PCGHRyseSon of Rome benchmarks

[Silverforce11]

May pick up the game when it hits $5-10. Looked really boring on consoles. Doesn't cryengine usually run better on AMD? So its not that surprising.

Well no, Crysis usually ran better (or much better) on NV hardware, every since Crysis 2.

 

[Carfax83]

Its beating it with a huge clock speed deficit, clearly this game engine is extremely well threaded for that to happen. Kudos to Crytek. I've always considered Crytek to be a poor game developer but a great game engine maker.

I never said it wasn't well threaded, only that there are other factors involved as well like the microarchitecture and the enormous cache..

Also, Star Citizen is so hyped I don't know if it can live up to it.

I just want to play the single player campaign. I hope they don't neglect that portion, as not everyone has the time or inclination to play online.

 

[Kallogan]

that game looks stunnin, will not even try on my 750M lol

 

[bystander36]

Its beating it with a huge clock speed deficit, clearly this game engine is extremely well threaded for that to happen. Kudos to Crytek. I've always considered Crytek to be a poor game developer but a great game engine maker.

Also, Star Citizen is so hyped I don't know if it can live up to it.

While newer Crytek games haven't been that great, I still found Farcry (the original before it was sold to another dev) to be my favorite 1st person shooter of all time. Perhaps it was due to the game mechanics related to what their engine could do, but it was a great game. Crysis was still good, but I'd rather them lose the suit. I loved using cover for stealth in the way Farcry did.

 

[TreVader]

So is the 980 smoke and mirrors and really slower than the 780ti?

You've got Titan killing the 980, I guess because DP?


Are these compute shaders done in DP?


This card is weird. It over clocks really well, but you can't over clock it because of TDP limit? It's faster than a 780ti, but slower in new games? I'm on the fence.

 

[toyota]

So is the 980 smoke and mirrors and really slower than the 780ti?

You've got Titan killing the 980, I guess because DP?


Are these compute shaders done in DP?


This card is weird. It over clocks really well, but you can't over clock it because of TDP limit? It's faster than a 780ti, but slower in new games? I'm on the fence.

where do you see the Titan? there is a 780 ti there which is slower than a Titan but slightly losing to the 980. plus the 780 ti being used has a higher factory percentage oc on it than the 980.

 

[x3sphere]

This card is weird. It over clocks really well, but you can't over clock it because of TDP limit? It's faster than a 780ti, but slower in new games? I'm on the fence.

No idea what you're on about here. Many people are getting 20% overclocks or close to it, which is usually about as good as it gets on air with any recent GPU. The TDP limit is presumably preventing us from overclocking it further though.

 

[TreVader]

No idea what you're on about here. Many people are getting 20% overclocks or close to it, which is usually about as good as it gets on air with any recent GPU. The TDP limit is presumably preventing us from overclocking it further though.

Actually it's a very poor overclocker because the 780ti can hit 1300 often and scales better.



I've seen benchmark after benchmark showing the 780ti catching up and even surpassing the 980 when compared OC to OC.


It posts high clocks, but throttles for TDP and generally gains much less than 780/ti

 

[x3sphere]

Actually it's a very poor overclocker because the 780ti can hit 1300 often and scales better.



I've seen benchmark after benchmark showing the 780ti catching up and even surpassing the 980 when compared OC to OC.


It posts high clocks, but throttles for TDP and generally gains much less than 780/ti

I wouldn't say 1300 Mhz is a common OC for the 780 Ti.

 

[TreVader]

I wouldn't say 1300 Mhz is a common OC for the 780 Ti.

True, but I wouldn't call it rare either. It's a good OC.


My point isn't that the 980 is bad i am having a hard time deciding if it's good or not. The revelations that all nvidia did was put more aggressive turbo and power gating to lower power consumption but only in games, instead if the implied magic, has me wondering if it's a gimmick.

I would be upset if I bought a card that nvidia will let some jerk OC to 2GHz as a stunt, then voltage lock my card so I then can never hope to replicate that.


This is why I never buy nvidia unless apple makes me. They are greedy and obv don't give a crap about over clocking unless it's for a Publicity stunt to make money.

