500gb or 1tb m2. SSD Recommendation?

Discussion in 'Hardware Components and Aftermarket Upgrades' started by Drew1, Aug 5, 2019.

  1. yrekabakery

    yrekabakery Notebook Virtuoso

    Reputations:
    1,084
    Messages:
    2,947
    Likes Received:
    2,847
    Trophy Points:
    181
    Perf/W is much higher on the 970 EVO than the 660p

    [​IMG]

    [​IMG]
     
    Aivxtla likes this.
  2. Aivxtla

    Aivxtla Notebook Evangelist

    Reputations:
    512
    Messages:
    531
    Likes Received:
    706
    Trophy Points:
    106
    I'm sorry but the 660P is a terrible value at its price point compared to other TLC drives that are priced maybe $50-70 higher for 2TB. 1/2 to full drive performance will also be worse once you get past empty drive tests, not to mention the much worse write endurance which for most people will be ok I guess but I'd still go for a TLC over a QLC for the extra leeway in case you end up doing a lot of video editing/exports or something of the sort where you end up doing a lot of writes/transfers. The 660P will drop to like 100MB/s after very long writes, uncommon for normal use and not too terrible vs HDDs but still much worse than a TLC drive.

    Get a MyDigital BPX Pro 1920 GB for $230, 1920GB Inland (Micro Center's brand) for $235, 2TB Sabrent Rocket for $250 (seen it for ~$200 a while back) or sometimes I've seen the Corsair MP500 for around $230-240.

    As for Samsung, EVO+ is replacing the EVO and that one can sustain above 1GB/s on long writes after pseudo SLC buffer is over. And if you really want to compare either Samsung or any of the other TLC drives to a 660p do a 1/2 drive full test and see what happens..... Same goes for low end TLC drives (controller wise) like the EX950 (A weak Silicon Motion controller like the 660p) vs 970 EVO/MP500.
     
    Last edited: Aug 21, 2019
    Robbo99999 likes this.
  3. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    5,057
    Messages:
    12,464
    Likes Received:
    2,448
    Trophy Points:
    631
    WD Black or Blue, Samsung (all), Adata XPG series are all worst in real-world battery life on notebooks than the Intel 660p is. With the 660p sometimes even giving a more responsive system too. Numbers, spec's and scores don't mean anything if they don't carry through to actual workloads and workflows.

     
    Vasudev likes this.
  4. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    5,057
    Messages:
    12,464
    Likes Received:
    2,448
    Trophy Points:
    631
    If all you're doing is running 'The Destroyer' benchmark, sure it is. Not reflected in any way on a mobile system in the actual workloads though where maximum data rates don't necessarily mean lowest power consumption, overall.

     
    Vasudev likes this.
  5. tilleroftheearth

    tilleroftheearth Wisdom listens quietly...

    Reputations:
    5,057
    Messages:
    12,464
    Likes Received:
    2,448
    Trophy Points:
    631
    Yeah, that might be so based on your limitations. But I have never found any SSD worthy of using without OP'ing by at least 25% and consider 33% OP'ing my personal minimum for sustained performance over time.

    A 2TB nominal SSD is just a 1.25TB (at best) SSD for me in a notebook. At worst, it is closer to 750GB when used in my desktop workstations as a dedicated cache drive (Temp, Scratch Disk, etc.).

    The 'price' for an SSD to run to its full potential is not merely the 'cost' of the drive... Never has been, never will be.

     
    Vasudev likes this.
  6. ole!!!

    ole!!! Notebook Prophet

    Reputations:
    2,505
    Messages:
    5,832
    Likes Received:
    3,792
    Trophy Points:
    431
    destroyer benchmark is like queue depth of 5-6 which is unrealistic to consumer. even anadtech's "light" benchmark runs at around 3-4 QD.

    majority of consumer runs at most QD 1-2 like 95% of the time. i'd say buy the cheapest known brand SSD and just go on with your normal day tasks and be fine for yrs.
     
    tilleroftheearth and Vasudev like this.
  7. Vasudev

    Vasudev Notebook Nobel Laureate

    Reputations:
    9,206
    Messages:
    10,869
    Likes Received:
    8,186
    Trophy Points:
    931
    The same OP values are applied on Seagate Enterprise SSD. A 8TB with OP for sustained write speed and much endurance at 6.xx TB.
    I wonder if you're behind their FW design!:eek:
     
    tilleroftheearth likes this.
  8. yrekabakery

    yrekabakery Notebook Virtuoso

    Reputations:
    1,084
    Messages:
    2,947
    Likes Received:
    2,847
    Trophy Points:
    181
    That’s not true. Almost 3/4 of The Destroyer is at QD 1-2.

    F6249523-0C2B-4E5B-BC51-CC11B070563C.jpeg
     
    Vasudev and Aivxtla like this.
  9. Chastity

    Chastity Company Representative

    Reputations:
    1,295
    Messages:
    6,545
    Likes Received:
    336
    Trophy Points:
    251
    For me, the 2tb 660p is being utilized for game installs, which it is currently half full. Since I am not using it for large file editing, who cares! :) It works, and works well, and uses less power to boot.
     
    tilleroftheearth and Vasudev like this.
  10. ole!!!

    ole!!! Notebook Prophet

    Reputations:
    2,505
    Messages:
    5,832
    Likes Received:
    3,792
    Trophy Points:
    431
    that is the issue, the individual tests yes at only QD1/2, say for example me running VLC for watching a video, but if 3-4 VLC ran at the same time, it is QD4-8.

    some of the tests are likely ran at the same time to reduce time, which is a mistake. now as far as i remember it use to be like that back 4-5 yrs ago, if they fixed it since then it is QD1-2. we can actually figure this out pretty easily by looking at optane storage review of the same tests light/destroyer.

    optane when mixed read/write 80/20 or 70/30 doesnt lose performance like flash cells due to optane is technically a type of memory? so flash you'll get say 700 MB/s when read/write active at same time while optane remains at 1900 MB/s.

    when doing destroyer test we just need to look at performance of optane vs flash. if flash is within half the performance then we know something is maybe inconsistent.

    theres also a lot going into it. sequential R/W and random R/W, if sequential for blocksize over 64kb is considered as multiple QD due to window nature is 64 i think unless specified in raid. also we have no idea if these tests are done with read only, then write only, if so it is very unrealistic.
     
    Last edited: Aug 21, 2019
    tilleroftheearth likes this.
Loading...

Share This Page