Star Trek Online DX11 Performance Review @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,724
Star Trek Online DX11 Performance Review - Star Trek Online has recently added DX11 Beta support, and we wanted to know how AMD's latest Radeon HD 7970 compared to NVIDIA's GeForce GTX 580 in the game and if DX11 provided any kind of performance difference over DX9. Our results were not what we expected.
 
Star Trek Online is programmed for nvidia cards. My 9800gt kept up with my 4870x2 when I played the game.
 
It is an nvidia sponsored game...

I don't think anyone is surprised AMD cards have a bit of trouble running it.

Skipping this game, even if it is now f2p.
 
Not surprise with the results..

Green shirt (NVIDIA)
capt-kirk_chair_1196285758.jpg


Red shirt (AMD)
red_shirt_star_trek.jpg
 
"Negative scaling." That's a euphemism I plan to try and incorporate into my vocabulary somehow.

"My salary is exhibiting negative scaling."

"Marital relations showed negative scaling with the late-night gaming feature enabled."

Oh, the possibilities!
 
Being that they start today witht he F2P i'll wait a few weeks to go back in. Quit a year ago out of sheer boredom. You can only click on the same Klingon/Romulan/Cardassian computer console so many times before it becomes a chore.

Having said this one of my BO was named Nina Hartley. Good times........
 
Good results for Nvidia, the implementation looks good for them at least.

I'd guess it'd improve if AMD cared enough to get driver hax out for this game.
 
Any chance you guys checked, even really quick, to see if the 69xx or 58xx series cards experienced the same drop? I'm curious to see if it's just an issue related to the new card and driver.
 
Hey HardOCP guys, are you going to do the same for World of Warcraft when the new 64-bit client comes out? Perhaps compare performances between AMD and Intel CPUs, as well as AMD and Nvidia video cards, with the new client? I would find that interesting enough to read all the way through. I don't play STO, despite being a big Trek fan, so this wasn't all that interesting to me.
 
I don't think AMD will improve these results considering the game is dead in the water. Always sucks when performance is completely lopsided towards one GPU manufacturer.
 
dgingeri said:
Hey HardOCP guys, are you going to do the same for World of Warcraft when the new 64-bit client comes out? Perhaps compare performances between AMD and Intel CPUs, as well as AMD and Nvidia video cards, with the new client? I would find that interesting enough to read all the way through. I don't play STO, despite being a big Trek fan, so this wasn't all that interesting to me.
Click to expand...

I wouldn't count on it considering EverQuest / World of Warcraft is about as un-Hard a game can possibly get. :p

d20.JPG


warrior_nerd.jpg


online+video+games+and+virtual+worlds+-hectorlorenzo.blogspot.com-wowdanddad.jpg


:p
 
Zarathustra[H];1038273540 said:
Why no apples to apples in DX11?
Click to expand...

The reason I ask is that it looks sort of as if DX11 grings the GTX580 up and the 7970 down, such that they meet and perform similarly.

Would love to compare them in the same chart with the same settings.
 
Wow...finally somebody is trying to do DX11 right. This game is practically a poster child for it. Unfortunatley, as long as the current gen of consoles is stuck in DX9 land...we are screwed. Probably be DX12 before the next gen of consoles comes out *shrug*.

One thing about the review...anytime a frame rate rockets to 797 vs 295...of course the averages are going to be screwed up. Therefore the sector space on page 4 just doesn't make sense. It isn't 4%...it is a ton more than that and the graph shows it. Honestly, in this case any FPS over 120 should be capped since the answer a red-herring for all intents and purposes. I konw you pointed out the min frame rate..but just say "the 797 vs 295 is screwing the average value"
 
Trepidati0n said:
Wow...finally somebody is trying to do DX11 right. This game is practically a poster child for it. Unfortunatley, as long as the current gen of consoles is stuck in DX9 land...we are screwed. Probably be DX12 before the next gen of consoles comes out *shrug*.

One thing about the review...anytime a frame rate rockets to 797 vs 295...of course the averages are going to be screwed up. Therefore the sector space on page 4 just doesn't make sense. It isn't 4%...it is a ton more than that and the graph shows it. Honestly, in this case any FPS over 120 should be capped since the answer a red-herring for all intents and purposes. I konw you pointed out the min frame rate..but just say "the 797 vs 295 is screwing the average value"
Click to expand...

Good post, and I agree :).
 
What about Radeon 5850/5870/6xxx cards? Anyone know of any DX9/DX11 comparisons on those?
 
Last edited:
Are you guys going to do the 6970 comparisons still for other games? I know you usually only use the latest hardware out, but I don't think there will be too many people buying the 7970 considering my 6970 plays everything maxed out.
 
Last edited:
Been playing for about 180 days now. I'm a lifetime subscriber as well. Love the game & love the graphics. Hoping more content gets injected into the game soon tho.
 
Zarathustra[H];1038273540 said:
Why no apples to apples in DX11?
Click to expand...

Because DX11 is broken on the 7970, best to run it in DX9 right now. Once this gets fixed, we can re-visit and do a DX11 comparison.

weebling1 said:
The-Doctor. ??

Who?
Click to expand...

Correct

AceCR42 said:
I don't think AMD will improve these results considering the game is dead in the water. Always sucks when performance is completely lopsided towards one GPU manufacturer.
Click to expand...

