REGISTER | LOGIN
Breaking News
News & Analysis

Intel Pumps Up PCs with Optane

Sub-$100 M.2 cards accelerate hard drives
3/27/2017 12:01 PM EDT
13 comments
NO RATINGS
Page 1 / 2 Next >
More Related Links
View Comments: Newest First | Oldest First | Threaded View
Page 1 / 2   >   >>
Ron Neale
User Rank
Author
Re: Retention and endurance
Ron Neale   3/31/2017 4:47:06 PM
NO RATINGS
Resistion: I have a calculation on the Optane (tm) cell write/erase lifetime from the DWPD 100GB/day and the endurance figure of 182.5TB I have requested Intel confirm my number before I comment any further on the subject, next week.  My number is for the 32GB version, although that should not change the cell w/e lifetime.

 

resistion
User Rank
Author
Re: Retention and endurance
resistion   3/31/2017 4:10:39 PM
NO RATINGS
I would divide 100 GB by 16 GB to get how many full 16GB array writes per day (only several).

rick merritt
User Rank
Author
Re: Resistance
rick merritt   3/29/2017 8:53:21 PM
NO RATINGS
@fragro: Great points!

Supposedly there was a workshop at the recent Open Compute Project meeting on the future of hard drives in the data center, but I have not been able to get any info on it.

If anyone was there, let me know here or at rick.merritt@aspencore.com

Ron Neale
User Rank
Author
Re: Retention and endurance
Ron Neale   3/28/2017 12:05:20 PM
NO RATINGS
Resistion:- Interesting to describe 100GByte writes/day as endurance. As you say endurance is usually the term used to describe the number of write/erase cycles to failure, or the write/erase lifetime. The reciprocal of GByte Writes/day is days per/GByte write, a WRITE time. In this particular case reducing Bytes to bits and days to seconds provides a write time of 0.108microsecs/bit (I think!) or much less if you write in parallel. A little more clarity please Intel or somebody on this definition of endurance in this new world of Optane(tm) where the reciprocal of write time is apparently endurance.

Shinobee
User Rank
Manager
Re: Retention?
Shinobee   3/28/2017 7:08:42 AM
NO RATINGS
Hi there.

 

I think generally (by convention not by official nomenclature), retention would be the time until any memory cell loses its data due to ambient temperature, radiation etc.  I think the number of write cycles until the cell doesn't work anymore is usually referred to as "endurance."

resistion
User Rank
Author
Re: Retention?
resistion   3/28/2017 6:48:20 AM
NO RATINGS
The number of usable write cycles is the endurance.

fragro
User Rank
Author
Resistance
fragro   3/28/2017 6:13:10 AM
NO RATINGS
It will take time to get used to this new animal, which is not just about speed. This technology will require a rethinking of the current computer in its current design.

I cannot help feeling that hdd is the next floppy disk, and that even the current SSD will soon play second fiddle to this new class of storage/processing hardware (that's what it will be). 

Many of us forgot that even HDDs were not common in "PCs" at the beginning. The current SSD is still marketed as a hybrid next to hdd and now we are already watching the next step in the evolution which will be even more of a sonic boom in the industry.

Interesting times indeed.

Violoncelles
User Rank
Author
Re: Retention?
Violoncelles   3/28/2017 6:09:16 AM
NO RATINGS
In DRAM , the retention is a time after which the cell has lost its information.
in NVM , the retention is the number of write cycles after which the cell cannot work anymore (cannot make the difference between 0 and 1)

resistion
User Rank
Author
Re: Retention?
resistion   3/28/2017 3:56:30 AM
NO RATINGS
If it's worth bragging they would mention it. I think it is tricky for them to report, if it is PCM-based.

ubm112211
User Rank
Author
Re: Retention?
ubm112211   3/28/2017 3:49:50 AM
NO RATINGS
this thing is nvm. what u think the retention should be?

Page 1 / 2   >   >>
Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed