Information

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



Discussion

Discussion on Article:
Radeon HD 5800 3-way CrossFire vs. GeForce GTX 480 SLI: Final Battle

Started by: Dribble | Date 07/06/10 02:04:46 AM
Comments: 16 | Last Comment:  01/19/12 12:45:15 PM

Expand all threads | Collapse all threads

[1-10]

1. 
You didn't benchmark eye infinity/nvidia surround?

That is the main reason for buying that many cards, and you cannot assume that performance at 25*16 will continue the same way with with eye infinity/surround.

Hence the whole review is a little pointless imo.
0 0 [Posted by: Dribble  | Date: 07/06/10 02:04:46 AM]
Reply
- collapse thread

 
Currently we have only one monitor, while Eyefinity/Surround benchmarks requires three.
0 0 [Posted by: Vader@Xbit  | Date: 07/06/10 05:30:11 AM]
Reply
2. 
In plus of the 3 monitors benchmarks, you also didn't use more than 4xAA...So disappointing. Using 3way crossfire or SLI it's all about the quality. It was rather interesting to see how 3 cards do with 8xAA, 16xAA, 24xAA and even 32xAA for extreme quality. Good thing that there are other sites that provide real testing for such configs...
0 0 [Posted by: TAViX  | Date: 07/06/10 05:56:27 AM]
Reply
- collapse thread

 
it's all about the quality


Especially in Crysis Warhead...
0 0 [Posted by: Vader@Xbit  | Date: 07/06/10 10:23:25 AM]
Reply
3. 
Third card does not scale as well as the second. Seems like the CPU is becoming a bottleneck. Is it possible to test with CPU (i7) north of 4.2GHz; maybe even a 980X @ 4.5GHz.
I'm with the previous posters. $1,000+ graphics subsystem justifies $750 monitor setup (3xAsus 25" at 1920x1200, 2ms). You need to test higher res./IQ. Crysis might be the exception but there are plenty of games that ran 100+fps @4xAA. Mind you Crysis also CPU bottlenecks. (If you do not believe me OC the CPU and rerun it.) The game is a resource hog.
0 0 [Posted by: jonup  | Date: 07/06/10 11:16:48 AM]
Reply
- collapse thread

 
Is it possible to test with CPU (i7) north of 4.2GHz; maybe even a 980X @ 4.5GHz


At this time, no. Later maybe - with 980X or more powerful CPU, but without overclocking. As for FSAA 8x - we are considering this.
0 0 [Posted by: Vader@Xbit  | Date: 07/06/10 02:58:32 PM]
Reply
4. 
Thanks for the swift and honest reply! What's the problem with OCing? 3-and-more-GPU setups become CPU bottlenecked, just too much overhead to be handled. Try it offline and if the performance difference is significant enough reconcider your decission.
0 0 [Posted by: jonup  | Date: 07/06/10 03:32:31 PM]
Reply
- collapse thread

 
Anyway, we'll have to wait for new CPU for our main testbed.
0 0 [Posted by: Vader@Xbit  | Date: 07/07/10 12:01:27 PM]
Reply
5. 
Eeerrmmmm...hi guys, i know it's probably very simple and this question sounds rediculous, but I just can't understand your "summary diagrams". They just look like a whole lot of red and green stripes to the left or right of a line.
To coin a phrase:-
Can you explain it to me like I was a 6 year old?
0 0 [Posted by: SgtBilko  | Date: 07/06/10 10:32:48 PM]
Reply
- collapse thread

 
It is simple. Look, green bar means "3-way CrossFire vs. GeForce GTX 480 SLI", so, 43.2 in Just Cause 2, for example, means, that 3-way CrossFire 43.2 % faster, than GeForce GTX 480 SLI in that game. Same principle for other games and cards.
0 0 [Posted by: Vader@Xbit  | Date: 07/07/10 08:59:26 AM]
Reply
 
The summary diagrams show you the percent of difference between the main card tested and its opponents. ( The main card is named in bold at the top )

There are usually 4 summaries, 1 for each resolution. If the bar is going backwards ( negative area of a graph) then the card is loosing. If the card is winning then bars are going forward towards the positive area of the graph.

The summaries allow you to quickly see how many / and how large of wins the card has at a resolution your interested in.

If your ever confused assume the ati is always loosing at HAWX , allows you to quickly figure out which way the bars are going lol
0 0 [Posted by: cashkennedy  | Date: 07/08/10 10:22:09 AM]
Reply
6. 
Asus Ares Ati 5970 in CF Crysis ~ 50 Fps at 2560x1600 4xAA http://www.techpowerup.co...SUS/ARES_CrossFire/7.html better than Nvidia GTX 480 in SLI .(34 fps)
0 0 [Posted by: Blackcode  | Date: 07/08/10 03:22:05 AM]
Reply
7. 
the CPU should be clocked at 3.8ghz (easy to do with just base clock and muti alone) you may have been CPU limited at some points or a lot of points

6 core 980x is not needed But can OC more then an 930 based cpu but i use 3.8 as 99% of users can do 3.8 with just 2-4 setting changes, base clock 191x20 (if its not asus motherboard you may need to set the votages as well), disable tubo and power management stuff
0 0 [Posted by: leexgx  | Date: 07/08/10 07:25:00 AM]
Reply
8. 
"The results of the multi-GPU subsystems are approximations calculated by adding up the power consumption of their constituents in 3D mode. It is clear that the three AMD RV870 processors are going to be more economical than the pair of Nvidia GF100 chips."

Power Consumption diagram does not show values for 5970 that is constituent of tested configuration. The same for Noise Level diagram.
0 0 [Posted by: KonradK  | Date: 07/19/10 10:26:20 AM]
Reply
9. 
When are you guys going to do an Eyefinity vs Surround review? Single monitors are NOT what these setups are for!
0 0 [Posted by: Heatlesssun  | Date: 07/26/10 09:15:47 PM]
Reply
10. 
I have this 5870 however I am trying to find replacement fans for it as they rattle. They are 80mm Everflow T128010SH fans however I couldn't find them anywhere. Any ideas where I can get a pair? Also I couldn't find any reference for the Everflow company so I couldn't contact them as I couldn't find them. Some help would be appreciated.

Thanks!
0 0 [Posted by: Frenchyaz  | Date: 01/19/12 12:45:15 PM]
Reply

[1-10]

Back to the Article

Add your Comment


Our Sponsors

Crucial.com Western Digital G.Skill