Another Godrays question

[WFRJ]

I have been reading about god-rays for months have been playing with mine set to low for the FPS, but I finally started doing some personal tests. (Already posted about load times)

Now I have a question:

I set a save then I ran 5 hours of real time gaming in the downtown area with god-rays off and got an average of 39 FPS

got out of the game and restated with god-rays set to low and got an average of 31 while covering the same routes (as close as I could)

got out again and set them to ultra and got an average of 36.

 

Huh??? Why??? Ultra, of course, looks better AND it ran better. Very confused here.

[zanity]

FO4 is insanely badly coded. It as if Beth decided to give up on the engine, and allow it to go out as an unoptimised mess- wait, that's EXACTLY what they did. Beth's AAA talent has been focused on OTHER things, including a completely NEW engine, since shortly after Fallout 4 began development years back.

 

FO4 was designed to be the last gasp on the PS3 (yes THREE). Only after most of the dev period was over did Beth finally decide to cancel the PS3 version. No way was Beth going to try to get a NEW engine running on the PS3- but the old Skyrim engine already did this job well enough. So FO4 was doomed, from day one, to be a Skyrim 'reskin'.

 

I always meant to try default Ultra godrays on my 6870 to see if the rumours of less of a hit than expected were true, but there were just too many other tweaks of greater importance to get the game running well on a 1GB VRAM card, so I kept the low GR setting.

 

The biggest engine sin of FO4 is the texture streaming function inhereted from John Carmack's Megatexture library. Turning this off gives a massive boost in performance and best quality texture stability.

 

Next comes the horrific CPU (yes CPU) rendered shadows- but this can be excused as a feature introduced in Skyrim in a desperate attempt to update the engine.

 

Godrays, of course, were part of a VERY profitable arrangement with Nvida to bang some late stage new-tech lighting into the game, and this was a worthwhile thing. But the downside is 'gameworks' shenanigans, where Nvidia uses code ownership to synthetically boost performance on its latest GPUs- as a marketing sales mechanism. Godrays were no where as ruinous as the Gameworks features Nvidia paid to have included in the latest Batman game, but issues remain that are 100% down to Nvidia's cynical war against PC gamers.

 

TL;DR- BETTER settings tend to run faster for many people than 'lower' settings on many PC configurations (sometimes with a little assistance from 'tweaks'). This is a DIRECT consequence of Beth engaging in SPEC INFLATION to please the requirements of hardware manufacturers. Beth, for instance, up-sampled many of the textures so they'd need up to FOUR TIMES more VRAM than the actual image data they contained, just to allow a 'minimum' GPU VRAM spec requirement that made the 4GB cards from Nvidia and the 8GB cards from AMD seem like a good idea. Dirty tricks to the max.

Edited June 24, 2016 by zanity

[NorthWolf]

I have been reading about god-rays for months have been playing with mine set to low for the FPS, but I finally started doing some personal tests. (Already posted about load times)

Now I have a question:

I set a save then I ran 5 hours of real time gaming in the downtown area with god-rays off and got an average of 39 FPS

got out of the game and restated with god-rays set to low and got an average of 31 while covering the same routes (as close as I could)

got out again and set them to ultra and got an average of 36.

 

Huh??? Why??? Ultra, of course, looks better AND it ran better. Very confused here.

 

The problem could either be:

 

1. Your graphics drivers aren't really optimized to work with the low setting and thus the ultra setting is giving you better results due to optimization.

 

OR, more likely...

 

2. Your testing methodology is flawed. It might seem like 5 hours should average out enough that comparing results will yield valid results, but unless you were taking virtually the same route every time you tested during those five hours it's not particularly meaningful. I would suggest finding an area that displays poor performance with godrays on the low setting. Do a quicksave and then start your test: walking 10-30 seconds in a straight line with autowalk. Then load the quicksave with godrays on ultra and try the same test. If you still get similar results then my only guess is that the above is the case.

[WFRJ]

Thanks for the replies and info.

 

My method was to save. Start playing and walking the downtown always trying to turn right. Started a noon (ish) game time and made a note of my FPS often till I got 500 samples.

restarted from scratch (even rebooted) and tried to follow the same path and recording fps places. Was very very bored by the end of the third walk thru I must say.

 

So you say the card may not be optimized for low but for ultra?? Sound counterintuitive to me. Most odd. Bad specs indeed Zanity!

I run a 980ti.

 

Still, O can live with mid 30s in town and middle 50s in the country.

[NorthWolf]

So you say the card may not be optimized for low but for ultra?? Sound counterintuitive to me. Most odd. Bad specs indeed Zanity!

 

It's just a guess as to why you would see such counter-intuitive behavior. You'd have to find somebody who actually knows what goes into graphical optimization to know if my suggestion has any merit.