Intel Optane P4800x vs P900

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

ekiro

New Member
Mar 16, 2019
20
2
3
Other any other difference between the Intel Optane P4800x & P900 other than their endurance rating (20.5 vs 5.11 PBW), MTBF (1.6m vs 2m), and Enhanced Power Loss Data Protection? I also read that the P4800x has the option to change the sector size. I've also seen some benchmarks where the P4800x performed slightly better on some tests.

The reason I ask is because I purchased a P4800x and believe I may do fine with just a P900. The return window is still open. I'm curious what you guys think.
 
  • Like
Reactions: cdoublejj

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,625
2,043
113
Yep, agree with Patrick we do the same, it's even nice and affordable to throw 1 in an E3 setup :D
 

ekiro

New Member
Mar 16, 2019
20
2
3
I deploy 900p and 905p drives on any machine that is running databases at this point. We are budget constrained so two 900P / 905P in RAID 1 at prices lower than a P4800X is worth the slight feature differentiation.
My server does not support a U.2 drive. I only can go with the AIC version of the NVMe drive, so RAID1 is not an option for me. My server is a HPE ProLiant DL325 G10. AMD EPYC 7401P /w 8x 32GB RAM. Based on my resesrch the P900 is almost the same performance wise. The only difference that appealed to me is their Enhanced Power Failure protection. Right now I can't make up my mind if I should just return the P4800x and get the P900 or not... hoping to hear some more input from you guys.

Also I heard that there are better drives available for the P4800x over the P900. Not sure if that's just a Windows thing of not. I will be running the latest Ubuntu. Does Ubuntu have the optimal software/drivers to maximize the P4800x's performance?
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
I'm nowhere near as knowlegable as the above rockstars, but I'm happy to share my thoughts as I have 4 905s.

3 points that haven't been attended to that I'll comment on:
  • The sector size change you are referring to is facilitated by isdct, which does not function with the 900p. Frankly, I miss using it previously with a P3700 and find the extremely limited SMART data provided by ESXi (vs. isdct) to be quite annoying. Further, isdct allows for control of a number of variables beyond just sector size which can't be changed on the 900p.
  • I'm not sure if you are referring to the lack of an ESXi VIB for the 900p or not, but one is available for the P4800x but not the 900p. Of no real consequence I suppose.
  • But here is a kicker for you while on the topic of ESXi:
    • Use the 900p in ESXi = void your warranty
    • Additionally, the Product will not be subject to this Limited Warranty if used in:(i)any compute, networking or storage system that supports workloads or data needs of more than one concurrent user or one or more remote client device concurrently;(ii)any server, networking or storage system that is capable of supporting morethan one CPU per device; or (iii)any device that is designed, marketed or sold to support or be incorporated into systems covered in clauses (i) or (ii)
    • https://www.intel.com/content/dam/s...-storage/intel-client-ssd-module-warranty.pdf
    • I would imagine this would absolutely impossible to enforce and if I had to guess why this clause exists = it would be to ensure "Enterprise" and "Ethusiasts" stay in their respective lanes and given the similar performance you mentioned but the 900p being much cheaper, it may be simply a scare tactic to ensure enterprises are not purchasing nearly equally capable 900ps, which I imagine could stand to hurt INTL's top line and margins given the premium they sell for.
    • I highly doubt this clause exists such that a STH home labber running a 900p in ESXi would have their warranty voided.
(hopefully this is somewhat helpful)
 
  • Like
Reactions: cdoublejj

zack$

Well-Known Member
Aug 16, 2018
701
315
63
But what about the p4801x?? Now available on eBay, Amazon and Newegg (to name a few sellers) in m.2 and u.2 at 100gb. Sure there is some performance drop in writes (estimated half of 900p) but pricing is pretty aggressive for those who, so far, have been undecided and wanted plp and higher write endurance.