Personally, don't think the game is dead in the water. With F2P new blood is being injected into the game, and DX11 is just at the beginning, Cryptic seems poised to take further advantage of DX11 in the game, and I look forward to see what that is.

-MikeHawk- said:
Been playing for about 180 days now. I'm a lifetime subscriber as well. Love the game & love the graphics. Hoping more content gets injected into the game soon tho.
Click to expand...

Near 700 days right now, almost there to get my Android BO
 
I too am curious about AMD performance using older hardware and DX9 vs. DX11.

I will say though that at 1080p and using my 6950 2GB (MSI Twin Frozr III so factory O/C to 850MHz) the game runs basically perfect with everything set to max whether in DX9 or DX11 mode. I don't have problems with frame rate no matter where I go or what I do in the game. If you have a 6950 or better, there are no practical performance concerns whatsoever. Also it's nice that there is an option of a frame rate limiter, so you can cap it at 30/60/120.

I do notice that the GPU temps are up to 4C lower with the game in DX11 mode, but there are some visual artifacts that are not present in the DX9 mode (some textures are all black).

I have been using DX11 since the patch came out, maybe I should try going back to DX9 for a little again.
 
Question: Is DX11 broken also on 6970

Answer: Yes

Did a quick test at Starfleet Academy with 6970 and Cat 12.1a

DX9 - AVG 50.6 FPS
DX11 - AVG 34.3 FPS

Very Broken on 6900 series as well with Cat 12.1a in DX11
 
fdiaz78 said:
Being that they start today witht he F2P i'll wait a few weeks to go back in. Quit a year ago out of sheer boredom. You can only click on the same Klingon/Romulan/Cardassian computer console so many times before it becomes a chore.

Having said this one of my BO was named Nina Hartley. Good times........
Click to expand...

If your account was a paid one at any time, you could have started F2P earlier (apparently it was January 5th)...which I found out on Jan. 7th. Fortunately for previously payed players, they aren't completely stuck with F2P limitations (same applies to if you have 20 hours of play time in, or have Cryptic points on the account).
 
Just tried DX11 mode on my laptop (G73SW w/GTX460M), and ended up with a crashlog.
 
Should've added in the previous post...

Why does Cryptic send BOTH a mini-dump AND a full dump? The mini-dump weighed in at around 170 MB and the full dump is 560MB?!?
 
You're probably right that there was some issue with the test, but it could still be correct that the overall average between the two were within 4%. A max is just a max fps, it could have occurred for like 1/10th of a second which would barely register on a 5 minute test.

Trepidati0n said:
One thing about the review...anytime a frame rate rockets to 797 vs 295...of course the averages are going to be screwed up. Therefore the sector space on page 4 just doesn't make sense. It isn't 4%...it is a ton more than that and the graph shows it. Honestly, in this case any FPS over 120 should be capped since the answer a red-herring for all intents and purposes. I konw you pointed out the min frame rate..but just say "the 797 vs 295 is screwing the average value"
Click to expand...
 
Zarathustra[H];1038274578 said:
I wouldn't count on it considering EverQuest / World of Warcraft is about as un-Hard a game can possibly get. :p

d20.JPG
Click to expand...

BULL!

Heh! The gentleman taking a well-deserved rest on the giant D20 is one of our best Demo Team members for Catalyst Game Labs.

I only see him twice a year (Origins and GenCon). But every time I do, he's busting ass. This is one of the first years we were actually able to MAKE him take some time out for sleep.

Absolutely hilarious to see him turn up here on H.

As for the nVidia bias. Not surprised.

City of Heroes (Cryptic's first big MMO) just works with nVidia cards. Sure, there's a few gotchas, but it's an 8 year old game now.
But the AMD support is a fucking NIGHTMARE. We have people essentially playing "pick a driver" with their AMD/ATI cards. And it's anyone's guess which one is going to be stable for your setup. If any.

Champions Online (same engine as STO) more or less performs in-line with STO. Big surprise.
 
boden11 said:
You're probably right that there was some issue with the test, but it could still be correct that the overall average between the two were within 4%. A max is just a max fps, it could have occurred for like 1/10th of a second which would barely register on a 5 minute test.
Click to expand...

Except that when you look at the FPS chart it's obviously throwing off the average to a huge degree. Any chart where measurements go out of bounds (for any length of time) will alter the resulting calculated min/max/avg numbers. Normally it's not a big deal since if they do go out of bounds, it's usually not very far off. But shooting up to 797fps by rendering a giant black void, loading screen or menu is what's borking up this specific chart.

There's a ~20% performance difference between the two cards on that test, on average. Not 3%.
 
Huh, just playing devil's advocate this time, but being a beta dx11 setting and they admitting that they require a special driver backdoor to fine tune the performance, couldn't it be that it isn't amd driver's fault per se but that they have actually gotten part of the backdoor code working on nvidia but not yet on amd?

Still a good idea to poke amd about this.
 
I always thought AMD had better looking graphics, when it worked. I just couldn't put up with the broken drivers all the time. I gave up and went back to Nvidia.
 
Well a AMD beta tester has drivers that seem to work great with this game.

Jan 19, 2012
7950's in Crossfire
DX9 at 1920 x 1200, 8x AA, 16xAF and all the stuff on highest = 90fps
DX11 same settings 129fps

6990 is 120fps with the same settings.
Click to expand...
 
You must log in or register to reply here.
Back
Top