Prey benchmarks [techpowerup]

[frowertr]

Or you could have bought it from Steam at its normal price where the money would actually go to the developer and find you didn't like the game in the 2 hours you get to play it then get a full refund.

Its not worth more than $30. No game is. I refuse to "support" developers at that price point. I want to enable the price down, not up. Sorry not sorry...

 

[Malogeek]

The point is that the GTX 1060 having more 'headroom' to maintain a 60fps lock at 1440p, like they talk about in the video, isn't something you can claim without locking the fps in the first place and measuring the %tage of time spent below 60fps.

Why do you need to lock it to 60fps if they're measuring it the whole time and can see if it dips below 60fps at any point? Are you saying if it's 60fps capped the game will perform differently?

With it unlocked, someone who wants to run 1080 at 100Hz can determine if that's possible as well.

 

[tamz_msc]

Why do you need to lock it to 60fps if they're measuring it the whole time and can see if it dips below 60fps at any point? Are you saying if it's 60fps capped the game will perform differently?

With it unlocked, someone who wants to run 1080 at 100Hz can determine if that's possible as well.

All I'm saying is that if someone wants to experience it at 60FPS, then they should collect the data for a fairly short(1min) run at a locked 60fps and measure how many frames out of 3600 fell outside the 16.67 millisecond target, instead of saying subjective things like the 'GTX 1060 is more likely to hold its 60fps lock than the RX 580'.

You've got the tools to measure it, why not use them?

 

[Bacon1]

Its not worth more than $30. No game is. I refuse to "support" developers at that price point. I want to enable the price down, not up. Sorry not sorry...

So then wait for a sale. Buying keys from key reselling sites is often not legit. Might as well just pirate it if you aren't willing to pay the actual developers. There are many legit sites outside of steam / origin.

You've got the tools to measure it, why not use them?

Yeah, its sad when sites have done 90% of the work and don't finalize it

 

[Bacon1]

People found more missing shadows from DF's video:

 

[tamz_msc]

Was about to post this, seems to be a case of either settings not being applied properly or NVIDIA just not rendering shadows correctly. It's funny seeing the specular lighting being repeated down the corridor till the exit sign where there shouldn't be any lighting in the first place.

 

[Malogeek]

Unless it's the complete opposite and there's some kind of GI not being rendered on the AMD

 

[Bacon1]

Unless it's the complete opposite and there's some kind of GI not being rendered on the AMD

Should at least be a soft shadow as the light is inside the room it's also missing the shadow from the other side of the entry but wasn't circled

 

[Bacon1]

Sorry for OT regarding CDKeys, back on topic.

DF seem to have a smoother experience with the 1060 on Prey?

The Nvidia side is also missing character shadows. Watch the opening sequence on both and you'll see the character moving through the door and you can see their shadow when going to pickup the wrench on the 580 but not on the 1060. I wonder if they have shadows at high vs VH or something or if the character model isn't being rendered on Nvidia's side for optimization. Starts about 1 minute into the video, maybe someone can get some screenshots or a gif

Edit: Made some gifs, neat website!

https://gifs.com/gif/prey-pc-runs-beautifully-gtx-1060-rx-580-tested-plus-budget-pc-analysis-xG4Q9r

https://gifs.com/gif/prey-pc-runs-beautifully-gtx-1060-rx-580-tested-plus-budget-pc-analysis-3lD7EQ

 

[railven]

Tinfoil other way, wonder if this is AMD's version of NV's schtick.

Would be hilarious if shadows or more IQ options are exclusive to AMD. No need to hamper your rivals performance, just lock out some casual IQ details when their vendor ID shows up.

 

[Bacon1]

Tinfoil other way, wonder if this is AMD's version of NV's schtick.

Would be hilarious if shadows or more IQ options are exclusive to AMD. No need to hamper your rivals performance, just lock out some casual IQ details when their vendor ID shows up.

Seems like a bad idea, as many people won't notice (hell DF didn't seem to mention it and they usually do lots of IQ checks... they also missed out when testing Hyper Shadows on 970 vs 390 though....) but will look at the FPS numbers so you'd just make your stuff look worse. I mean look at the gifs I posted, had to even notice that the character's shadow isn't there unless you watch it a few times. It's very subtle.

Edit: Yep, if you look at when they used Shadows: High for 1440p, both cards are missing the shadows in that scene with the enemy (image with circles posted above) and neither appear to have the character shadow either in the opening sequence.

So looks like they tested NV with shadows: High, or for some reason the shadows were capping at high for NV during their 1080p testing.

https://gifs.com/gif/prey-pc-runs-beautifully-gtx-1060-rx-580-tested-plus-budget-pc-analysis-pg5Gor

You can also see shadows from the enemy are missing. So seems like character / enemy based shadows as well as the ground shadows are from Very High as they are both missing later in the video when testing 1440p on High for both cards

 

[railven]

Seems like a bad idea, as many people won't notice (hell DF didn't seem to mention it and they usually do lots of IQ checks... they also missed out when testing Hyper Shadows on 970 vs 390 though....) but will look at the FPS numbers so you'd just make your stuff look worse. I mean look at the gifs I posted, had to even notice that the character's shadow isn't there unless you watch it a few times. It's very subtle.

