Information

Did you know you could become an X-bit labs fan on Facebook or follow us on Twitter?



Discussion

Discussion on Article:
3DMark 2013 Graphics Cards Performance Review: AMD and Nvidia Based Products in the New Benchmark

Started by: beck2448 | Date 02/27/13 06:41:28 AM
Comments: 22 | Last Comment:  04/30/13 03:18:00 PM

Expand all threads | Collapse all threads

[1-13]

1. 
Crossfire has a serious problem. Until it's fixed their fps numbers cannot be taken at face value. http://techreport.com/blo...ptures-crossfire-and-more
1 2 [Posted by: beck2448  | Date: 02/27/13 06:41:28 AM]
Reply
2. 
These tests back up what we know about the competing GPU stacks:

1. AMD needs at least one sku between 7770 and 7850, if not two splitting the difference.

2. AMD needs a part similar to 680 (or the original 925/5500 7970) that uses less than 225w.

Interesting the powertune ratings of 7770 and 7850 are 80w and 130w. 7870 175w, and 7950 is 200 or 225; depending on it you count the original version or those with boost.

It's almost like there are VERY APPARENT holes (one or two between 80-130w/one at 150w/one at 200w or 225w) waiting to be filled. I bet AMD is completely clueless to this fact...and will in no possible way fill them any time next quarter.

Because the 7000 series is stable, you see...
1 0 [Posted by: turtle  | Date: 02/27/13 07:43:47 AM]
Reply
- collapse thread

 
something called 7790 is supposedly coming in, like, 10 days, priced ~130eur. it's even supposed to sport GCN2 architecture - i didn't pay much attention on this rumour, until sale started to list it as 'available' with a pricetag
0 0 [Posted by: snakefist  | Date: 03/09/13 01:59:19 AM]
Reply
3. 
The reason the 7770 and the 7750 did so bad is because you tested the 1gb version (against a 2gb 650ti) when running benchmarks at such a high resolution. Its pointless having a 1gb card in this test.
3 2 [Posted by: azkor00  | Date: 02/27/13 08:03:07 AM]
Reply
- collapse thread

 
No, 650Ti is just faster card than 7770 and especially 7750.
1 2 [Posted by: rrr  | Date: 02/27/13 08:14:12 AM]
Reply
4. 
show the post
1 6 [Posted by: tbaracu  | Date: 02/27/13 09:05:45 AM]
Reply
5. 

I know for sure AMDis stronger in Fire-Strike, and suddenly comes the site xbitch-labs to say another thing... putting Nvidia in front lol.
2 4 [Posted by: tbaracu  | Date: 02/27/13 09:08:13 AM]
Reply
- collapse thread

 
Here's something new! You got to be the first in MONTHS to say that Xbit favors Nvidia. How does it feel all alone on an Island? They're just numbers acquired by running an app to observe hardware AND DRIVERS performance.
0 0 [Posted by: MHudon  | Date: 03/27/13 08:09:12 AM]
Reply
6. 
The graphics card performance ratings have always been unreliable because the benches are poorly designed and often favor one GPU design over another. The GPU makers figured out how to inflate the benches years ago and they all do it. Only a fool would take the numbers as being accurate, but there are a lot of naive PC enthusiasts.
4 0 [Posted by: beenthere  | Date: 02/27/13 09:33:06 PM]
Reply
7. 
I think the top AMD cards results are a little low...Just compare it with similar results from other sites if you don't believe me. They even used cheaper CPU's to test.
Just saying...
2 0 [Posted by: TAViX  | Date: 02/28/13 12:33:26 AM]
Reply
- collapse thread

 
show the post
0 4 [Posted by: Jordan  | Date: 02/28/13 04:14:05 AM]
Reply
 
I don't think this is done by purpose, but the default settings in drivers you use for AMD cards are NOT the same as the ones in nvidia drivers. You are doing this for years now, so is no surprise here.
2 0 [Posted by: TAViX  | Date: 03/06/13 08:53:35 AM]
Reply
 
