The Skyrim Special Edition Graphical Setting Performance Testing Spreadsheet --- Or --- How to get those dank Eff-Pee-Esses

	A	B	C	D	E	F	G	H	L	M	N	O	P
1		Hello! Welcome to my spreadsheet. I hope the information contained herein helps you optimize your own Skyrim client, and gain some of those sweet, sweet FPS. Please note that pictures and GIFs have been linked to the "Client State" column entries, where applicable. They are denoted by underlined text.
2
3
4	ABSOLUTE FPS VALUE TABLE This table lists the raw FPS numbers recorded in each test, along with the FPS gain/loss over the AVERAGE.								FPS CHANGE OVER BASELINE TABLE This tables lists the FPS gain/loss over the BASELINE value for each test individually.
5	Client State	Looking Down Towards Ralof	Looking Straight Down	Looking at the Horizon	Looking at the Cave Rock	AVERAGE FPS	FPS CHANGE OVER BASELINE		Client State	Looking Down Towards Ralof	Looking Straight Down	Looking at the Horizon	Looking at the Cave Rock
6
7	BASELINE Setup (All subsequent FPS measurements will be compared to THESE values)								BASELINE Setup (All subsequent FPS measurements will be compared to THESE values)
8	Ultra Everything, TOAA, Shadows at 4096 (Ultra) (THIS IS BASELINE. Shadows will always be at 4096 and AA will always be TOAA unless stated otherwise.)	60	82	69	78	72.25			Ultra Everything, TOAA, Shadows at 4096 (Ultra) (THIS IS BASELINE. Shadows will always be at 4096 and AA will always be TOAA unless stated otherwise.)	60	82	69	78
9
10	Non-Baseline Setups								Non-Baseline Setups
11	Client at Lowest Possible Settings, All Graphic Options Turned off or to Low	120	120	120	120	120	47.75		Client at Lowest Possible Settings, All Graphic Options Turned off or to Low	60	38	51	42
12	Client at "Optimal" Settings by GeForce Gaming Experience Optimizer	68	95	78	84	81.25	9		Client at "Optimal" Settings by GeForce Gaming Experience Optimizer	8	13	9	6
13	Client with Both GeForce and Steam Overlays Running	60	82	69	78	72.25	NO EFFECT	This is another thing which a lot of redditors swear impacts performance -- so much so that it's actually included in the subreddit's official modding guides --, but I've personally never noticed an impact on performance over many many hours of play. I suspect it will only affect systems whose CPU's are already near 100% load with the game.	Client with Both GeForce and Steam Overlays Running	0	0	0	0
14
15	Modifications on the BASELINE (Each line entry below involves starting with the BASELINE setup, and then changing a single property, as stated in the "Client State" column. Changes do NOT roll over from one column to the next. Each line entry tests only a SINGLE variable.)								Modifications on the BASELINE (Each line entry below involves starting with the BASELINE setup, and then changing a single property, as stated in the "Client State" column. Changes do NOT roll over from one column to the next. Each line entry tests only a SINGLE variable.)
16	Shadows at 2046 (High)	65	92	74	82	78.25	6	Reducing the shadow resolution from 4092 to 2046 or 1024 is definitely noticeable if you stop and stare at a shadow. Given the FPS gains that can be had, though, it's still worth reducing.	Shadows at 2046 (High)	5	10	5	4
17	Shadows at 1024 (Low)	67	95	76	84	80.5	8.25		Shadows at 1024 (Low)	7	13	7	6
18	Without Any Anti-Aliasing	64	88	74	82	77	4.75		Without Any Anti-Aliasing	4	6	5	4
19	With FXAA instead of TOAA	62	85	72	80	74.75	2.5		With FXAA instead of TOAA	2	3	3	2
20	Without Screen-Space Ambient Occlusion	60	82	67	78	71.75	NO EFFECT		Without Screen-Space Ambient Occlusion	0	0	-2	0
21	Without Snow Shader	64	92	71	84	77.75	5.5	The Snow Shader should be off if you use Rudy ENB. Without an ENB, though, it does make a noticeable visual difference.	Without Snow Shader	4	10	2	6
22	Without Screen-Space Reflections	69	92	77	90	82	9.75		Without Screen-Space Reflections	9	10	8	12
23	Without Lens Flare	63	84	72	80	74.75	2.5	Lens Flare is one of those stupid effects where you lose FPS when its on, even if there's no actual lens flare being rendered! EVEN IF THERE'S NO SUN!	Without Lens Flare	3	2	3	2
24	Without 64-Bit Rendering Targets	64	88	73	81	76.5	4.25	I can personally not see any visual change when toggling the 64-bit render targets on and off, and so recommend turning it off for its sizeable FPS gain. Your mileage may vary.	Without 64-Bit Rendering Targets	4	6	4	3
