Got the Intel 660P M.2 1TB SSD today and its not that slow than some say

[Øyvind Skald 118]

Yes.

This is primary made for storage and low cost. Low cost because of 6 layers of each cell. So, the SSD works a lot more per bit than a single layer storage cell. But the price. This was 1 234 kr.  that is $ 143.79. And to compare prices, the Samsung 860 EVO Series 1 TB is priced at 1 647 Kr ($ 191.92) in the cheapest stores in Norway.

I am using it for Native Instruments/Komplete Samples alone, so I don’t care about Windows boot up times. And it will not be written to on a daily basis. Just the few times I use it I need it to be fast. The technology is also that it has a cache of 1-layer cells. So, the data you use most will be moved into that. It also manage size and what cels on the chips to be this 1 layer cels and it is by use. Only thing important in that regard is that it needs alone time to do such work. I haven’t tested and found out exactly what will make it do this task. But I don’t think Intel will make the CPU run any tasks when this drive has to run catch up.

This was also a OEM thing. No paper box or anything written other than those on the plastic holder. I can’t say you all can buy this and feel safe. Because I haven’t tested the his for more than opening some projects and done one test with Samsung Magician, Samsung’s own SSD tool. But one thing for sure. It is great that SSD storage gets cheaper. And its good that they think of storage too. Not only Windows performance with Random (IOPS), but more storage and fast enough that it is much better than an HDD.

Here is the test of the Intel 660P

Here is a test my Samsung 960 EVO 500GB (Mind you, it is not a fresh installation of Windows on it.)

Here are some pictures of the thing…

Edited February 21, 2019 by ØSkald

[Øyvind Skald 118]

Linus tells so much better the tech stuff than me in this video.

 

[Grem 2,530]

Is that a drive with their Optane technology?

[synkrotron 578]

13 hours ago, ØSkald said:

Linus tells so much better the tech stuff

The guy is a star

7 minutes ago, Grem said:

Is that a drive with their Optane technology?

I thought that the Optane stuff was for conventional HDD's

I could be wrong...

[Grem 2,530]

Just now, synkrotron said:

I thought that the Optane stuff was for conventional HDD's

 

Naw, the ptane stuff is a new non-volatile memory cell technology. When introduced it was suppose to be magnitudes faster than present nand technology. But so for, for the last two years I believe they have released Optane drives that are no faster than present day SSD's. They have just never lived up to the hype that was stated in the original announcements. But still cost way more.

 

But let me say that they are really fast. Just not as fast as we were told. Imagine that.

[synkrotron 578]

Ah, okay, you know more than me Mr Grem

When I was spec'ing up my new workstation there was an option for Optane Memory but there was a note against it "use with mechanical HDD."

Hence my misunderstanding above.

Watching the LTT video...

[Øyvind Skald 118]

3 minutes ago, Grem said:

 

Naw, the ptane stuff is a new non-volatile memory cell technology. When introduced it was suppose to be magnitudes faster than present nand technology. But so for, for the last two years I believe they have released Optane drives that are no faster than present day SSD's. They have just never lived up to the hype that was stated in the original announcements. But still cost way more.

 

But let me say that they are really fast. Just not as fast as we were told. Imagine that.

I think that Optane is great drives. It's  just that Samsungs has worked on their own tech, and Samsung seems to be faster. But does that med there is no potentional in improvement in Optane? I think they both can do much more, but for now, Samsung is leading.

[Grem 2,530]

The very first Optane drives were specifically meant to be cache drives for HDD. They did help in that situation, but not enough to justify the cost to everyday users.

 

Later drives  were meant more for system use and mass storage.

[Grem 2,530]

3 minutes ago, ØSkald said:

But does that med there is no potentional in improvement in Optane?

 

Agreed. Based on the initial announcements we should have Optane drives that smoke anything Samsung has. But that hasn't happened. Yet.

 

It just may be a matter of perfecting the manufacturing process. Intel is just being really quiet about it.

[Øyvind Skald 118]

Here is the specs for my new Drive:

https://www.intel.com/content/www/us/en/products/memory-storage/solid-state-drives/consumer-ssds/660p-series/660p-series-1-tb-m-2-80mm-3d2.html

[mettelus 424]

That video is a great overview of how the internal caching functions, both in theory, and under (extreme) testing. His point about "intended use" is also very valid in that the average user isn't going to be stress-testing writes to the drive, which allows the drive to clean up its caching between uses.

13 hours ago, ØSkald said:

I am using it for Native Instruments/Komplete Samples alone

For this, you will not see significant performance issues for reading if the cells have only been written to once (when you are done copying them). But as shown in the video, if you try to write ALL libraries to the drive at once (his test was a 230GB file copy), you will see the write hit when the cache gets saturated. Basically, the drive is setting aside a portion of itself to use SLC (1/4 density) as a "fast cache" to give the slower QLC writes time to catch up. If it gets filled quicker than the QLC writes, you will start seeing QLC speeds until the drive has enough idle time to "re-orient itself." Initially the drive is 250GB as "SLC," but as you fill it that cache will drop proportionally to 1/4 of the free disk space.

Once written, it shouldn't be anything noticeable to you for reading samples from it, which is your intended use.

[Øyvind Skald 118]

1 minute ago, mettelus said:

That video is a great overview of how the internal caching functions, both in theory, and under (extreme) testing. His point about "intended use" is also very valid in that the average user isn't going to be stress-testing writes to the drive, which allows the drive to clean up its caching between uses.

For this, you will not see significant performance issues for reading if the cells have only been written to once (when you are done copying them). But as shown in the video, if you try to write ALL libraries to the drive at once (his test was a 230GB file copy), you will see the write hit when the cache gets saturated. Basically, the drive is setting aside a portion of itself to use SLC (1/4 density) as a "fast cache" to give the slower QLC writes time to catch up. If it gets filled quicker than the QLC writes, you will start seeing QLC speeds until the drive has enough idle time to "re-orient itself." Initially the drive is 250GB as "SLC," but as you fill it that cache will drop proportionally to 1/4 of the free disk space.

Once written, it shouldn't be anything noticeable to you for reading samples from it, which is your intended use.

Yeah. I did se that. It  stopped up writhing. was really slow at times. I think it took it 2 hour just to write the  NI library. But it seems fast for loading it.

[craigb 3,298]

Linus' Tech Tips is one of the few YouBoob channels I subscribe too.  I loved his "As Fast As Possible" series.