7970 Vs 770 - Stock & OC Gaming Benchmarks

LtMatt · 24 Jun 2013 at 15:11

Gregster said:
This is why I never trust paid reviewers. That Hitman bench is a standardised in game bench result, so all those fluctuations are clearly wrong and the same goes for Tomb Raider.

A few brown envelopes fly about me thinks

The problem is that not every site uses the in game benchmark. If they did then our results would more often match up. I hate benching games without an in game benchmark, its not consistent enough. At least i know with Hitman, Tomb Raider, Sleeping Dogs, Bioshock etc you get the same results everytime.

jimfowler_82 · 24 Jun 2013 at 15:12

ah yes i forgot i have a SC version for a moment

LtMatt · 24 Jun 2013 at 15:14

jimfowler_82 said:
ah yes i forgot i have a SC version for a moment

Jim see how you stack up to a clocked 7970 in Tomb Raider.

Copy these settings

1080p, Ultimate preset, FXAA only, and use TressFx. Overclock that 770 as far as it will go and post your benchmark score.

jimfowler_82 · 24 Jun 2013 at 15:14

Will do give me 10 mins

Gregster · 24 Jun 2013 at 15:15

LtMatt said:
The problem is that not every site uses the in game benchmark. If they did then our results would more often match up. I hate benching games without an in game benchmark, its not consistent enough. At least i know with Hitman, Tomb Raider, Sleeping Dogs, Bioshock etc you get the same results everytime.

Agreed.

When we did the BF3 bench, if someone was pap at it and someone was awesomes, the results would not tally up even on the exact same systems.

I do wish more devs would implement a standardised bench like Tomb Raider/Bioshock/Metro LL etc.

LtMatt · 24 Jun 2013 at 15:16

jimfowler_82 said:
Will do give me 10 mins

Atta boy. This is what Scotty managed with his 770.

whyscotty said:
2500k @ 5.0

770 Lightning 1385/8210

Uploaded with ImageShack.us

Gregster said:
Agreed.

When we did the BF3 bench, if someone was pap at it and someone was awesomes, the results would not tally up even on the exact same systems.

I do wish more devs would implement a standardised bench like Tomb Raider/Bioshock/Metro LL etc.

Yes will be delighted if they get a proper bench in bf4, really hope they do. Will settle some arguments for sure.

jimfowler_82 · 24 Jun 2013 at 15:18

67.8fps avg

jimfowler_82 · 24 Jun 2013 at 15:18

Im a bit scared to change voltage yet so thats on default voltage 106% power +100gpu +400mem

LtMatt · 24 Jun 2013 at 15:29

Definitely cpu bottlenecked at 1080p with x4 AA. Even with the cpu at 5ghz usage was still only 67-77%.

With x8 usage was a bit higher, 87-97%, though it only crept up to 97% at the end.

jimfowler_82 said:
67.8fps avg

Not bad, you need to find a few more fps to tie with scotty and double that to catch the 7970 though.

jimfowler_82 · 24 Jun 2013 at 15:35

With a little more tweaking

CPU @ 4.4ghz
GPU 1320mhz | Memory 1987mhz

jimfowler_82 · 24 Jun 2013 at 15:40

Guys this could end up an expensive new habit for me chasing numbers lol.

LtMatt · 24 Jun 2013 at 16:06

jimfowler_82 said:
With a little more tweaking

CPU @ 4.4ghz
GPU 1320mhz | Memory 1987mhz

jimfowler_82 said:
Guys this could end up an expensive new habit for me chasing numbers lol.

Good fun innit.

Here's a cheeky little 7950 entry at 1225/1750 to beat Jim's 770. Might be able to squeeze a bit more out of it yet, so scotty should prepare an orifice in case this £200 gpu takes the lead.

jimfowler_82 · 24 Jun 2013 at 16:12

Thats an impressive score for a 7950 but isnt TressFX like running PHYSx on an AMD card?

LtMatt · 24 Jun 2013 at 16:18

jimfowler_82 said:
Thats an impressive score for a 7950 but isnt TressFX like running PHYSx on an AMD card?

No as TressFX and Direct Compute are open to everyone. Its basically just part of Directx 11.

Taken from Toms

"TressFX will provide realism to hair rendering in ways that otherwise can only be done with CGI or pre-rendered scenes. The trick to this is DirectCompute, allowing much better visuals with a minimal impact on CPU performance. The technology is similar to Nvidia's PhysX, making the GPU do the computations rather than the CPU."

Taken from Wiki

"Microsoft DirectCompute is an application programming interface (API) that supports general-purpose computing on graphics processing units on Microsoft Windows Vista, Windows 7 and Windows 8. DirectCompute is part of the Microsoft DirectX collection of APIs and was initially released with the DirectX 11 API but runs on both DirectX 10 and DirectX 11 graphics processing units."

I need to try and beat Scottys score.

jimfowler_82 · 24 Jun 2013 at 16:32

Ill be ramping up the cpu to 4.5ghz and trying again with beta drivers to try to get another 1fps

LtMatt · 24 Jun 2013 at 16:33

jimfowler_82 said:
Ill be ramping up the cpu to 4.5ghz and trying again with beta drivers to try to get another 1fps

That's the spirit. Not sure i can better my score so if you get another 1fps you will win.

Uncle Petey · 24 Jun 2013 at 16:38

Sorry for the delay

1250/1900

1300/1800

Yay! An extra 2fps min and 1.4fps avg! So worth it.

LtMatt · 24 Jun 2013 at 16:43

Uncle Petey said:
Sorry for the delay

1250/1900

1300/1800

Yay! An extra 2fps min and 1.4fps avg! So worth it.

Well done Petey. Now see if you can break my sleeping dogs records posted earlier.

Did you have to drop the memory down to get 1300 stable? I find 1823 is a good number but if you can go higher then that should help so long as error correction is not kicking in.

Uncle Petey · 24 Jun 2013 at 16:48

Nah, I used Trixx to get the core voltage over 1.3 but there's no memory voltage adjustment on there unless there's a modded/newer version about.
Can I get AB to run over 1.3v?

I haven't really been keeping up with stuff tbh

jimfowler_82 · 24 Jun 2013 at 16:54

Sorry mate but...