Asus GTX 670 Listed on Amazon

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Don Karnage

Platinum Member
Oct 11, 2011
2,865
0
0
Never assume overclocking. Worst thing you can do.

Kinda weak numbers for my favorite game

1336900308CAz2bNKppI_3_5.gif
 

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
Kinda weak numbers for my favorite game

1336900308CAz2bNKppI_3_5.gif

At that resolution you need multiple cards for BF3. An Average of ~60fps is just barely cutting it IMO.

Also as has been noted in other threads, HardOCP uses faulty testing methodology and incredibly wacky graphs.
 
Last edited:

Axon

Platinum Member
Sep 25, 2003
2,541
1
76
I was also disappointed in the BF3 performance, but other benches seemed better. I play at 1920 however, so my read is different.
 
Last edited:

Grooveriding

Diamond Member
Dec 25, 2008
9,107
1,260
126
At that resolution you need multiple cards for BF3. An Average of ~60fps is just barely cutting it IMO.

Also as has been noted in other threads, HardOCP uses faulty testing methodology and incredibly wacky graphs.


hehe, man I've heard it all when it comes to [h]. Display something positive about nv; great review, well done chap. Display something positive for AMD; flawed methods, bad review etc. lol

The 7970 is super fast if you put a big overclock on it and brings it to parity in the games it is slower than a 680 in and even faster in the games it is faster than the 680 in (Crysis, Metro) Even if you overclock the 680, as that review shows

It looks like the 670 is so much slower than th 680 and 7970 in that review because the memory couldn't clock over 6.4 Notice the 680 with 7ghz memory speeds keeping up with the 7970.

1336900308CAz2bNKppI_3_5.gif


Kyle claims in that review it is the extra CUDA cores in the 680 helping it, that will be some of it but, GK104 is hungry for memory bandwidth and the 670 is coming up short in that review because of that as well. Especially with him benching 2560x1600.
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
hehe, man I've heard it all when it comes to [h]. Display something positive about nv; great review, well done chap. Display something positive for AMD; flawed methods, bad review etc. lol

The 7970 is super fast if you put a big overclock on it and brings it to parity in the games it is slower than a 680 in and even faster in the games it is faster than the 680 in (Crysis, Metro) Even if you overclock the 680, as that review shows

It looks like the 670 is so much slower than th 680 and 7970 in that review because the memory couldn't clock over 6.4 Notice the 680 with 7ghz memory speeds keeping up with the 7970.

http://www.hardocp.com/images/articles/1336900308CAz2bNKppI_3_5.gif

Kyle claims in that review it is the extra CUDA cores in the 680 helping it, that will be some of it but, GK104 is hungry for memory bandwidth and the 670 is coming up short in that review because of that as well. Especially with him benching 2560x1600.
yeah he should do a clock for clock comparison of the gtx670 and gtx680. he will then see less than 5% difference in most cases as the cards are indeed bandwidth starved. once you get the memory speed up there you will see those extra cores start to scale much better.

I was actually thinking about getting a gigabyte gtx670 but the two reviews I have seen could not get the memory above 6400 either.
 
Last edited:

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
And you claim this or that to prove a point. It's not just because it's showing a good AMD card...I do own a 6950...

The problem is Hardocp thinks things are playable when they are not. 60fps average in BF3 is not that hot...not on MP.
 

Don Karnage

Platinum Member
Oct 11, 2011
2,865
0
0
And you claim this or that to prove a point. It's not just because it's showing a good AMD card...I do own a 6950...

The problem is Hardocp thinks things are playable when they are not. 60fps average in BF3 is not that hot...not on MP.

I have no hiccups at 55fps on ultra at 1920X1080 in a 64 player map. Game is silky smooth
 

mple

Senior member
Oct 10, 2011
278
1
71
I bit on the 680 DC2T model that popped up on Newegg a few hours ago. Hope I don't regret spending the extra $120 for 7% performance
 

Grooveriding

Diamond Member
Dec 25, 2008
9,107
1,260
126
And you claim this or that to prove a point. It's not just because it's showing a good AMD card...I do own a 6950...

The problem is Hardocp thinks things are playable when they are not. 60fps average in BF3 is not that hot...not on MP.

I'm not really arguing that point. For a MP FPS to be comfortable I never want to go below 60fps as a minimum personally.

Just saying that the 7970 is peforming as it is there because that is the reality of the cards performance, same can be said for the 680, 670. [h]'s benchmarks are fine and give good results.

Though I don't agree with his conclusion about the cuda cores being accountable for the large difference between the 670 and 680. It's the 670's poor clocking memory that is slowing it down and 1600P is likely making the memory speed more of a handicap than it would be at 1080P. imo.
 

mple

Senior member
Oct 10, 2011
278
1
71
I'm not really arguing that point. For a MP FPS to be comfortable I never want to go below 60fps as a minimum personally.

Just saying that the 7970 is peforming as it is there because that is the reality of the cards performance, same can be said for the 680, 670. [h]'s benchmarks are fine and give good results.

Though I don't agree with his conclusion about the cuda cores being accountable for the large difference between the 670 and 680. It's the 670's poor clocking memory that is slowing it down and 1600P is likely making the memory speed more of a handicap than it would be at 1080P. imo.

http://www.techpowerup.com/reviews/ASUS/GeForce_GTX_670_Direct_Cu_II/31.html

1890mhz memory OC from the 670 DC2T :O
 

cmdrdredd

Lifer
Dec 12, 2001
27,052
357
126
I'm not really arguing that point. For a MP FPS to be comfortable I never want to go below 60fps as a minimum personally.

Just saying that the 7970 is peforming as it is there because that is the reality of the cards performance, same can be said for the 680, 670. [h]'s benchmarks are fine and give good results.

Though I don't agree with his conclusion about the cuda cores being accountable for the large difference between the 670 and 680. It's the 670's poor clocking memory that is slowing it down and 1600P is likely making the memory speed more of a handicap than it would be at 1080P. imo.

Right. The thing with overclocked reviews though is Video Cards don't overclock like your CPU does. I know people take it into account for their own usage situation and that's fine. I've personally never been able to overclock any video card I own to the levels the review sites got with the same card. So I don't really put much thought into overclocked results.