Edit: Yep, if you look at when they used Shadows: High for 1440p, both cards are missing the shadows in that scene with the enemy (image with circles posted above) and neither appear to have the character shadow either in the opening sequence.

So looks like they tested NV with shadows: High, or for some reason the shadows were capping at high for NV during their 1080p testing.

NV sponsored games on Ultra look amazing* but run like a turd. It be a similar approach, granted if that what were happening and not just user error, in my opinion.

NV sponsored game: 20 FPS but you get all the NV goodies.
AMD Sponsored Game: 55 FPS but you get, welps shadows!

*Amazing used subjectively. Some settings aren't worth the performance hit.

 

[amenx]

People found more missing shadows from DF's video:

Havent read all the YT comments but did anyone mention @ 3:43?

Oops.. what happened to the shadows on AMD's side?

 

[Bacon1]

Havent read all the YT comments but did anyone mention @ 3:43?



Oops.. what happened to the shadows on AMD's side?

You could have just read my comments above. That is from 1440p which they turned shadows down to High from Very High. But it seems like NV was always on High and only AMD was tested on Very High 1080p.

 

[Malogeek]

So basically they messed up the settings for the comparison.

 

[Krteq]

Hmm, what about this - 1:30 - shadows and lightning in that room

 

[Bacon1]

More missing shadows from Joker's video:

Chairs not shadowed on 1060:
https://youtu.be/3_RUg4kydzI?t=273

But @ 1440p test they do show:
https://youtu.be/3_RUg4kydzI?t=348

Then missing again on his 1080 Ti 1440 vs 4k test:
https://youtu.be/3_RUg4kydzI?t=391

People have said that some settings don't apply correctly on launch and only when you disable/re-enable them while in game, so maybe that is part of the reason.

Either way it makes trying to do comparisons / benchmarks without video or IQ checks pretty much impossible. Hopefully they release a patch soon.

 

[swilli89]

Hmm, what about this - 1:30 - shadows and lightning in that room

So multiple sources now showing that nvidia's "game ready" driver either is of bad quality and buggy or maliciously causing the card to not render all of the shadows to boost performance.

 

[tamz_msc]

So multiple sources now showing that nvidia's "game ready" driver either is of bad quality and buggy or maliciously causing the card to not render all of the shadows to boost performance.

Or the game might be behaving in an odd manner when it comes to applying changed settings.

 

[Bacon1]

Or the game might be behaving in an odd manner when it comes to applying changed settings.

One of the posters on reddit mentioned that changing shadows while in game has no effect until you re-load the game. So you could have it on low, change to Very High and play thinking it was on very high, but it would still be actually rendering in low. Probably what happened to a lot of reviewers when swapping resolutions and settings.

Sadly the reviewers aren't catching it and doubt many of them will redo their testing with the proper settings.

 

[tamz_msc]

One of the posters on reddit mentioned that changing shadows while in game has no effect until you re-load the game. So you could have it on low, change to Very High and play thinking it was on very high, but it would still be actually rendering in low. Probably what happened to a lot of reviewers when swapping resolutions and settings.

Sadly the reviewers aren't catching it and doubt many of them will redo their testing with the proper settings.

They got their page/youtube views. From their perspective they have no incentive to redo the tests. They'd much rather move on to new content.

 

[Head1985]

https://www.purepc.pl/karty_graficz...c_wymagania_niskie_ale_grafika_slaba?page=0,9

 

[railven]

I just realized this game is running off CryEngine. For some reason I assumed they were using a Bethesda inhouse engine.

Reading the stuff Bacon1 found, this reminds me exactly of of ArcheAge, which also uses the CryEngine. The issues are almost to a T. I ended up rage quitting due to the shadows and their random behaviors.

Sometimes I'd have to quit the game 5 times for settings to take affect. You can just toss ArcheAge + Shadows into Google and get a boat load of hits.

EDIT:
Some references:
Shadows:
https://www.google.com/search?q=arc...me&ie=UTF-8#safe=off&q=archeage+shadow+issues

Settings not Saving:
https://www.google.com/search?q=arc...1.69i57j0l5.3663j0j7&sourceid=chrome&ie=UTF-8

 

[William Gaatjes]

Just a silly question : In the scenario that the gpu driver in general would replace game shaders for optimized versions, how would the driver know that you changed a in game setting ?
How does the game inform the driver that different shaders are used or more shaders are used for example for shadows.
Is there some kind of directx function for that to let the driver know that image quality settings have changed ? Is the driver itself probing the shaders on the fly to see if it needs to use different shaders ?

I noticed that some games when applying graphic changes, need to almost restart the game while other games just change on the fly.

 

[dogen1]

Just a silly question : In the scenario that the gpu driver in general would replace game shaders for optimized versions, how would the driver know that you changed a in game setting ?
How does the game inform the driver that different shaders are used or more shaders are used for example for shadows.

If the game is using different shaders, it's sending different shaders to the driver. It's that simple.