kirock said:
OK, I'm not an expert on this but hopefully I get most of this right.
PCI was the original bus protocol and was used for most periphrals including vid cards until around mid 90's. Then the AGPX2 came out, it had a different voltage level from the PCI bus and I think it transferred info on both the up and the down clock cycle. (Thus the X2 part). Then the AGP bus took over and we had an evolution of X4 and X8 up to today.
But meanwhile, back at the PC labs, they worked out a way to speed up the original PCI bus (with its much simpler protocol) and BAM, PCI-E X16 hit the mobo market. The X16 means it's 16 times faster then the original PCI, this means it's X2 faster then the AGPX8 bus. Whew, still with me?
Now this is the part where I'm not really sure of the details. There is a change with the PCI-E X16 protocol from the original. I think it's voltage for sure but also other stuff to do with the clock transitions again , but not as complicated as the AGP bus. SO (stay with me), there is a PCI-E X1 bus which will soon be the replacement of all PCI slots. This new bus is the same protocol as the PCI-EX16 but without the expensive ICs(chips) needed to move it at X16, and this make things sooooooo much easier/cheaper to build a mobo! The article I read some months back basically said the new PCI-E X1 will replace all PCI slots within the next few years.
With that being said, I wouldn't buy a PCI-E X1 video card, it's still 16 times slower then the X16 slot and 8 times slower then the AGPX8.
Almost there.
Let's start from the beginning. On the First mobo type, IBM placed ISA. On the Second mobo type, Intel placed PCI. Then Intel improved the PCI, and called it PCI-X. However, PCI-X was still slower than AGP, so a new Bus was created: PCI-Express.
PCI-Express is NOT a faster kind of PCI Bus, just like PCI was NOT a faster kind of the ISA Bus. PCI-Express is a completely new kind of Bus, and although it is somtimes made compatible with AGP (nVidia makes chipsets that allow AGP cards to work simultaneously with PCI-Express), it has a completely differernt protocol.
The Protocol:
PCI-Express is compromised of "lanes" that carry the data back and forth, and allow components to talk at a rate of 500MBps full-duplex (250MBps one way). The "x1" after a PCI-Express slot lets you know that the slot "owns" only one lane; the "x16" slot owns all 16 lanes. This is why graphics cards are not run in the "x1" lane, because 16 lanes are much better at full-duplex than 1 lane.
If you have questions about this stuff, and want more info, check out
www.answers.com and
www.wikipedia.org
By the way, you will never hear of PCI-X because its technology has been incorporated into the current PCI, and the difference is so marginal that it's not worth it to advertise a mobo as PCI-X compatible and confuse the masses.
Also, a PCI-Express x32 is coming out soon. Watch for it, as it will allow you to run graphics cards as fast as you do with SLI, except you will only use one slot... Just imagine, using SLI with two graphics cards with 2 PCI-Express x32 slots... WOW