25	Without Precipitation Occlusion	60	82	69	78	72.25	NO EFFECT		Without Precipitation Occlusion	0	0	0	0
26	Without God Rays (Off)	69	97	80	90	84	11.75	God Rays do actually make a huge visual difference, but only when turning them on or COMPLETELY off. By setting them to "low" quality, you can retain 95% of the visual effect, while still gaining some frames.	Without God Rays (Off)	9	15	11	12
27	With Shadow Distance at Lowest (Med)	60	82	69	78	72.25	NO EFFECT		With Shadow Distance at Lowest (Med)	0	0	0	0
28	With Distant Object Detail at Lowest (Low)	62	83	71	79	73.75	NO EFFECT		With Distant Object Detail at Lowest (Low)	2	1	2	1
29	Baseline Tested in Windowed Mode and Borderless Windowed Mode	60	82	69	78	72.25	NO EFFECT		Baseline Tested in Windowed Mode and Borderless Windowed Mode	0	0	0	0
30	With Skyrim 2020 Parallax Textures by Pfuscher Enabled, but NO ENB files	49	75	57	72	63.25	-9	The loss of FPS from high-res textures confused some people, so it's worth stating that I was NOT running out of VRAM (only 4.65GB out of 8 were in use) or system RAM. High-res parallax textures simply require more computational processing.	With Skyrim 2020 Parallax Textures by Pfuscher Enabled, but NO ENB files	-11	-7	-12	-6
31	With Noble Skyrim 2K Textures Enabled	60	80	69	77	71.5	NO EFFECT		With Noble Skyrim 2K Textures Enabled	0	-2	0	-1
32	With DynDOLOD on Medium Settings, with 256 Textures	59	80	66	75	70	-2.25		With DynDOLOD on Medium Settings, with 256 Textures	-1	-2	-3	-3
33	With DynDOLOD on Low Settings, with 128 Textures	60	80	69	76	71.25	NO EFFECT		With DynDOLOD on Low Settings, with 128 Textures	0	-2	0	-2
34	With Video Driver Thread Optimization Disabled	60	82	69	78	72.25	NO EFFECT		With Video Driver Thread Optimization Disabled	0	0	0	0
35	Baseline With ONLY ENB Binaries Installed	53	72	60	70	63.75	-8.5	This is widely-known, but still worth stating: simply having the ENB binaries installed will kill nearly 10 FPS, even if you aren't using a preset, and aren't benefiting from any visual improvements.	Baseline With ONLY ENB Binaries Installed	-7	-10	-9	-8
36	With ENB Binaries and Rudy ENB Preset, with ALL Preset Effects Disabled	60	80	70	76	71.5	NO EFFECT		With ENB Binaries and Rudy ENB Preset, with ALL Preset Effects Disabled	0	-2	1	-2
37	With ENB Binaries and Rudy ENB on Full	29	32	34	33	32	-40.25	Holy fuck, Rudy is expensive...	With ENB Binaries and Rudy ENB on Full	-31	-50	-35	-45
38	With ENB Binaries and Rudy on Full, but with Game-Space AO, Reflections, Lens Flare, and 64-bit targets DISABLED	34	37	41	40	38	-34.25		With ENB Binaries and Rudy on Full, but with Game-Space AO, Reflections, Lens Flare, and 64-bit targets DISABLED	-26	-45	-28	-38
39	With ENB and Rudy on Full, but all Other Game Settings on Lowest Possible	40	41	48	43	43	-29.25		With ENB and Rudy on Full, but all Other Game Settings on Lowest Possible	-20	-41	-21	-35
40	With ENB and Rudy on Full, and Skyrim 2020 Parallax Textures by Pfuscher	25	30	30	32	29.25	-43		With ENB and Rudy on Full, and Skyrim 2020 Parallax Textures by Pfuscher	-35	-52	-39	-46
41	Baseline, but set Object Fade to 15 (Ultra=30)	60	82	69	78	72.25	NO EFFECT			0	0	0	0
42
43	Multivariable Testing on the BASELINE with BethINI (Click HERE for a 1-Second Overall Comparison GIF) (Click HERE for a 3-Second Overall Comparison GIF)
44	BethINI Ultra	62	93	69	85	77.25	5	Alright so, here's where we need to have a little chat: Up to this point, all testing has been single-variable testing. That is to say: I toggle one graphical feature at a time, and measure its impact on performance. Now though, in this section, BethINI is making changes to dozens of variables simultaneously..... and yet, the results are not as great as you'd think. If you were to manually add up the FPS gains from each change made individually, as reported in the section prior, you'd expect to see much larger gains, but in reality, you get diminishing returns. This is the nature of graphics processing: The product is LESS than the sum of its parts. If you have five graphical features which individually cost 5 FPS, you would think that turning them all off would net you 5x5 = 25 FPS, but in reality, you only gain like 16.5 FPS, or something. It's just the unfortunate reality.
