I am wondering how old this document is? The paper claims to talk about the state of the art of FPGAs. However, it does not mention PICe 3.0 at all. It talks about FPGAS that attach directly to the CPU socket or to the DIMM slot on the motherboard (see the last section for the DIMM slot attachment). It does not talk about PCIe 3.0 . This make me think the paper was written before that that standard arrived. The document in question can be found here. http://www.scientific-computing.com/features/feature.php?feature_id=241 Any help appreaciated. Thanks in advance. Respectfully, Newport_j
Newport_j wrote: > I am wondering how old this document is? The paper claims to talk about > the state of the art of FPGAs. However, it does not mention PICe 3.0 at > all. Right under the abstract: Scientific Computing World: June/July 2009
Please excuse. You are right. I found this site at the last minute when typing this message. The document that I originally printed out had no date on it. This one does. Duh. Thanks! But stilll there is little mention of PCIe and a lot about FPGAs plugging into motherboards and into DIMM slots. It seems to me to be unecessary, if you have PCIe 3.0. Any thoughts. Respectfully, Newport_j
This paper is not about the state of the art, it is about promoting FSB solutions. For example they are comparing Memory Latencies with old PCI latencies, at a time when PCI was already in the prozess to be replaced with PCI Express. They also never talked about transfer rates, only boosting about "zero copy architecture". 2009 PCIe 3.0 was still in specificaton phase, only now you can get FPGA with hardcore support for it.