In answer to your questions...
PCI-E card for PCI-E slot...
The cards memory (vRAM) makes no difference for compatibility, it can have DDR2, GDDR3, 4 or 5. "hd" isn't a memory standard...
Depending on how your third question is interpreted;
(a) PCIe x16 (2.0) has a theoretical bandwidth twice that of 1.0, 500mb/sec as opposed to 250mb/sec -multiply each by the 16 PCI lanes (hence PCIe
x16) to get the theoretical bandwidth of the slot (8Gb/sec vs 4Gb/sec). Data transfer rate and base clock frequency are also doubled.In reality there is little difference between the two since most of the graphics functions are handled within the card itself. Slot bandwidth only comes into play where the graphics card and CPU need to communicate in an intensive manner.
(b) A PCIe x16, as shown above, offers a much greater bandwidth than does a single lane PCIe x1. Such a large reduction will impact graphics performance in a great amount of gaming scenarios. See
this article for a comparison of real world PCIe bandwidth (The article is based on the then current 1.0 specification)....
Always place the graphics card into the primary PCIe x16 slot if you have two or more available. This slot is invariably the PCIe x16 slot closest to the chipset/CPU.
.