cat 12.9 bf3 benches ?

[The Alias]

I was lurking on some forums and alot of people say their bf3 frame rate has gone up quite a bit . Is there any place that has done official benches in the game and compared amd cards to their nvidia counterparts ?

 

[VulgarDisplay]

The forums I've seen stated that they got around a 5-7% increase. Not massive, but it further closes the gap with Nvidia. I'd say they would be about neck and neck in BF3 if the 5-7% number is true. Very good news for AMD owners especially since BF3 is the only game out right now that needs the extra FPS worth playing, and it's the game that a lot of Nvidia GPU owners touted as the reason to go nvidia.

 

[f1sherman]

I was lurking on some forums and alot of people say their bf3 frame rate has gone up quite a bit

http://benchmark3d.com/amd-catalyst-12-9-beta-9-001-benchmark/2

don't discard everything you read in driver threads, but you won't lose much if you do

 

[Olikan]

http://benchmark3d.com/amd-catalyst-12-9-beta-9-001-benchmark/2

don't discard everything you read in driver threads, but you won't lose much if you do

this bench was done with a HD6950...not 7xxx series

 

[f1sherman]

sure... and if you add all the times that new driver was released and someone reported Image Quality and FPS increase,

by now you should have something like 4k x 4k crispness running at hundreds of fps

 

[zaydq]

Well... I don't own BF3 so I can't say much, but I do have 12.9 and I noticed a few higher scores in 3dmark11 and heaven (could be driver cheating though?) . I've heard from a couple of members an increase in BF3 frames.... could be true.

 

[RussianSensation]

I find it ironic that when a new architecture like Fermi came out, we fully expected NV to increase performance by about 10% with drivers to optimize for it. However, when AMD does the same thing with a brand new GCN architecture, it's "OMG, it took them 8 months to get their cards to perform well enough." Talk about a double standard. New major architectures like Fermi and GCN are expected to benefit significantly from driver updates, unlike refreshed architectures. Kepler is not a new architecture per say, but just a reworked Fermi, which is why we have seen almost no driver increases from NV this generation. That's more or less expected and it's also expected that a new architecture like GCN will have more issues initially and at the same time experience far larger performance gains - we have seen both of these scenarios play out this year.

Even with Cats 12.8, the performance in Armored Kill is now just 3 fps apart between 7970 and 670 and 7970 GE and 680:

I expect by the end of the year that gap is closed completely, but NV has done little to make up the massive performance gaps it has in Alan Wake, Anno 2070, Sleeping Dogs, Metro 2033, Crysis 1 / Warhead, SKYRIM with Mods, etc. etc.

Looks like BF3 is the last major game where 680 has any lead at all and that lead has now dwindled to less than 5% in the majority of recent testing.

 

[Mastercommander]

Unfortunately I don't like any of those games except Skyrim and Sleeping Dogs, which run great already @ 1080p on my 660 Ti

 

[3DVagabond]

OCaholic has done a performance review of the 12.9's on the 7970. Seems like a solid release. Shows small improvement across the board. No obvious glitches or bugs.

 

[notty22]

I find it ironic that when a new architecture like Fermi came out, we fully expected NV to increase performance by about 10% with drivers to optimize for it. However, when AMD does the same thing with a brand new GCN architecture, it's "OMG, it took them 8 months to get their cards to perform well enough." Talk about a double standard. New major architectures like Fermi and GCN are expected to benefit significantly from driver updates, unlike refreshed architectures. Kepler is not a new architecture per say, but just a reworked Fermi, which is why we have seen almost no driver increases from NV this generation. That's more or less expected and it's also expected that a new architecture like GCN will have more issues initially and at the same time experience far larger performance gains - we have seen both of these scenarios play out this year.

Even with Cats 12.8, the performance in Armored Kill is now just 3 fps apart between 7970 and 670 and 7970 GE and 680:

I expect by the end of the year that gap is closed completely, but NV has done little to make up the massive performance gaps it has in Alan Wake, Anno 2070, Sleeping Dogs, Metro 2033, Crysis 1 / Warhead, SKYRIM with Mods, etc. etc.

Looks like BF3 is the last major game where 680 has any lead at all and that lead has now dwindled to less than 5% in the majority of recent testing.

edit: removed comment. The comment I responded to was posted days ago, not worth the effort pointing out my issues with these responses.

 

[blastingcap]

OCaholic has done a performance review of the 12.9's on the 7970. Seems like a solid release. Shows small improvement across the board. No obvious glitches or bugs.

Good to hear though frankly with such small improvements I'd rather stay on tried-and-true driver releases just in case there are hidden bugs in 12.9. According to their chart, that's less than 1 frame per second difference in BF3 at the settings they used.

 

[3DVagabond]

Good to hear though frankly with such small improvements I'd rather stay on tried-and-true driver releases just in case there are hidden bugs in 12.9. According to their chart, that's less than 1 frame per second difference in BF3 at the settings they used.

It's the first time I've seen nothing go down across the board. Seems like they've got everything clicking with these. I'd like to see crossfire benches to see how they are.

 

[Shmee]

have they fixed poor crossfire performance on ETQW yet?

 

[3DVagabond]

Wouldn't know. I've never even seen it tested. :\

 

[Shmee]

It was mentioned as a win7 issue in 12.8. cant wait for a fix.

 

[VulgarDisplay]

have they fixed poor crossfire performance on ETQW yet?

250fps on a single card isn't enough? You need 500fps?

 

[KompuKare]

OCaholic has done a performance review of the 12.9's on the 7970. Seems like a solid release. Shows small improvement across the board. No obvious glitches or bugs.

That's a handy site. I knew there were major improvements since the release drivers but them having benches with every driver since 11.12 is nice.

Strange that there were different version of 11.12 though. They have beta 12.9 scoring 124.4% vs 11.12 but surely the 11.12RC ('included in the box of 7970') would have been a more useful baseline.

See the synthetics have gone up 148.01% vs 11.12. So it seems AMD are continuing to waste resources on those silly benches - guess with even Intel optimizing so much for synthetics they really have no choice.

 

[Shmee]

250fps on a single card isn't enough? You need 500fps?

I wasn't getting anything near that at my res with these drivers. I suppose I could check my configs and perhaps tweak them, see if it helps.