hpr-knowledge-base/hpr_transcripts/hpr2903.txt

Episode: 2903
Title: HPR2903: What is PMEM
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr2903/hpr2903.mp3
Transcribed: 2025-10-24 12:58:14

---

This is HPR Episode 2903 entitled, What is B-M? It is hosted by AWP and in about 8 minutes long and carrying a clean flag.
The summary is persistent memory, B-M also known as storage class memory.
This episode of HPR is brought to you by an honesthost.com.
Get 15% discount on all shared hosting with the offer code HPR15, that's HPR15.
Better web hosting that's honest and fair at An Honesthost.com.
Good day. My name is JWP and I want to talk to you about persistent memory or P-M, also known as storage class memory.
What is persistent memory? B-M is a next generation technology whose data transfer speed is as good as D-RAM, about 50 to 300 nanoseconds, 100 times faster than a SSD.
Unlike D-RAM, it can retain the data after reboots.
In detail, persistent memory P-M is a solid state high performance by the addressable memory device that resides on the memory bus.
That's a key thing for later on. We're talking about something that resides on the memory bus of the computer.
Being on the memory bus allows P-M to have D-RAM access to data, which means that it's nearly the same speed and latency of D-RAM and the non-voleterative of a NAND flash.
It's called this N-V-DEM, a non-vodal dual line in memory module, and the Intel 3D X-Point DEMs, also known as Octane DC, persistent memory, are two examples of persistent memory technologies.
Samsung also makes a memory like this, and we also had a memory module like this.
Persistent memory such as the Intel Octane DC persistent memory provides a full-proof solution installed alongside traditional RAM.
P-M has many advantages of RAM, including low latency access, but it comes in greater capacities until Octane DC, for example, is available in 128, 256, and 512 sizes.
The benefits of persistent memory, persistent memory in the data center allows applications to run without incurring latency penalty without going to storage.
The main advantages include access to latencies less than that of flash SSDs, and increases the throughput more than flash storage.
It's cheaper than D-RAM. P-M is cacheable, and this is a huge advantage over the PCI interconnect, which cannot be cached in the CPU.
Real-time access to data allows ultra-fast access to large data sets, and data-persistent memory after the power and eruption, just like flash.
Some of the use cases for persistent memory, fraud detection, cyber-thread analysis, web-scale personalization, financial training, and Internet of Things IoT, which is very interesting.
Let's go a little bit more into the acronyms.
It says it's non-volatile, double-inline memory module. This means that it's in a dim hardware format.
Then it says, now let's talk about the non-volatile, non-volatile means you can plug it off and on again, and the information will still be there.
That's the key thing. It's in a dim format that allows you to have a function where it'll keep the data there.
Let's look into persistent memory versus N-V-RAM.
Non-volatile random access memory retains its information, even if there's no power. If the power is lost, the data is written to disk.
You don't lose the data because it can be recovered from N-V-RAM. N-V-RAM uses the backup battery to keep data persistent.
During this time, it can flash the data out to a flash device that is attached directly. In most cases, N-V-RAM resides on the PCIe bus.
P-M or N-V-D-RAM minus N. That's another acronym. If you see N-V-D-M with the minus N, it really means P-M.
It can also be backed up by the battery, but it resides on the memory bus. Again, it has the advantages that we talked about before with the BN on the memory bus.
Where is P-M going? It's no wonder the sort of N-V-RAM computing has exploded in recent years. According to Gardner, 75% of the cloud-native application development will use N-V-RAM by 2019 and by 2021, at least 20% of large and global organizations will adopt platforms using N-V-RAM technologies.
What are some of the drawbacks of P-M? P-M is a local store. It's on one host. It doesn't really go across many hosts.
Hosts, failures can result in loss of availability. In case of catastrophic errors, you may lose all the data and must take manual steps to reformat the P-M.
That's a basic overview of P-M and how it works. Again, the leading technology guys are Intel Octane. Do you see that's one that I'm sort of interested in?
Samsung is actively selling it also. Over the years, it's been several stuff. It's not a new thing. Even back in the 80, 86 days with the OSI,
I used to have an in-memory disk that I would use sometimes with a Microsoft database that was incredibly fast.
It's just an evolution of the technology. Again, probably the most interesting is that this resides on the memory bus rather than on the PCI bus.
All right, I put all the links and the notes for y'all. If you need anything, please give me a please contact me at jwpfi.html.com. Thank you very much. Bye.
We are a community podcast network that releases shows every weekday, Monday through Friday. Today's show, like all our shows, was contributed by an HPR listener like yourself.
If you ever thought of recording a podcast, then click on our contributing to find out how easy it really is.
Nekka Public Radio was founded by the digital dog pound and the infonomicon computer club and is part of the binary revolution at binwreff.com.
If you have comments on today's show, please email the host directly, leave a comment on the website or record a follow-up episode yourself.
Unless otherwise status, today's show is released on the creative comments, attribution, share a like, 3.0 license.
You