jmke

For the DDRdrive to perform as well as it does with only one lane of PCI-Express, it must employ extremely low latency tech. There are as few components as possible between the PCI bus and the RAM, and those parts are as speed optimized as possible. With such a fast turnaround time, we found that unleashing the true potential of the DDRdrive required moving to a board without a beefy video card. Newer video cards like the Geforce GTX series tend to assume they own the PCI-E bus, meaning other devices have to repeatedly arbitrate to communicate with the CPU. The added latency of waiting for a hefty video card to release the bus adds up, leading to an unnecessary bottleneck at extremely high IOPS. Given the target market for the X1 is the enterprise sector, it is not unreasonable to assume these units would be put to use in displayless server environments.


We reached over 280k IOPS with IOMeter. Not quite 300k, but given the added overhead by IOMeter itself, it is safe to say the control panel application was being truthful. The DDRdrive really shines in single sector random access, literally blowing everything else out of the water. At such insanely high IOPS, the CPU quickly becomes the bottleneck. For reads, the dual X1 setup gets slightly lower values due to the pair being installed in PCI-E 1x slots as opposed to the single unit installed in the PCI-E x16 slot. x16 is not significant, but the lower latency link to the CPU is. For writes, the extra bandwidth helps the dual setup achieve higher IOPS, but is still ultimately bottlenecked by the CPU.

The DDRdrive X1 (Limited capacity of 4GB) is shipping as of today for a cost of $1495.

http://www.pcper.com/article.php?aid...e=expert&pid=1

__________________