Announcement

Collapse
No announcement yet.

Intel SSD 760p 256GB NVMe SSD For $99 USD On Linux

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • piotr.dobrogost
    replied
    If 760p is budget version of 900p it would make sense to compare it not only with 960 pro but also with its budget version 960 evo…

    Leave a comment:


  • numacross
    replied
    Originally posted by Jabberwocky View Post

    960 EVO 500GB vs 900P Optane 480GB
    Code:
    Read: 2171 vs 2309
    Write: 1320 vs 1950
    4k Rnd Read: 48 vs 189
    4k Rnd Write: 105 vs 170
    Mixed: 1282 vs 2217
    4k Rnd Mixed: 37 vs 126
    Price: 230 vs 600
    To me it looks like it's a lot cheaper and better to use RAM for low latency storage combined with SSD for large storage. Optane might be good for managing database indexes, but still looking at what they promised. It's a joke.
    You can't really "store" things in RAM since it's not persistent. As I wrote before this product is not meant for general consumers and if you look at the price of "proper" server SSDs with SAS/U.2 and internal power protection you'll see that 900P's price is not that outrageous.



    Originally posted by Jabberwocky View Post
    ​​​​​​​July 2015:
    [...]
    ¯\_(ツ)_/¯
    While the marketing blurb of "up to 1,000 times faster" is not yet achieved it might have something to do with the product being first generation. Especially the controller might not yet be up to par. All in all I stand by what I wrote about this tech being interesting since NAND-based SSD landscape is filled with error correction voodoo

    Leave a comment:


  • artivision
    replied
    Originally posted by numacross View Post

    But how do you want Michael to test this? If you ask the SSD to read a block it has previously been ordered to delete all you're going to get is zeroes. This is because the controller is aware where the data actually is according to it's internal bitmaps. What you're saying is true but I'm not aware of any way of reading this short of reverse-engineering the controller and reading the NAND directly which is out of reach of most attackers. Not to mention almost all modern SSDs are doing self-encryption internally. That's why an ATA Secure Erase command takes no time to execute - it's only dropping the key and the bitmaps.
    Firs of all you don't put any effort to understand what others say. Is it really possible that i said, that Michael had any other way to test those specific SSDs? Would the results for each have been different? That i said is to test those SSDs vs others made for privacy or the closest he can find and in any way possible. Or at least make a review for a new tech like that, companies must see that there is demand. And no i don't feel safe with encryption only.

    Privacy like that:
    http://ieeexplore.ieee.org/document/...6/?reload=true
    https://gizmodo.com/after-a-preset-t...-ma-1712508562
    Last edited by artivision; 29 January 2018, 04:10 PM.

    Leave a comment:


  • Jabberwocky
    replied
    Originally posted by numacross View Post

    Optane is not slow, in fact it's the fastest SSD you can get. It's just that it's very expensive to get the proper, multi-channel version (the 900P). This product is not for normal consumers since the differences between it and a NAND SSD are not really that big, however it is noticeable even while casually using Windows.

    Where it shines is: awesomely low access times, bit-addressability (in contrast to traditional which has to hack away in the controller to extract performance out of the page-based NAND) and consistency in performance (https://img.purch.com/r/600x450/aHR0...dlMDYxLnBuZw==).

    Bear in mind that this is the first generation of this product that has to compete against a well established technology (NAND-based SSDs) that had it's own problems in the beginning

    For a newcomer it's doing great. I'm looking forward to next generations of this product.
    960 EVO 500GB vs 900P Optane 480GB
    Code:
    Read:          2171 vs 2309
    Write:         1320 vs 1950
    4k Rnd Read:     48 vs  189
    4k Rnd Write:   105 vs  170
    Mixed:         1282 vs 2217
    4k Rnd Mixed:    37 vs  126
    Price:          230 vs  600
    To me it looks like it's a lot cheaper and better to use RAM for low latency storage combined with SSD for large storage. Optane might be good for managing database indexes, but still looking at what they promised. It's a joke.

    July 2015:
    Intel and Micron today unveiled their all-new memory technology called 3D XPoint (pronounced "cross-point"). This is a new class of memory that can be used both as system memory as well as nonvolatile storage. In other words, 3D XPoint can be used to replace both a computer's RAM and its solid-state drive (SSD).

