Ive Been An Experianced Hardware Expert For About 19 Years But There Are A Few Things I Am Not Sure About....
I Bought A GTX460SE At An Excellent Price ($129.00) It Is Identical To A GTX460 1GB In Everything But The Amount Of Shader Processors (288 vs. 384) Even The Stock Clocks And The RAM Type And Bus Are The Same (256-Bit DDR5) And It Comes Within 5% Or So In Performance In Realworld Processing Loads. However.....
I Am Upgrading My Main Panel From A Dell 2001FP To A Dell 2408FPW (From 1600x1200 To 1920x1200) Witch Im Sure Will Not Require Me To Run My Games At Any Lower Settings Than I Did At 1600x1200 But Id Still Like Some Extra Performance To Help Take Up Any Slack And I Doubt A New Card And Any Unessesary Spending Will Have Any Tangible Benefit, Thus The Following......
Im Currently Overclocked To 800MHz GPU/1600MHz Shaders/900MHz RAM And After Hours Of Gaming Or Intensive GPGPU (CUDA/OpenCL Etc.) Processong Shows No Instability And Rarely Gets Over 50C So I Know It Can Go Even Farther, So I Must Ask Out Of The Three Variables (Core/Shader/Memory) Witch Of Them Has The Most Impact On Overall Shader Performance?
This Will Help Me In The End Get The Most Out Of My GPU, Thank You.
Assuming it overclocks the same I run my EVGA GTX 460 1GB at 820-1640-2000. It will go higher but that's fine for what I do with it.
Since you are a hardware "expert" you should know that the core and shaders are linked and overclocking them nets a much higher benefit than overclocking the memory. You should also know that no two overclocks are the same, and if you want to find the limits of your card it takes trial and error just like with any overclock.
So I Must Ask Out Of The Three Variables (Core/Shader/Memory) Witch Of Them Has The Most Impact On Overall Shader Performance?
Makes no sense? The shaders obviously have the most impact on shader performance.
I have seen some people get as high as 900-1800-2000 with a GTX 460 1GB. Pretty good for a card that is 675-1350-1800 stock. You are just going to have to overclock, test and repeat. You won't know until you try.