Show description

Intel and AMD have their own systems, but both work in the same way – the socket holds the processor, with the chip clamped into position. That keeps the silicon secure, and it means a cooler can be attached over the top. Intel processors lay on top of a flat surface, and the processor talks to the mobo through 30 golden pads that link to the socket. The chips have notches that line up with the socket so installation has to be correctly orientated. AMD chips are trickier. They interface with the motherboard using hundreds of tiny metal pins that line up with depressions on the socket.

Larger heatsinks are sometimes too long to fit inside smaller cases, and double-width coolers can also block motherboard slots below the card. That’s not the only physical consideration. More powerful cards require more electricity, with some top chips needing two eight-pin power plugs to function. Most power supplies come with these connections, but it’s worth checking in advance. Every graphics card has a variety of outputs that connect to monitors. 1 may one day supercede these specific formats; the sheer bandwidth available makes it a very strong option.

Nvidia’s talking about NVLink offering DRAM-class speed and latency, which will allow for the use of Pascal’s unified memory across the entire PC. It will also improve performance between GPUs, so multi-GPU systems could end up getting a far more linear scaling in terms of gaming speed. As well as the NVLink connection on the GPU itself, it will also require dedicated silicon in the CPU if it wants to bypass the PCIe interface completely. From the outset though, that looks likely to be restricted to supercomputer-class high performance computing (HPC); Intel is unlikely to start dropping Nvidia silicon into its designs.