Intel Optane 900p SSD Released in AIC and U.2 Form Factors

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
Remember, many NVMe SSDs have write cache.
I hear ya :/ that's my conclusion I could come to as well...

Would be a nice article comparing the write cache capacity for the 750 400 /800 / 1.2 and same thing for p3700 / p3600 and the new optanes 900p / 4800 / 4800x as well as a general performance comparison of them all too :)
 

William

Well-Known Member
May 7, 2015
789
252
63
66
From Charles Wirth - FUGGER over at XS.
XTREMESYSTEMS

I talked with him on the phone about these drives a week or so ago. He compared the speed difference is like... remember when you first went from HDD's to a SSD ? This has the same effect... but SSD's to Optane.
 
  • Like
Reactions: Stux and T_Minus

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
From Charles Wirth - FUGGER over at XS.
XTREMESYSTEMS

I talked with him on the phone about these drives a week or so ago. He compared the speed difference is like... remember when you first went from HDD's to a SSD ? This has the same effect... but SSD's to Optane.
I was HOPING for such a notice!

Going to run one in my new windows/desktop/workstation system that's for sure!!!
 
  • Like
Reactions: William

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
LOL! I have a 750 400gb in there now, gonna yank it sell it, and get one of these new ones that's for sure, sadly will be gone for a week coming up for work/conf so will have to wait tell Mid Nov. :( LOL!!
 

William

Well-Known Member
May 7, 2015
789
252
63
66
Always something in our way to new Toys !!!
Dang It !!
I have to wait also :(
 

CookiesLikeWhoa

Active Member
Sep 7, 2016
112
26
28
35
So would these work with something like a SM 2028U-TN24R4T+? Would be interesting to get a couple to use as a SLOG for an external HDD array.
 

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
From Charles Wirth - FUGGER over at XS.
XTREMESYSTEMS

I talked with him on the phone about these drives a week or so ago. He compared the speed difference is like... remember when you first went from HDD's to a SSD ? This has the same effect... but SSD's to Optane.
Woke up this morning to look at those benchmarks again :) very nice.

Anyone compared the 480 to the 280 that you know of?
 

BackupProphet

Well-Known Member
Jul 2, 2014
1,092
650
113
Stavanger, Norway
olavgg.com
I'd like to see more Enterprise level testing because all the sites are comparing to consumer drives, and this images is kind of scary...

https://img.purch.com/r/600x450/aHR...QL1kvNzIyODA2L29yaWdpbmFsL2ltYWdlMDA5LnBuZw==
I have no idea how they get those numbers, but most consumer focused review have absolutely no idea about how to test drives. I highly recommend everyone to benchmark with fio, which is the ultimate storage performance benchmark tool. storagereviews use fio exclusively.

Fio was originally written to save me the hassle of writing special test case
programs when I wanted to test a specific workload, either for performance
reasons or to find/reproduce a bug. The process of writing such a test app can
be tiresome, especially if you have to do it often. Hence I needed a tool that
would be able to simulate a given I/O workload without resorting to writing a
tailored test case again and again.

A test work load is difficult to define, though. There can be any number of
processes or threads involved, and they can each be using their own way of
generating I/O. You could have someone dirtying large amounts of memory in an
memory mapped file, or maybe several threads issuing reads using asynchronous
I/O. fio needed to be flexible enough to simulate both of these cases, and many
more.

Fio spawns a number of threads or processes doing a particular type of I/O
action as specified by the user. fio takes a number of global parameters, each
inherited by the thread unless otherwise parameters given to them overriding
that setting is given. The typical use of fio is to write a job file matching
the I/O load one wants to simulate.
Benchmarking is hard! If a site want to differentiate from competition, being transparent, with detailed and good explanations will make you look more trust-worthy.
 

Waldek

New Member
Apr 3, 2016
17
2
3
50
Woke up this morning to look at those benchmarks again :) very nice.

Anyone compared the 480 to the 280 that you know of?
Allyn at PC Perspective has (Intel Optane SSD 900P 480GB and 280GB NVMe HHHL SSD Review - Lots of 3D XPoint! | PC Perspective). This is 3D XPoint not NAND, so the rule "the bigger, the faster" does not apply here. In the tests, there seem to be tiny differences between the models, but I guess this is either test noise or some thermals influencing.
But no one so far seems to be confirming the experience being any similar to the HDD->SDD, still the latency sensitive workloads should see tangible improvement.
 
  • Like
Reactions: T_Minus

Patrick

Administrator
Staff member
Dec 21, 2010
12,513
5,805
113
But no one so far seems to be confirming the experience being any similar to the HDD->SDD, still the latency sensitive workloads should see tangible improvement.
HDD -> SSD was a 10-100x improvement. SSD -> NVMe SSD was a 2-5x improvement. NVMe SSD to Optane NVMe is more like the latter rather than the former.

Remember, you can make QD1 synthetic workloads last 15 minutes but at some point, if you are only moving 5MB or 50MB of data, being 3x faster on an operation that takes <1s is not too noticeable.

I have thought about getting a pair of 480GB drives for primary hosting on the forums but I think I prefer cheaper 800GB drives. Not having to deal with capacity issues is a good thing and the NVMe NAND SSDs have enough DRAM to make things run relatively smoothly.

Just a note here: if you are using our standard fio / iometer test scripts they are not indicative of real-world workstation performance. I do not think there are many folks who are hammering consumer SSDs to steady state using 4K full disk random writes for 15 minutes before they even start recording data. If you have free space on your disk, you are not doing full disk.
 
  • Like
Reactions: T_Minus

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
Thanks for that article.

Comparing to 4800X Here:
Intel Optane SSD DC P4800X 375GB Review - Enterprise 3D XPoint | Saturated IOPS Performance - 4KB, 8KB Random, 128K Sequential

It looks to me like the 900P do 50,000 IOPs @ QD1 while the 4800X does 100,000 at QD1 and then comparing reads the 4800X is a bit faster in that regard too.

I wonder if the 4800X improvement is due to controller/firmware for Enterprise or a combination of that + higher over provisioning? (Does that still matter for optane?)
 

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,641
2,058
113
For anyone wondering you can't use Newegg Store Credit Card for Pre-Orders :(

Guess I'll be waiting until it's "IN STOCK" in a couple days...

Oh, well... gonna get 2x 280gb to start I don't need much capacity in my desktop or testing for that matter. If all goes well maybe some more.
 

MiniKnight

Well-Known Member
Mar 30, 2012
3,072
973
113
NYC
For anyone wondering you can't use Newegg Store Credit Card for Pre-Orders :(

Guess I'll be waiting until it's "IN STOCK" in a couple days...

Oh, well... gonna get 2x 280gb to start I don't need much capacity in my desktop or testing for that matter. If all goes well maybe some more.
Amazon gift cards and free prime shipping for me! Too bad it's not available yet there.
 

azev

Well-Known Member
Jan 18, 2013
769
251
63
Does this optane have the same CPU/mobo requirements like the original optane drive ?
I am curious whether this will work with E5 v1 cpu systems ?
 

Patrick

Administrator
Staff member
Dec 21, 2010
12,513
5,805
113
Does this optane have the same CPU/mobo requirements like the original optane drive ?
I am curious whether this will work with E5 v1 cpu systems ?
P4800X just ideally needed PCIe 3.0