View Single Post
Old 12-24-11, 08:41 PM   #73
shadow001
Registered User
 
Join Date: Jul 2003
Posts: 1,526
Default Re: Official Radeon 7970 Review and discussion thread

Quote:
Originally Posted by i SPY View Post
And its cooler compared to 40nm


7970 Thermal test
http://www.hardware.fr/articles/848-...nfrarouge.html






Now if it were 1125mhz by default then it would own really bad

But still 200mhz OC with default voltage is amazing imo.

My GTX580's are running at 950 Mhz clock, but it is with a voltage bump and water cooled, not the stock cooler ...There probably wouldn't be a huge difference in performance with the HD7970 if it's running at stock clocks now that i think about it..


That actually brings up a point about architecture efficiency overall, since both my GPU's(being water cooled) and the Tahiti are running at roughly the same clock, they both use a 384 bit memory bus, though tahiti's memory is rated for higher speeds and has higher memory bandwith, but the performance gap seen compared to a stock GTX580 would be smaller(they run at 772 Mhz for the core afterall), and keep in mind that Tahiti has a billion or so extra transistors in it's core over fermi.


I fully realise that the HD7970's are on their first offcial driver release and the card will get faster by the time Kepler releases, and the drivers for the GTX580 are mature and extracting pretty much all the speed the core has by now(so it's a large advantage there), but it makes me wonder that Nvidia wouldn't need to pull a miracle to make something faster than an HD7970 either....


A higher clocked Fermi, with the same amount of shaders, same transistor budget and same 384 bit bus but running with faster GDDR5, clocked at 1 Ghz core using the 28nm process would be a pain in the ass for the HD7970, never mind adding more shaders, texture units, rops or moving to a wider bus.....
shadow001 is offline   Reply With Quote