45	BethINI High	74	109	82	98	90.75	18.5
46	BethINI High, Terrain Manager Max Distance at 300K	73	108	81	97	89.75	17.5
47	BethINI Medium	86	120	95	106	101.75	29.5
48	BethINI Low	120	120	120	120	120	47.75
49
50	Rudy ENB Preset Options Testing (Each line entry below involves starting with a "Rudy BASELINE", which is just the original BASELINE configuration, with the addition of the ENB Binaries, and the Rudy ENB Preset, along with the diabling of Skyrim's Internal Ambient Occlusion (as per Rudy's installation instructions.) All FPS gains/losses in this section are referenced against the Rudy BASELINE, not the original BASELINE. (Click HERE for a 1-Second Overall Comparison GIF) (Click HERE for a 3-Second Overall Comparison GIF) ---- (Click HERE for a 1-Second GIF comparing Rudy's AO to the Game's built-in AO) (Click HERE for a 3-Second GIF comparing Rudy's AO to the Game's built-in AO)
51	BASELINE Skyrim, With Rudy ENB on Full, No Game-AO (This is the Rudy BASELINE)	29	32	34	33	32
52	Rudy ENB, No AO of Any Kind	37	43	41	45	41.5	9.5
53	Rudy ENB, Game AO on, Rudy AO off.	37	42	40	44	40.75	8.75	The game's built-in AO is super-efficient for what you get, but IMO doesn't look nearly as nice as Rudy's.
54	Rudy ENB, Both Game AO and Rudy AO on Simultaneously	28	31	33	32	31	-1
55	Rudy ENB, No Game-AO, Self-Intersecting Rudy AO Disabled	33	37	37	38	36.25	4.25	Self-Intersecting Ambient-Occlusion is a weird one: whether the game looks better or worse with it is debatable, and turning it off can net you around 4 FPS... so... you should turn it off, right? Well hold on -- you'll notice that a few boxes down from here, I tested the effect that reducing the Ambient-Occlusion's "Resolution Scale" and "Texture Scale" has on performance. Reducing both values to 0.1 gains about 6.25 FPS... but, interestingly, doing so completely negates any gains that can be had from disabling the Self-Intersecting AO. The two properties are mutually exclusive! What I mean is: once I reduced the res and texture scales to 0.1, turning Self-Intersecting AO on or off had absolutely no impact on performance. That 4.25 FPS gain that could previously be had, disappeared completely. Personally, I think that the game looks better with the Self-Intersecting AO on, while the res and texture scales seem to have virtually no visual impact, so, contrary to the popular suggestion, I'd recommend leaving Self-Intersecting AO on, and dropping those two values to 0.1, instead. You get a GREATER FPS gain, while still being able to enjoy Self-Intersecting AO!
56	Rudy ENB , Water Tessellation Disabled (Tested at BASELINE testing location)	29	32	34	33	32	NO EFFECT	Water Tessellation was a bit weird, because several people told me I could gain a few FPS by turning it off, but I didn't detect even the slightest change in framerates, even while staring right at lake Ilinalta, or even while standing in the water, looking down at it! Might as well keep it on, methinks. (For what it's worth, I was using Realistic Water Two for the testing.)
57	Rudy ENB Water Tessellation Disabled (Tested at Lake Ilinalta)	29	32	34	33	32	NO EFFECT
58	Rudy ENB, No Depth-of-Field	31	34	37	36	34.5	2.5
59	Rudy ENB, Rudy AO Resolution Scale and Texture Scale both at 0.1	35	40	38	40	38.25	6.25	The Ambient-Occlusion resolution scale is a great place to save some FPS. Lowering it results in virtually no visual change or loss of quality, but nets you over 6 FPS. PLEASE NOTE THOUGH that once you make this change, you need to ALSO set "SamplingRange" to 1.0 under SSAO_SILL in enbseries.ini to avoid a weird visual artifact.
60	Rudy ENB, Rudy AO Resolution Scale and Texture Scale both at 1.0	21	22	28	24	23.75	-8.25
61	Rudy ENB, BethINI High	30	33	35	34	33	1
62	Rudy ENB, Changing AO Amount and AO Intensity to Any Value	29	32	34	33	32	NO EFFECT
63
64	Resolution Testing (Each line entry below is testing the FPS gains that come from decreasing the game's resolution from 3440x1440p to 2560x1080p. The FPS Change column states which BASELINE is being used as reference.) (Click HERE for a 0.5-second GIF showing the loss of quality from upscaling)
65	BASELINE at 2560x1080p	85	114	96	109	101	28.75 over BASELINE	This is another example of a time when the net results are less than the sum of their parts: Reducing the resolution of my client nets me nearly 29 FPS when playing on the BASELINE setup. This makes sense: 2560x1080p is only 56% the number of pixels for the graphics card to render as 3440x1440p. It also makes sense to assume that the same gains would be seen regardless of what graphical settings you got running, and yet.... with an ENB running, dropping the resolution only gained me 13.75 FPS over its full-res equivalent. The thing is, if we check the math, an increase of 13.75 over 32 (43%) is almost the exact same net gain as 28.75 over 72.25 (40%). It might be counterintuitive, but it checks out: the gains are proportional, not absolute.
66	Rudy BASELINE on Full at 2560x108p	45	51	51	52	49.75	17.75 over Rudy BASELINE
67	Rudy BASELINE on Full and Skyim 2020 4K Parallax Textures at 2560x1080p	39	50	44	50	45.75	13.75 over Rudy BASELINE
68
69
70	Final Setup Testing (Using what was learned from all of the tests above, the game client was optimized to gain as many FPS as possible, while retaining critical visual effects, like Rudy's Ambient Occlusion and Depth of Field. The final result of this optimization is a gain of an incredible 19.37 FPS over the identical setup without optimization -- with virtually no perceptible losses in visual fidelity.)												In Summary: If you are the average player, like me, and would like to run an ENB with some 4K textures, begin your next Skyrim Special Edition client setup by doing the following: 1) Use BethINI, and click the "high" preset button on the program's landing page. 2) Navigate through the other tabs in BethINI, and disable Lens Flare, disable Anamorphic Lens Flare, disable the Snow Shader, and set Screen-space reflections to a high divisor, or just disable it entirely. Additionally, set God Rays to low quality, and disable 64-bit render targets. 3) Install the ENB binaries and Rudy's ENB preset, then go into your enbseries.ini file, navigate to SSAO_SSIL, and set Resolution Scale to 0.1 and Texture Scale to 0.1. Next, set SamplingRange to 1.0. Be sure to save your changes. 4) ??? 5) Profit.
71	Rudy ENB with: BethINI High, no Lens Flare, No Anamorphic Lens Flare, No Screen Space Reflections, No Snow Shader	47.5	55	51	57	52.625	20.625
72	Rudy ENB with: The above ^^ and Skyrim 2020 4K Parallax by Pfuscher	42	53.5	44.5	55	48.75	16.75
73	Rudy ENB with: The above ^^ and Obsidian Weathers	41.5	54	44	55	48.625	16.625 over Rudy BASELINE, 19.37 over the same setup before optimizing	I can't lie - a gain of 19.37 FPS over what I started with nothing short of amazing, especially when considering that these gains were had with virtually no perceptible losses in visual quality. While my end result of ~ 41.5 FPS is definitely nothing fantastic, it's certainly playable, and enjoyable. Especially considering that this testing was done in the dense forests of Falkreath Hold.
74	Rudy ENB with: The above ^^ and "The Complicated" change to Windows Registry settings	95	103	99	107	101	69 over Rudy BASELINE	I made some good gains by optimizing the Skyrim client itself, but I was able to get AN ADDITIONAL 50 FPS!!! By making some modifications to Windows Regsitry settings. To say I'm speechless is putting it mildly. It's a bit hard to explain all the steps in text here, though, so I made a video to walk you through the process: https://tinyurl.com/SkyrimFPSGuide
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100