Here's a good point. Most users actually keep those default settings. Comparing the numbers you get from products that are strait out of the box shows exactly what most buyers will experience. This makes the test pertinent for all but some enthusiast. Still very interesting don't you think?
0 0 [Posted by: MHudon  | Date: 03/27/13 08:13:28 AM]
Reply
8. 
show the post
0 3 [Posted by: nerbne  | Date: 03/03/13 07:00:03 AM]
Reply
9. 
After testing with 3DMark 2013 I find the same issue that has existed all along - that the benchmark is unreliable in that it reports as much as a +/- 3% difference with no changes to the system. That's a poor benchmark by any standards.
2 0 [Posted by: beenthere  | Date: 03/04/13 07:20:21 AM]
Reply
10. 
actually this sites review of the amd vs nvidia is correct... only firestrike extreme "removes" the Influence of the cpu why?... i have a 8350 cpu, when i compare the scores on this web site to my own, the physics score of this web site test platform is around 16500, whereas for the fx8350 about 8900. thus the xblitlabs platform has a faster cpu, but it does not affect the firestorm extreme benchmark.

which means my score is higher than xbitlabs on firestrike extreme because my gpu (zotac 680 amp!) (1240/7000) is faster... BUT only on that benchmark because the other benchmarks (cloud gate, icestorm ) the cpu affects the score.

my scores for the benchmarks are : (i am using the asus for comparasion because it is closest to my card... you will notice that all of the benchmarks except for firestrike extreme are faster than my setup (fx8350, zotac 680 amp!)
AMD icestorm xbitlabs :amd 7970 (166983) my 7970 (115926) http://www.3dmark.com/is/275520
NVIDIA icestorm xbitlabs gtx680 (165605) my gtx680 (124476)
http://www.3dmark.com/is/269064

AMD firestrike extreme xbitlabs :amd 7970 (3558) my 7970 (3592) http://www.3dmark.com/is/256103
NVIDIA firestrike extreme xbitlabs gtx680 (3363) my gtx680 (3588)
http://www.3dmark.com/is/247884

the point I am making is not that one is faster than the other but that the firestrike extreme compares graphics cards only (cpu does not affect the score) ... but does it mean anything?, well it is a benchmark...
1 0 [Posted by: haar  | Date: 03/10/13 10:34:48 AM]
Reply
- collapse thread

 
"but does it mean anything?, well it is a benchmark..." I totally agree!!! These are just numbers and there's a lot of way to interpret them an that's all there is to it! Personally, I like those benchmark to quantify the improvement made by overclocking a product. I prefer looking at the big picture shown by many different results and whose better than Xbitlabs.com to provide these?
0 0 [Posted by: MHudon  | Date: 03/27/13 08:22:50 AM]
Reply
11. 
Hey Xbit, when you getting TITAN?!? I'm losing my patience!!! :D
1 0 [Posted by: Bo_Fox  | Date: 03/10/13 11:47:12 AM]
Reply
- collapse thread

 
Titan is only "worth it" if you buy 2 (since a 690 would be the smarter single card choice), and there aren't going to be many people with $2k to spend on graphics. Really rather see them focus time testing other things.
0 0 [Posted by: AnonymousGuy  | Date: 03/12/13 11:54:14 AM]
Reply
 
Send them a sample, I'm sure the'll gladly test it just for you
0 0 [Posted by: MHudon  | Date: 03/27/13 08:24:45 AM]
Reply
12. 
Run two different video cards one from AMD one from Nvidia by installing the AMD in the first slot and resetting the CMOS

Source: http://fbappointments.com...md-and-nvidia-video-cards
0 0 [Posted by: Randy Appointment  | Date: 04/30/13 03:17:57 PM]
Reply
13. 
Run two different video cards one from AMD one from Nvidia by installing the AMD in the first slot and resetting the CMOS

Source: http://fbappointments.com...md-and-nvidia-video-cards
0 0 [Posted by: Randy Appointment  | Date: 04/30/13 03:18:00 PM]
Reply

[1-13]

Back to the Article

Add your Comment


Our Sponsors

Crucial.com Western Digital G.Skill