Madshrimps Forum Madness

Madshrimps Forum Madness (https://www.madshrimps.be/vbulletin/)
-   WebNews (https://www.madshrimps.be/vbulletin/f22/)
-   -   AMD Radeon HD 6850 and 6870 3DMark Performance Numbers (https://www.madshrimps.be/vbulletin/f22/amd-radeon-hd-6850-6870-3dmark-performance-numbers-75220/)

jmke 9th October 2010 12:35

AMD Radeon HD 6850 and 6870 3DMark Performance Numbers
 
2 Attachment(s)
A lucky person in China got his hands on two of AMD's upcoming video cards, he ran 3DMark06 and 3DMark Vantage using the test setup details below:

CPU:i7 940 2.93G CPU: i7 940 2.93G
Motherboard: X58 (Gigabyte)
DDR3 1066 6GB (7-7-7-19) Memory: DDR3 1066 6GB (7-7-7-19)


The cards tested were:

HD6870
Core frequency: 900
Memory Frequency: 4200
Number of stream processors: 960
3DMark Vantage: P16270 & X7538
3DMark06: 19480

HD6850
Core frequency: 775
Memory Frequency: 4000
Number of stream processors: 800
3DMark Vantage: P14872 & X6549
3DMark06: 18750

if we put these numbers in perspective, the future is looking bright, knowing that HD 6870 is to replace HD 5770 and HD 6850 the HD 5750.


*Obtained on Stefan's Core i7 920@3.2Ghz test setup


- translated source

K404 9th October 2010 13:43

20% increase in SP, ~5% increase in core MHz..... 15% increase in score. It's not looking *that* good.... it implies there's no optimisations going on, just a raw increase in SP.

15% increase in performance is still nice of course, but will probably be 15% more expensive than a 5770 to start with....

AMD back on a red PCB?


Will this be the 2010 equivalent of the 4890?

Stefan Mileschin 9th October 2010 15:40

I wonder what drivers he used for the cards :)

Oberon 9th October 2010 18:31

Remember, AMD is supposed to be going to a 4 shader design instead of the old 4 simple + 1 complex design, so shader numbers can't be directly compared. The best comparison we have here is the 5770 to 6850; you've got the "same" number of shaders and the 6850 comes in at roughly a 10% clockspeed deficit, but scores roughly 10% higher in 3D06. That's a nice little improvement.

*EDIT* Of course, this is just if the testbeds are comparable. Either way, I won't exactly be disappointed if these numbers are correct.


All times are GMT +1. The time now is 15:05.

Powered by vBulletin® - Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO