Samsung lifts lid on 1.6TB flash whopper spaffing data at 3GB/s

Samsung has developed its first NVMe-connected 3,000MB/s solid-state drive (SSD). And it is extending its three-bit-per-cell range of flash storage with a 1TB whopper. NVMe stands for Non-Volatile Memory Express, and is a standardised protocol for applications to access data stored on PCIe-connected SSDs via a software driver …

[cyke1]

for prices, http://www.pcper.com/news/Storage/Samsung-SSD-EVO-suggested-pricing-revealed

[Anonymous Coward]

To get the UK price, just change the $ sign to a £ sign.

[Antonymous Coward]

Interesting link...

$250GB < 2*$120GB

$500GB < 2*$250GB

and... for the top of the line 1TB drive...

$1TB < 2*$500GB!!!!!!!!!!one!

I'd love it to come true but I can't imagine that glaring badthink making it past the whalesong and cocaine department.

[Anonymous Coward]

Shut up...

And take my money!

[Steve Davies 3]

Re: Shut up...

That well be an arm and two legs please?

[Anonymous Coward]

Re: an arm and two legs

If I throw in a kidney, can I have the "Enterprise" 1.6Gb one, please?

[Anonymous Coward]

My newly aquired 250GB Samsung 840 SSD feels old already.

[Neil Barnes]

no word on the 840 EVO's price and affordability

But to be fair, '1TB', 'Flash', and 'affordable' are not usually words found in close conjunction with each other...

[Anonymous Coward]

Re: no word on the 840 EVO's price and affordability

It won't be cheap enough for the Wintel world. They won't adopt anything of any decent quality unless it is dirt cheap.

Having DMA seems to be a luxury in Wintel land, despite it being on the Amiga computer in 1985.

[Sandtitz]

Obvious troll is obvious

Oh yes, the wonderful world of Amiga and the zealot users. Apparently paying less is stupid these days.

The PC has always had DMA channels. That was when "Commodore" actually meant either PET2001 or VIC-20.

Amiga users really had to donate livers and pay through the nose to have equipment that was pretty much standard on PC's back then - real time clock, hard drives, even a secondary disk drive was much more expensive than the PC equivalent. Not to mention the paltry 512kB memory standard on A1000/A500.

I should know, I too was on the Commodore bandwagon back then.

[JEDIDIAH]

Re: Obvious troll is obvious

You're on crack. Simply on crack. There was no point in time where Amiga kit was more expensive than PC kit.

When half a meg was "paltry", you would be hard pressed to find more on a clone. If you did manage to find a better equipped (upgraded) clone, then it likely wasn't cheap either.

It took PC gear a long time to catch up to Amiga and everyone else. It took PC operating systems even longer.

What memory you had on an Amiga you could use without any hacks or shenanigans.

[Annihilator]

Shortsighted developers

I suspect the reason we're waiting for prices is that Samsung are waiting for their developers to recompile their pricing calculator, changing the declaration of int_price to long instead of short....

[jcrb]

And the write performance is?

Not even mentioned.

Not even sequential write numbers?

Going to guess that means they are so horrific that they would rather leave us guessing than even just quote sequential write bandwidth much less random IOPs.

Mixed performance numbers? Yeah I'm guessing those might not be so great either

[Michael Duke]

Re: And the write performance is?

The SSD provides up to 98,000/90,000 random read/write IOPS and 540MB/sec sequentially reading, 520MB/sec sequentially writing. These numbers vary with product capacity.

[Annihilator]

Re: And the write performance is?

You have to read articles veeerrrrrry carefully before jumping in with a comment moaning there's a salient fact missing. Usually turns out it's there, as in this case.

[jcrb]

Re: And the write performance of the XS1715 is?

Not even mentioned for the XS1715

Not even sequential write numbers for the XS1715

Going to guess that means they are so horrific that they would rather leave us guessing than even just quote sequential write bandwidth much less random IOPs for the XS1715

Mixed performance numbers? Yeah I'm guessing those might not be so great either for the XS1715

There I fixed my post... happy now?

[Alan Brown]

Re: And the write performance is?

It's not the sequential write times which matter as much as the latency scatter time.

840s are generally ok, but they do exhibit occasional LONG write latency.

[Anonymous Coward]

How much is it? Oh. NVM...

[Anonymous Coward]

"The Korean biz claims 740,000 random read IOPS for this NVMe product, not quite matching the 750,000 of Micron's P320h and P420m."

Oh yes, those extra 1% random read IOs - and who knows how each manufacturer measures those. Samsung's claims are essentially matching those of Micron, not "not quite matching".