVG 56.1

---------------------------------------------------------------------------------------------------------------------------------------------------------------
Batman - 1080p - Full settings 4xmsaa
Single 290 @977/1250
i7 2600k @stock
1600mhz 16GB Ram

6% difference?

DVR off - Min 78 - Max 165 - AVG 120

BatmanOrigins.exe_2014-08-03-17-08-47-771.png

DVR On - Min 76 - Max 156 - AVG 113

BatmanOrigins.exe_2014-08-03-17-13-24-388.png

---------------------------------------------------------------------------------------------------------------------------------------------------------------
Batman - 1080p - Full settings 4xmsaa
Single 290 @977/1250
i7 2600k @stock
1600mhz 16GB Ram

DVR off - Min 23 - Max 178 - AVG 122

BatmanAC.exe_2014-08-03-12-35-39-158.png

DVR on - Min 13 - Max 167 - AVG 115

BatmanAC.exe_2014-08-03-12-33-07-425.png

On this Bench the min is very random one, when the bench changes scene the min FPS really takes a hit.
---------------------------------------------------------------------------------------------------------------------------------------------------------------
AMD VCE on Dxtory

Tomb Raider max settings with 2xssaa

Dxtory stock encoder with quality low

Dxtory with AMD VCE

Rusty0611 · 5 Aug 2014 at 17:35

GoogalyMoogaly said:
Surely it'd be:
((x / y) - 1) * -100

Not sure you can get more accurate than that if you're calculating a perfomance hit (x = DVR fps, y = normal fps).

Firstly you don't need brackets around x and y. Secondly that isn't the correct way to do it using that method. You'd do the bigger number divided by smaller to work out the proportion between the two.

The 'correct' method is (using your terminology of x and y):

(y - x) / ((y + x)/2) * 100

pgi947 · 5 Aug 2014 at 18:06

TIL; Maths

LtMatt · 5 Aug 2014 at 18:11

Rusty0611 said:
( x / y -1) * 100

Or you can use the more accurate method detailed in the other thread.

GoogalyMoogaly said:
Surely it'd be:
((x / y) - 1) * -100

Not sure you can get more accurate than that if you're calculating a perfomance hit (x = DVR fps, y = normal fps).

Rusty0611 said:
Firstly you don't need brackets around x and y. Secondly that isn't the correct way to do it using that method. You'd do the bigger number divided by smaller to work out the proportion between the two.

The 'correct' method is (using your terminology of x and y):

(y - x) / ((y + x)/2) * 100

Gregster · 5 Aug 2014 at 18:18

An apt Gif

Rusty0611 · 5 Aug 2014 at 18:30

LOL Matt.

gerardfraser · 5 Aug 2014 at 19:08

Thought I would post this here for compare.

Did some Crossfire testing with AMD GVR.More performance hit with Crossfire than single cards,but still not too bad.

Breakdown : percentage hit on average values.
Tomb Raider - 21.6%
Thief - 12.4%
Sniper Elite V2 - 12.5%
Hitman Absolution - 9.7%
Batman Arkham Origins -6.9%
Sleeping Dogs - 11.2%
Bioshock Infinite - 18.1%