Wednesday, May 22nd 2024
PC DDR6 Memory to Offer 10-times the Bandwidth of DDR4: Synopsys
The next-generation PC DDR6 memory standard (not to be confused with GDDR6), will offer a 10-times increase in bandwidth over DDR4, according to a presentation by Synopsys, a major vendor of memory controller and PHY IP blocks. The initial draft of DDR6 specification by JEDEC is expected to be ready within 2024, with version 1.0 of the spec ready by mid-2025. Speeds (data-rates) of DDR6 start at DDR6-8800, and range up to DDR6-17600 in the first generation; with future generations of DDR6 going all the way up to DDR6-21333 (or 21 Gbps). This is exactly 10 times the bandwidth of DDR4-2133, the initial speed of DDR4 that debuted with 6th Gen Core "Skylake" processors, almost a decade ago. It hence makes sense for a memory specification 10 years since to offer such a linear scaling in bandwidth.
Synopsys also talks about LPDDR6 in this presentation, the future low power memory standard for thin-and-light computing devices and smartphones. LPDDR6 will have an introductory data-rate of LPDDR6-10667 over a 24-bit memory channel, with two 12-bit sub-channels. The highest defined data-rate for LPDDR6 is expected to be LPDDR6-14400 (likely 14466 MT/s). Besides generational increases in bandwidth, both PC DDR6 and LPDDR6 are expected to introduce several security and energy-efficiency features, including an "efficiency mode" that reduces idle power draw for the memory devices.
Source:
TweakTown
Synopsys also talks about LPDDR6 in this presentation, the future low power memory standard for thin-and-light computing devices and smartphones. LPDDR6 will have an introductory data-rate of LPDDR6-10667 over a 24-bit memory channel, with two 12-bit sub-channels. The highest defined data-rate for LPDDR6 is expected to be LPDDR6-14400 (likely 14466 MT/s). Besides generational increases in bandwidth, both PC DDR6 and LPDDR6 are expected to introduce several security and energy-efficiency features, including an "efficiency mode" that reduces idle power draw for the memory devices.
31 Comments on PC DDR6 Memory to Offer 10-times the Bandwidth of DDR4: Synopsys
Recall how fast DDR5 was supposed to go, by JEDEC specification. I don't remember seeing the faster numbers often, outside some server and workstation modules in non-XMP/EXPO context.
The performance for the baseline spec for each RAM generation usually doubles and this the case here as well - The DDR6 baseline of DDR6-8800 is a bit over 4 times faster than DDR4-2133 and roughly twice as fast as DDR5-4000.
Bulldozer and its two follow-ups, the APU versions wanted 1866 and 2133. Having 34GB/s of video memory (although shared with the CPU) was a pretty big boost to performance compared to Intel's IGPs typically 1066/1333/1600. None of their CPUs officially supported 2133 DDR3, but every chip going back to Sandy Bridge supported it.
I still use my 2500K @ 4.7GHz and 1.38V with 32GB DDR3 2133 running 10-12-11 CL1. It's four dual rank sticks, so not 100% ideal, but pretty close! 9.37ns latency :cool:
Anyway, yeah - so while DDR4 did debut at 2133, DDR3 already existed at 2133, and it wasn't an overclocked variant. OC DDR3 actually went all the way through and past DDR4 spec - into 4's overclocking range! I remember seeing a DDR3 3400 kit, 200 faster than DDR4's fastest speed of 3200!
Even today, the best seller is still DDR4-3200, and that won't change anytime soon, despite DDR5 or distant future DDR6.
www.amazon.com/Best-Sellers-Computer-Memory/zgbs/pc/172500
DDR3-2133 vs DDR6-8000
Assuming 64-bit memory bus:
8000MT/s is 8 000 * 64 = 512 000 Mb/s = 64 000 MB/s = 62.5 GB/s
2133MT/s is 2 133 * 64 = 136 512 Mb/s = 17 064 MB/s = 16.66 GB/s
DDR3 is not inherently dual-channel. Memory modules have been 64-bit for a while and having multiple channels is a memory controller thing.
IMCs on desktop tend to be dual-channel today but same modules are used for other configurations, usually in servers. For example DDR3 was used in quad-channel on X79 - Sandy-E/Ivy-E Xeons.
Latest Epycs can do 12-channel DDR5 if I remember correctly.
Sandy Bridge was speculated to be spec'd to run at up to 2133 in 09/10, but I think when Intel realised in that none of their CPUs until DDR4 would be pushing much past 2133, they decided to artificially limit Sandy to dual channel 1333 (so Ivy could be 1600 and Haswell/Broadwell could be 1866/2133. Broadwell on the crappy node shrink might have been why its max official supported speed remained at 1866 - idk for sure.
Also ... Does LPDDR with ECC even exist?Yeah but what flavour of "G" is that? When describing data rates, mega/giga/tera is always, universally, decimal (powers of 1000), not binary (powers of 1024). Your M here is 1,000,000 but your G is 1,024,000,000.