Intel’s Plans for 3DXP DIMMs Emerge • Hacker News Board

Re: Intel’s Plans for 3DXP DIMMs Emerge

#71

Post by amluto » Tue, Jul 24, 2018, 7:50 PM UTC

caf wrote: post #28

jzelinskie wrote: post #20

It seems obvious in retrospect, but persistent memory adds a pretty exciting new advantage for persistent data structures. Another thought: As potentially paradigm changing technology like this becomes available will it ever make sense to redesign the OS?

I used to think so - once you have persistent primary memory, you can install operating systems and applications directly into primary memory, so there's no longer any point in having a distinction between "install" and "run/boot". However, iOS and Android have shown that it's possible to do away with this distinction even with a traditional OS running underneath. So I now tend to think that instead what will happen…

> so there's no longer any point in having a distinction between "install" and "run/boot".

When Linux boots, the in memory state changes quite a bit. Even the actual code gets modified during boot. The whole process takes well under a second. Linux does support an “execute in place”, but it’s barely a win, and I don’t think it works on x86.

A more interesting idea is to put your OS installation on a DAX (direct access) filesystem.

Re: Intel’s Plans for 3DXP DIMMs Emerge

#72

Post by wtallis » Tue, Jul 24, 2018, 7:56 PM UTC

grogers wrote: post #64

One interesting thing for databases is that as nonvolatile storage latency decreases, traditional btrees get more attractive relative to newer log-structured designs. Especially if the write endurance is increased as well over current SSDs.

Or to put it another way: there's not a lot of reason to have two layers of log-structured storage. Your SSD already needs its own log-structured flash translation layer, and if that's tuned properly for your database workload, then another layer of the same kind of thing may not help much.

Re: Intel’s Plans for 3DXP DIMMs Emerge

#73

Post by AtlasBarfed » Tue, Jul 24, 2018, 9:48 PM UTC

kristianp wrote: post #57

Is there any tech that's faster than DRAM and cheaper than SRAM? There's a need to fill that gap.

L2/L3 cache?

Re: Intel’s Plans for 3DXP DIMMs Emerge

#74

Post by kristianp » Wed, Jul 25, 2018, 1:16 AM UTC

AtlasBarfed wrote: post #73

kristianp wrote: post #57

Is there any tech that's faster than DRAM and cheaper than SRAM? There's a need to fill that gap.

L2/L3 cache?

That's SRAM usually.

Re: Intel’s Plans for 3DXP DIMMs Emerge

#75

Post by magicalhippo » Wed, Jul 25, 2018, 7:25 AM UTC

dragontamer wrote: post #68

magicalhippo wrote: post #58

Earlier quoted context omitted.

But traditional operating systems still assume RAM contents is volatile (because currently it is), most filesystems assume disks are glacially slow etc. A traditional spinning rust HDD has an effective latency of ~10 ms. The NVME version of the 3DXP has an effective latency of ~50 us, or two orders of magnitude better. Not sure how low the DIMM version will go, but maybe another order of magnitude? If so, we're talki…

> But traditional operating systems still assume RAM contents is volatile (because currently it is) RAM is only volatile on startup. Certainly not when a VM hibernates and comes back.

> RAM is only volatile on startup.

No it isn't. Anything that needs to survive a power cycle needs to go to non-volatile storage. And this is assumed to be very, very slow.