CPU & GPU Binning

[mri5]

Can someone please tell me:

1. About CPU and GPU Binning?

2. Size of wafers?

3. How they test them?

4. How they bin and how they cut wafers etc?

All four of amd nvidia and intel please.

[Activate: AMD]

I can tell you what I know generally, but I'm not sure about all of them

1) CPU and GPU binning refers to the process of testing processors to determine which ones run better than others, then sorting them accordingly. Certain chips will be inherently capable of running at higher clock speeds at the same voltage, or lower voltages at the same clock speed. Binning is basically how manufacturers can pick the chips that run the fastest and sell them in the highest-end gear, while the slower ones get sold for less. Related, but not really "binning", is die harvesting. Chips in multi-core designs might have defects in certain parts of the chip (a core in a CPU or shader block in a GPU). These parts of the chip are then disabled and the resulting fully-functional part is sold as a lower-end part. Think GTX 670 vs GTX 680 or AMD's old triple-core CPU's, these were cases where a shader block or core was disabled. In short, its a way for manufacturers to maximize yields and create halo products by cherry picking dies and running them at speeds that would produce terrible yields if all chips were asked to run them.

2) 300mm is the standard right now as far as I know. 450mm is under development (last I heard)

3) I'm not entirely sure.

4) As far as how they bin, I'm sure its related to how they test because, as explained in 1), binning is just sorting out the chips that do the best in testing. Cutting wafers is usually done with a saw or possibly a laser

[mri5]

Thankyou mate @Activate: AMD