    The companies claim that 3D XPoint is a major breakthrough in memory process technology, the first new memory category since the introduction of NAND flash in 1989. It's said to be extremely fast and durable, up to 1,000 times faster (both in read and write speeds), and it will have higher endurance than existing NAND Flash memory currently being used in SSDs. What's more, it also has as much as 10 times greater density, leading to much more storage capacity in the same physical space, while remaining as energy efficient and affordable as existing NAND flash memory.
    ¯\_(ツ)_/¯

    Leave a comment:


  • numacross
    replied
    Originally posted by Pepec9124

    Have you never heard about wear leveling? SSDs are not actually removing data when you tell them to do so. You'd need to overwrite whole drive few times to be sure, or ask drive to drop encryption key if it has internal encryption.

    There are some interesting papers on that subject.
    But how do you want Michael to test this? If you ask the SSD to read a block it has previously been ordered to delete all you're going to get is zeroes. This is because the controller is aware where the data actually is according to it's internal bitmaps. What you're saying is true but I'm not aware of any way of reading this short of reverse-engineering the controller and reading the NAND directly which is out of reach of most attackers. Not to mention almost all modern SSDs are doing self-encryption internally. That's why an ATA Secure Erase command takes no time to execute - it's only dropping the key and the bitmaps.

    Leave a comment:


  • numacross
    replied
    Originally posted by Jabberwocky View Post
    Why is optane so slow? The only thing it's good at is doing random. Intel marketed the crap out of it since 2015. It was supposed to be "a revolutionary" product. A few years later and I don't see the insentive to buy this product, especially since you have a new (potentially worse) system in order to use optane?
    Optane is not slow, in fact it's the fastest SSD you can get. It's just that it's very expensive to get the proper, multi-channel version (the 900P). This product is not for normal consumers since the differences between it and a NAND SSD are not really that big, however it is noticeable even while casually using Windows.

    Where it shines is: awesomely low access times, bit-addressability (in contrast to traditional which has to hack away in the controller to extract performance out of the page-based NAND) and consistency in performance (https://img.purch.com/r/600x450/aHR0...dlMDYxLnBuZw==).

    Bear in mind that this is the first generation of this product that has to compete against a well established technology (NAND-based SSDs) that had it's own problems in the beginning

    For a newcomer it's doing great. I'm looking forward to next generations of this product.

    Leave a comment:


  • Davidovitch
    replied
    I just ran this test for my Ryzen 1800X + Samsung 960 EVO NVMe but realized after the run the temperatures were missing. What is actually the dependency for monitoring the temperature of an NVMe drive for PTS, is that nvme-cli?

    Does aynone knows how to run the test profile again but now using another drive of the same machine, is that by changing the environment variable PTS_TEST_INSTALL_ROOT_PATH to the relevant disks?

    Leave a comment:


  • niner
    replied
    Originally posted by artivision View Post
    Michael some times you do second class reviews. I believe that if you ask all Phoronix readers, a 90% will want a SSD privacy based review. For example: if i remove a file the simple way, for how long the SSD will keep my data and what is the potential for removed data to be recovered. I mean an SSD that can write 300MBps but the removed data are destroyed within month is a thousand times better than one that can write 500MBps but destroys removed data in a year or never. You should refer to technologies like that in your next article.
    If you need privacy, encrypt your data and/or drive. It's as simple as that. No need to hope for data to be really deleted within a month.

    The only reliable way to keep your data secure and prevent unwanted access is to encrypt it. Everything else is about as useful as closing your eyes.

    Leave a comment:


  • grok
    replied
    Originally posted by Jabberwocky View Post
    Why is optane so slow? The only thing it's good at is doing random. Intel marketed the crap out of it since 2015. It was supposed to be "a revolutionary" product. A few years later and I don't see the insentive to buy this product, especially since you have a new (potentially worse) system in order to use optane?
    It's mostly useless to normal people. However like RAM it's still fast on small sizes so it's possible to sell it as a 16GB or 32GB cache.
    It's also another talking point to sell Z chipsets. Makes me think of people clamoring for more and more PCIe lanes - I don't think you'll need them, I'd rather have 4x storage lanes multiplexed to 8x or 16x lanes. Or more 2x slots and 2x SSDs - I would hardly care about this 760p being "crippled" to 2GB/s

    Leave a comment:


  • fernie
    replied
    Not buying Intel ssd again.. first 600p was mostly not working (device disappearing suddenly), the replacement had serious issues corrupting file systems ( https://bugzilla.redhat.com/show_bug.cgi?id=1402533 ). It's since fixed with firmware update, but the amount of headache caused by it, not trusting to their drives anymore, did they test it at all before releasing..

    Leave a comment:

Working...
X