PCI-E 2.0 doubled effective throughput from 250GB/s to 500GB/s from PCI-E 1.0. For the most part, graphics cards today barely use up x8 bandwidth in a 2.0 slot, so in a full x16 1.0 slot, I doubt you'll see much detriment to the card performance. You're more like to see CPU bottleneck before you see PCI-E bottleneck.
It will be like running Crossfire in an x8 x8 setup. You'll probably see a slight bottleneck, a performance loss of 3-5% at most with a 1.0 slot , assuming the CPU isn't the bottleneck. In any case, the card will work well, even in a 1.0 slot.
If you have a PCIe 1.0 MoBo, I'm more afraid of the whole system being a "bottleneck" than the PCIe alone.
PCIe 1.0 means you'll be working with gen1 PCIe MoBo's hardware, which is quite slow by today's needs: nForce4, Intel 9xx and AMD5xx and 6xx series chipsets and a CPU that's very slow (most probably).
If you have a PCIe 1.0 MoBo, I'm more afraid of the whole system being a "bottleneck" than the PCIe alone.
PCIe 1.0 means you'll be working with gen1 PCIe MoBo's hardware, which is quite slow by today's needs: nForce4, Intel 9xx and AMD5xx and 6xx series chipsets and a CPU that's very slow (most probably).
Cheers!
That was what I was thinking.
Also, if you mean a 6990 (I'm not aware of a 5990), the PCEi 1.0 x16 may actually be a problem as well. It's a dual gpu card on a single card. That means 2 GPU's are sharing the bandwidth of a x8 slot.