Seems Intel has been keen to listen to the pro-sumer market!
 

ekiro

New Member
Mar 16, 2019
20
2
3
But what about the p4801x?? Now available on eBay, Amazon and Newegg (to name a few sellers) in m.2 and u.2 at 100gb. Sure there is some performance drop in writes (estimated half of 900p) but pricing is pretty aggressive for those who, so far, have been undecided and wanted plp and higher write endurance.

Seems Intel has been keen to listen to the pro-sumer market!
My server has no U.2/M.2 ports. Only AIC will work.
 

copcopcopcop

Member
Feb 2, 2017
36
11
8
36
But what about the p4801x?? Now available on eBay, Amazon and Newegg (to name a few sellers) in m.2 and u.2 at 100gb. Sure there is some performance drop in writes (estimated half of 900p) but pricing is pretty aggressive for those who, so far, have been undecided and wanted plp and higher write endurance.

Seems Intel has been keen to listen to the pro-sumer market!
I am considering grabbing two of these P4801x’s and connecting them to a supermicro AOC-SLG3-2E4 pcie card for a ZFS mirrored SLOG.

I was originally looking at the 900p but would prefer to have power loss protection, also 100gb should be fine for a SLOG device.
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
100gb Version has less max iops.
yeah, i still think the p4801x should do fine in a homelab environment, no?
I too have really pondered the answer to this question, considering ...
  • 900P: Sequential Write (up to): 2k MB/s / Random Write (100% Span): 500k IOPS
  • P4800X: Sequential Write (up to): 2k MB/s / Random Write (100% Span): 500k IOPS
  • P4801X: Sequential Write (up to): 1k MB/s / Random Write (100% Span): 250k IOPS
... at what point does the lower sequential write speed / random write IOPs "hold back" a pool of spinning rust? (perhaps I'm looking at this completely the wrong way)

Example:
  • NFS41 on zpool = RaidZ 3x4x10TB w/ 20GB Optane 900p vDisk as SLOG / dataset = 64K recordsize / sync = always
  • Tested with IOmeter v 1.10 / Win10 VM 8 vCPU / 8 GB RAM / Secondary Disk
  • Random Performance shown with Reads @ 100% Reads / Writes @ 100% Writes / Both @ 100% Random
Hypothetically, would I see MB/s top out at 1000 MB/s (with a P4801x) instead of 1,441 MB/s (with a 900p as tested with) as this IOmeter data shows?

 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
And perhaps to look at this from an IOPS perspective, same test parameters as above, but 4k only ...

IOPS peak at 30,185 @ QD16, which represents only 1 of the 4 writes required for RaidZ (in my case) ... 30K x 4 = 120K = starting to get somewhat close to spec for P4801x, whereas with the 900P (or P4800x), there is still plenty of headroom.

(again, I could be looking at this in the wrong manner)

 

zack$

Well-Known Member
Aug 16, 2018
701
315
63
Answer: SLOG benchmarking and finding the best SLOG

And perhaps to look at this from an IOPS perspective, same test parameters as above, but 4k only ...

IOPS peak at 30,185 @ QD16, which represents only 1 of the 4 writes required for RaidZ (in my case) ... 30K x 4 = 120K = starting to get somewhat close to spec for P4801x, whereas with the 900P (or P4800x), there is still plenty of headroom.

(again, I could be looking at this in the wrong manner)

 

zack$

Well-Known Member
Aug 16, 2018
701
315
63
P4801X 100gb:

Code:
/dev/nvd0
        512             # sectorsize
        100030242816    # mediasize in bytes (93G)
        195371568       # mediasize in sectors
        0               # stripesize
        0               # stripeoffset
        INTEL SSDPEL1K100GA     # Disk descr.
        PHKM90150015100D        # Disk ident.
        Yes             # TRIM/UNMAP support
        0               # Rotation rate in RPM

