Following the recent announcement of the HBM2 standard, JEDEC has now officially published details regarding GDDR5X graphics memory standard, which promises significant increase in memory bandwidth.

The main advantage for the new GDDR5X memory is that it will use a very similar memory architecture to the currently available GDDR5 memory, which should make the implementation quite easy. The new GDDR5X doubles the pre-fetch and thus providing double the bandwidth of the GDDR5 memory.

Currently Micron has managed to raise the bandwidth up to 10 to 12Gbps with GDDR5X but future version of the GDDR5X promises to even reach its theoretical maximum of 16Gbps. Bear in mind that GDDR5 memory is limited to 8Gbps.

Micron expects the GDDR5X to be available in the second half of 2016, or just in time for both Nvidia and AMD next-generation graphics cards. With both HBM and upcoming HBM2 ready for implementation and offering significantly higher bandwidth, it is quite likely that GPU manufacturers will use the GDDR5X for its mid-range segment as it would offer easier implementation and lower cost.



Source: JEDEC.org.