 

[lyssword]

looks awesome on my g3258/270x on medium settings

 

[Fox5]

seems like going from 4 to 8 cores on the AMD side is giving a 15-20% gain or so?
(if you consider the 4300 is running at lower clock, have less l3), if you compare 4 cores to 4 cores with HT on the Intel side with 2500 and 2600 is also no to far

but the 5690X is gaining a lot more compared to haswell, if you correct the clocks, but there are other big differences involved (memory/cache)

It does seem that Intel gains a lot more than AMD in this test. Even hyper threading shows big gains for Intel in this.

 

[psolord]

The latest patch mentions that it fixes Nvidia performance. Especially the Maxwell's.

Anyone tried it yet?

 

[Magic Carpet]

The latest patch mentions that it fixes Nvidia performance. Especially the Maxwell's.

Anyone tried it yet?

Yeah, would be good to know :whiste:

 

[Gloomy]

Everyone I know with an Nvidia card that owns the game has mentioned it is much smoother.

I know it's unscientific but since people are noticing it, it's no small change.

 

[Erenhardt]

I think these are post patch results:
http://www.pcgameshardware.de/Ryse-PC-259308/Specials/Test-Technik-1138543/

 

[Fire&Blood]

Some of the graphics arguments bounce between subjective and objective a lot, hard to make a case there. I think it looks great, can't say it's the best looking game but I am positive Ryse has the best looking vegetation I have ever seen and the other elements aren't too far behind either.

I haven't noticed any differences after the patch with a AMD card.
I still get impressive results maxed out @1440p with the sig rig. I have the temporal AA on, taxes about 2-3 fps. The CPU is locked and GPU is stock, 880/1250. I kept the settings stock for these shots but as reminder this is maxed out @1440p, yet I can easily keep the game above 40fps if I turn AA off and close in on 50 if I push the core above 1GHz. As it is , all stock and frame rate overlay is on but these are jpg's, just don't do in game image quality justice:

 

[RussianSensation]

^Rick, you have MSI Twin Frozr III HD7950 and you are running it at 880mhz? Is it the 2x8-pin version too?

Those were highly binned 7950 V2 cards, and overclocked to 1.2-1.3Ghz at 1.256-1.275V. Even at 1.225V, many of those cards hit 1.2Ghz. You are leaving a lot of performance on the table. I bet at your stock voltage you can do 1.15-1.175Ghz.

http://forums.anandtech.com/showthread.php?t=2259333

Both of my Sapphires can do 1.150Ghz at only 1.174-1.175V (essentially stock voltage of Tahiti XT chip). I need 1.225V to hit 1.20Ghz.

I think these are post patch results:
http://www.pcgameshardware.de/Ryse-PC-259308/Specials/Test-Technik-1138543/

Results still state October 9, 2014. Also, the description doesn't align with the scores.

"Updated on 23/10/14:
Crytek has now released a first patch for Ryse, which will bring next smaller bug fixes and new features (HUD and controller vibrations deactivated, support for "high resolution timer" mode) on Nvidia graphics cards in particular improvements. Our graphics guru Raff has the improvements of the patches seen on his private system, so we can give you a first impression.

With an FX-8350 (4.006 MHz, NB: 2671 MHz) as well as the GTX Titan (1,202 / 3,600 MHz) the gain is through the patch depending on the resolution and settings at 2.7 (1080p, TAA, 16: 1 AF) to 4.8 percent (1440p, SSAA + TAA, 16: 1 AF). The scene is not comparable to that of our original benchmarks, but represents a CPU lastigsten is - nevertheless you will find yourself even here in the GPU limit. According to Crytek to grab especially on GTX 900 graphics cards support improvements. With a GTX 980 OC was achieved in a fast pass at least an increase of 11.3 percent in Full HD, 2 × 2 SSAA and 16: 1 AF.

We will use this day to the delivery tomorrow, to create more detailed benchmarks for our print edition 12/2014."

These benches still in the article are from the original drivers and the original section of the game they tested. Looks like they haven't posted new data.

 

[Fire&Blood]

^Rick, you have MSI Twin Frozr III HD7950 and you are running it at 880mhz? Is it the 2x8-pin version too?