Synchronous random writes:
         0.5 kbytes:     28.0 usec/IO =     17.4 Mbytes/s
           1 kbytes:     27.0 usec/IO =     36.1 Mbytes/s
           2 kbytes:     27.5 usec/IO =     71.1 Mbytes/s
           4 kbytes:     23.4 usec/IO =    167.1 Mbytes/s
           8 kbytes:     30.4 usec/IO =    257.0 Mbytes/s
          16 kbytes:     44.1 usec/IO =    354.3 Mbytes/s
          32 kbytes:     64.6 usec/IO =    483.5 Mbytes/s
          64 kbytes:    103.7 usec/IO =    602.7 Mbytes/s
         128 kbytes:    161.1 usec/IO =    776.1 Mbytes/s
         256 kbytes:    285.8 usec/IO =    874.6 Mbytes/s
         512 kbytes:    527.5 usec/IO =    947.9 Mbytes/s
        1024 kbytes:    988.2 usec/IO =   1012.0 Mbytes/s
        2048 kbytes:   1905.6 usec/IO =   1049.5 Mbytes/s
        4096 kbytes:   3730.2 usec/IO =   1072.3 Mbytes/s
        8192 kbytes:   7398.6 usec/IO =   1081.3 Mbytes/s
P4801X 200gb:

Code:
root@freenas[~]# diskinfo -wS /dev/nvd0
/dev/nvd0
    512             # sectorsize
    200049647616    # mediasize in bytes (186G)
    390721968       # mediasize in sectors
    0               # stripesize
    0               # stripeoffset
    INTEL SSDPEL1K200GA    # Disk descr.
    PHKM8522004W200E    # Disk ident.
    Yes             # TRIM/UNMAP support
    0               # Rotation rate in RPM

Synchronous random writes:
     0.5 kbytes:     24.8 usec/IO =     19.7 Mbytes/s
       1 kbytes:     25.0 usec/IO =     39.1 Mbytes/s
       2 kbytes:     25.3 usec/IO =     77.2 Mbytes/s
       4 kbytes:     22.9 usec/IO =    170.5 Mbytes/s
       8 kbytes:     25.1 usec/IO =    310.8 Mbytes/s
      16 kbytes:     29.8 usec/IO =    523.5 Mbytes/s
      32 kbytes:     41.0 usec/IO =    762.9 Mbytes/s
      64 kbytes:     60.3 usec/IO =   1037.3 Mbytes/s
     128 kbytes:     96.6 usec/IO =   1293.5 Mbytes/s
     256 kbytes:    162.0 usec/IO =   1543.1 Mbytes/s
     512 kbytes:    291.4 usec/IO =   1715.6 Mbytes/s
    1024 kbytes:    551.4 usec/IO =   1813.6 Mbytes/s
    2048 kbytes:   1073.1 usec/IO =   1863.8 Mbytes/s
    4096 kbytes:   2109.1 usec/IO =   1896.5 Mbytes/s
    8192 kbytes:   4183.8 usec/IO =   1912.1 Mbytes/s
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
P4801X 100gb:

Code:
/dev/nvd0
        512             # sectorsize
        100030242816    # mediasize in bytes (93G)
        195371568       # mediasize in sectors
        0               # stripesize
        0               # stripeoffset
        INTEL SSDPEL1K100GA     # Disk descr.
        PHKM90150015100D        # Disk ident.
        Yes             # TRIM/UNMAP support
        0               # Rotation rate in RPM

Synchronous random writes:
         0.5 kbytes:     28.0 usec/IO =     17.4 Mbytes/s
           1 kbytes:     27.0 usec/IO =     36.1 Mbytes/s
           2 kbytes:     27.5 usec/IO =     71.1 Mbytes/s
           4 kbytes:     23.4 usec/IO =    167.1 Mbytes/s
           8 kbytes:     30.4 usec/IO =    257.0 Mbytes/s
          16 kbytes:     44.1 usec/IO =    354.3 Mbytes/s
          32 kbytes:     64.6 usec/IO =    483.5 Mbytes/s
          64 kbytes:    103.7 usec/IO =    602.7 Mbytes/s
         128 kbytes:    161.1 usec/IO =    776.1 Mbytes/s
         256 kbytes:    285.8 usec/IO =    874.6 Mbytes/s
         512 kbytes:    527.5 usec/IO =    947.9 Mbytes/s
        1024 kbytes:    988.2 usec/IO =   1012.0 Mbytes/s
        2048 kbytes:   1905.6 usec/IO =   1049.5 Mbytes/s
        4096 kbytes:   3730.2 usec/IO =   1072.3 Mbytes/s
        8192 kbytes:   7398.6 usec/IO =   1081.3 Mbytes/s
P4801X 200gb:

