Biggest, fastest: Leapfrogging in Samsung HBM2 • HWzone
Computersgraphic cards

The Biggest, Fastest: Leapfrogging in the Samsung HBM2 Memories

The Korean giant is taking advantage of its lead in the chip world to announce mass production for ever more agile graphic memories

Get updates from everyone in TelgramGet updates from us all at TelgramJoin the channel now

A few weeks after we began to fantasize about the improvements they would bring Memories of the GDDR6, With effective 450 GB bandwidths per second for mid-market tickets and almost 1,000 GB per second at the seemingly high end - comes as a reminder to remind us that the entire GDDR world is still a remnant of the past, trying to get the pace of The new star in town.

The new star is called HBM, or HBM2 to be precise, and has become more relevant than ever at the CES 2018 exhibition with the announcement of The Kaby Lake-G processors Which include cores Structured and advanced memories of this kind, as well as announcing processors Graphics for the Mobile Market - When it is not exaggerated to declare that much of the graphics cards for the high market and possibly even the medium market will soon contain this device, which can already reach up to a monstrous bandwidth of terabytes per second.

Prepare - chips Faster in the doorway

Samsung is not satisfied with the latest state of the art, and announces the start of mass production for a new generation of HBM2 chips, which bring even more impressive 2.4 gigabit per second performance per pin in the ultra-wide 1,024 bit interface of this standard (compared with 2 max) Gigabit per second to date), under a modest basic workload of 1.2 volts and with an increased storage density that allows for a total volume of 8 gigabytes per chassis . The end result of all this is the effective bandwidth of 307 gigabytes per second per package, when we already know that it is possible to combine two to four Ones with a single processing core - which means we can get graphical processing cores with volume soon Dedicated 32 gigabytes, and an impressive bandwidth of over 1.2 terabytes per second.

The GDDR6 chips have not yet arrived, Are already clarifying that they will remain the obvious choice for the most advanced end products on the market

All of this is certainly impressive, though it is worth noting that this is just an interim development, on the way to the arrival of the HBM3 memoir generation - perhaps as early as next year. These memories are expected to be another huge leap forward, with support for larger volumes up to 64 gigabytes (!), A "much lower" working voltage than HBM1.2 generation 2 volts, and essentially doubling the effective bandwidth for all Such a memory structure for 512 gigabytes per second, 2 times the base of HBM2 and 66 percent more than the innovative memories of . With an array of four buildings On the chip we will reach an imaginary bandwidth of 2 terabytes in a second (!), And if we look at arrays of six or eight structures this figure can reach 4 terabytes in a second - crazy.

there is also On the horizon, which makes the device even more flexible and efficient - and fits a much wider range of applications than the GDDR

In other words - there is still a lot more to strive for, and this race is between To- And Mikron is not supposed to show signs of slowing down.


תגובה אחת

    It seems to me (the price) more for calculations and work…
    Games to take advantage of HBM3?

Leave a Reply

Back to top button