Those were highly binned 7950 V2 cards, and overclocked to 1.2-1.3Ghz at 1.256-1.275V. Even at 1.225V, many of those cards hit 1.2Ghz. You are leaving a lot of performance on the table. I bet at your stock voltage you can do 1.15-1.175Ghz.

http://forums.anandtech.com/showthread.php?t=2259333

Both of my Sapphires can do 1.150Ghz at only 1.174-1.175V (essentially stock voltage of Tahiti XT chip). I need 1.225V to hit 1.20Ghz.



Results still state October 9, 2014. Also, the description doesn't align with the scores.

"Updated on 23/10/14:
Crytek has now released a first patch for Ryse, which will bring next smaller bug fixes and new features (HUD and controller vibrations deactivated, support for "high resolution timer" mode) on Nvidia graphics cards in particular improvements. Our graphics guru Raff has the improvements of the patches seen on his private system, so we can give you a first impression.

With an FX-8350 (4.006 MHz, NB: 2671 MHz) as well as the GTX Titan (1,202 / 3,600 MHz) the gain is through the patch depending on the resolution and settings at 2.7 (1080p, TAA, 16: 1 AF) to 4.8 percent (1440p, SSAA + TAA, 16: 1 AF). The scene is not comparable to that of our original benchmarks, but represents a CPU lastigsten is - nevertheless you will find yourself even here in the GPU limit. According to Crytek to grab especially on GTX 900 graphics cards support improvements. With a GTX 980 OC was achieved in a fast pass at least an increase of 11.3 percent in Full HD, 2 × 2 SSAA and 16: 1 AF.

We will use this day to the delivery tomorrow, to create more detailed benchmarks for our print edition 12/2014."

These benches still in the article are from the original drivers and the original section of the game they tested. Looks like they haven't posted new data.

Yep, I have some runs posted in that thread :biggrin: : http://forums.anandtech.com/showpost.php?p=33822143&postcount=238

I ran the game at stock clocks to make a point about Ryse efficiency and it's looks on a old rig that was fairly weak to begin with, at 1440p and maxed out to top it off. I think it's too bad that Crytek doesn't pump out games like Ubisoft, I would expect them to challenge everything on PC both in terms of efficiency and image quality. Again, the screens I posted don't do the game justice. I have run the game in past at 1175/1525, it closes in on 50FPS but to highlight the efficiency screens were taken at stock speeds.

 

[Erenhardt]

There are results after patch.
http://www.computerbase.de/2014-10/...n/#diagramm-grafikkarten-benchmarks-2560-1600

900 series get >10% boost with patch
lesser gains for the rest.

 

[96Firebird]

Pretty decent gains from most Nvidia cards, makes you wonder why a game patch was needed in the first place...

Summary (Card, Gain @ 1080p, Gain @ 1600p)...

Nvidia:
GTX 980, 11%, 13%
GTX 970, 12%, 12%
GTX 780Ti, 8%, 7%
GTX 770, 9%, 7%
GTX 760, 4%, N/A

AMD:
R9 290X, 2%, 4%
R9 280X, 1%, 2%
R9 285, 2%, 3%
R9 270X, 0%, N/A

 

[psolord]

There are results after patch.
http://www.computerbase.de/2014-10/...n/#diagramm-grafikkarten-benchmarks-2560-1600

900 series get >10% boost with patch
lesser gains for the rest.

thanks

cb needs to get a 290 ;p

 

[Muhammed]

looks like NV got the performance back on track:

http://pclab.pl/art60277-5.html
http://pclab.pl/art60277-6.html

 

[DiogoDX]

pclab

kkkkkkkkkkkkkkkkkkk

 

[papapapapapapapababy]

than cant be rigth... i ended the game with a 3.6 4160 + oced 750 Ti at 45+fps ( frapz) with everithing set to high and 80% resolution... maybe they didint use the lastest drivers @ patchs?


(btw it looks lovely but the world physics are a joke, the original crysis > this console game i blame the subpar amd cpu.. everithing in the world is statick xpts 4 the main dude
( bonus focus 4 breakin a base!)