Code:
root@freenas[~]# diskinfo -wS /dev/nvd0
/dev/nvd0
    512             # sectorsize
    200049647616    # mediasize in bytes (186G)
    390721968       # mediasize in sectors
    0               # stripesize
    0               # stripeoffset
    INTEL SSDPEL1K200GA    # Disk descr.
    PHKM8522004W200E    # Disk ident.
    Yes             # TRIM/UNMAP support
    0               # Rotation rate in RPM

Synchronous random writes:
     0.5 kbytes:     24.8 usec/IO =     19.7 Mbytes/s
       1 kbytes:     25.0 usec/IO =     39.1 Mbytes/s
       2 kbytes:     25.3 usec/IO =     77.2 Mbytes/s
       4 kbytes:     22.9 usec/IO =    170.5 Mbytes/s
       8 kbytes:     25.1 usec/IO =    310.8 Mbytes/s
      16 kbytes:     29.8 usec/IO =    523.5 Mbytes/s
      32 kbytes:     41.0 usec/IO =    762.9 Mbytes/s
      64 kbytes:     60.3 usec/IO =   1037.3 Mbytes/s
     128 kbytes:     96.6 usec/IO =   1293.5 Mbytes/s
     256 kbytes:    162.0 usec/IO =   1543.1 Mbytes/s
     512 kbytes:    291.4 usec/IO =   1715.6 Mbytes/s
    1024 kbytes:    551.4 usec/IO =   1813.6 Mbytes/s
    2048 kbytes:   1073.1 usec/IO =   1863.8 Mbytes/s
    4096 kbytes:   2109.1 usec/IO =   1896.5 Mbytes/s
    8192 kbytes:   4183.8 usec/IO =   1912.1 Mbytes/s
Thanks! I was looking just at page 9 (and didn't see those diskinfo benchmarks there). The 200 Gb looks darn close to 900p / P4800x ...

So the point is that spinning rust can't hit those speeds at lower recordsizes? Clearly the case with the 200Gb iteration, but what about the 100Gb?

(I for some reason thought this guy only came in 100 Gb versions - slaps head)
 

Davewolfs

Active Member
Aug 6, 2015
339
32
28
So did you end up going with the 4801X or 900p?

It’s interesting that Intel specifically states that one device has PLP while the other does not. Yet people seem to believe that PLP is not necessary.

Is there an edge case that Intel is aware of? Why would they advertise this otherwise?
 

Rand__

Well-Known Member
Mar 6, 2014
6,626
1,767
113
Original specs of 900p included PLP, that was removed in later revisions of the documentation
 
  • Like
Reactions: T_Minus

TrumanHW

Active Member
Sep 16, 2018
253
34
28
I SO wish I understood what was being said here.

Were those specs about the 900 and 905p ... and maybe in RAID or something
- where it said 100GB specs ...and then
- 200GB specs ..?
or is that saying that the performance of a larger capacity drive just doubles the stats!??

Very confusing.

What IS the difference between the 905p / p4800x / p4801x ...?

Were people talking about using them in a raid of 4 drives!? Geesh.

How does it work..?

Let's say I'm writing 420GB of data ... and I have a 375GB Optane + 48GB of RAM ...

Does that mean that no matter HOW slow the spinning drives are ... I could write 423GB (obviously a little less bc of overhead and formatted capacity, MiB vs MB, etc) ... but between the DRAM + Optane, they could literally hold 100% of the data being written to buffer the performance..? Is that literally how it works..?

Is that how you choose the size of an Optane sort of..? And then I'd need a

• ZIL
• L2arc


Thanks!