Hier een review van dit schatje Semiconductor Technology is fascinating. An engineer will come up with a new technology that allows for a significant increase in performance over existing technology, as with the NVidia GeForce. As the new platform matures, the speed and throughput of given device is realized and the limitations are exposed. At that point, another brilliant engineer sets out to improve again on the initial design in an effort to alleviate the bottlenecks and achieve yet an even higher level of performance. As we all know, this process happens at an alarming rate in the Computer and Networking Industries as well as Semiconductors.
Local memory bandwidth was an initial bottleneck for NVidia GeForce chipset based 3D Graphics cards. It was a bottleneck that was envisioned early on by NVidia however. They designed the initial GeForce architecture to be capable of interfacing to a new high performance SDRAM and SGRAM with double the data rate of current standard DRAM architectures. DDR SDRAM/SGRAM, or Double Data Rate Synch DRAM, allows any device interfacing to it to perform reads and writes to the DRAM on the rising and falling edge of the memory clock cycle. With each clock cycle, you can perform both a read from and a write to the memory. As a result the effective throughput of the DDR SDRAM is doubled. Hence the term, "double data rate".
Vind het hier
(ik blijf origineel in dit regeltje)