https://news.samsung.com/global/samsung-mass-producing-industrys-first-512-gigabyte-nvme-ssd-in-a-single-bga-package-for-more-flexibility-in-computing-device-design
How soon before we get that on a microsd card?
>>54839184
3 years
>>54839184
>LPDDR4 on an sd card
Never
>>54839191
Cool, adding extra RAM on a phone through a microsd card would be fucking baller.
>>54839140
Meanwhile youve still got to pay 200$ extra for 112GB when buying iphone.
I read about this yesterday, it's pretty crazy what they're able to do.
Imagine having this shit in future phones and other devices.
NVMe speeds in everything. It's going to be great.
They should start adding these to cameras.
I just wish they'd stop dicking around with these pathetic 128-500GB sizes and woud finally go well over the 2TB range.
Also what about the temps?
950 PRO can hit about 100C when under full load, imagine how toasty that little thing is going to get.
Put this thing inside a phone or something and you're going to burn a hole in your leg, if it starts accessing files while in your pocket.
>>54839322
>Also what about the temps?
Yeah, why is it so hard to add heatsinks to them?
>>54839322
>Also what about the temps?
They'll probably have embedded power management that'll throttle the chip to lower the temps. I wonder how they will perform under agressive throttling though.
>>54839696
>Yeah, why is it so hard to add heatsinks to them?
The whole point of a BGA package is to save space. Heatsinks consume space, and without proper airflow to dissipate the heat from the heatsink, they won't be as efficient as you could think.
>>54839216
>adding extra RAM
That on-chip RAM is to be used by the SSD controller, probably as a caching device, and it's 4 Gb (gigabit), which is 512MB, and that ain't much by today's standards.
>>54839140
>>54839146
will this SSD get up to 125C or will I have to take my money elsewhere?
>>54841940
So, you only buy SSDs that get up to 125C...?
>>54841983
Yes.
>>54839140
That is actually pretty damn impressive. Can't wait to see a slim 10'' core m laptop with this in it, it's going to be great. This might also enable some even more radical UMPC designs than the ones we've been seeing on kickstarter.
>>54842036
>it's going to be great.
It's going to be pretty much the same as one with an M.2 SSD.
This will only make a difference in phones, smaller tablets, etc.
>>54839140
>you ssd runs hotter then your gpu and cpu combined
>>54842245
>your SSD gives you third degree burns
>>54839925
>heatsinks consume space
rubbish
to the trash it goes
>>54839140
>BGA
Oh goodie, now you won't even be able to upgrade storage because that'll be soldered in too!
![greedy-jewish-person-cartoon[1].jpg](https://i.imgur.com/9tbZVZW.jpg "greedy-jewish-person-cartoon[1].jpg")
>>54842338
No upgrades for you goyim.
>>54839140
Holy fuck. I just took the PCI-E SSD from my old Asus EEE 900 and measured one of the 2 GB chips on it, this is just over 30% larger while holding 256x more data
>>54844635
mSATA ≠ PCIe
>>54845358
ain't that the truth, learned that the hard way.
>>54839925
>>54839696
>>54842326
The way many m.2 slots are mounted, it would actually be pretty easy to put a heatsink on the chips with an epoxy. I've thought about doing it but I have no thermal epoxy on hand (although I do happen to have appropriately sized RAMsinks).
With a BGA package it would basically look like a heatsink on a northbridge or southbridge controller, were it to be used in a desktop application.
That'd be kind of interesting to see a super integrated solution for thin mini-ITX boards, or even a really, really small AIO/desktop solution.
>>54845358
The older Asus EEEs used mini PCI-E for everything, not mSATA. Otherwise getting a larger hard drive would be cheaper.
>>54845423
I've heard people adding heatsinks to the m.2 drives, and apparently that has actually allowed pic related to work without throttling.
It's really strange that there's no heatsinks in these things by default.
They run hot as fuck.
>>54845481
mPCIe SSDs actually exist? I thought there were only mSATA drives
the next 5 years are going to be pretty amazing, exponential growth is real
>>54845732
Those drives were designed for burst throughputs of 2200MB/s not a constant flow of data at 700MB/s.
That's why it's 8W in a tiny un-heatsinked board.
>>54845767
Yeah, and they're a bit more expensive than mSATA drives.
>>54839140
>>54839146
>samshit shill
>literally bricks their own hardware with firmware fro m time to time
>>54845858
Got a link? I've never been able to find them. I want to toss an SSD in my laptop's WWAN mPCIe slot
>>54839140
This is sick. Expect next wave of laptops with even better battery life thanks to larger batteries.
Or who knows many someone will actually design a non thermally limited laptop now with decent cooling. Nah just kidding.
>>54839140
This is actually their second or third iteration of these chips, I think going back to early 2015 or mid 2014. Regardless, they've been making them for a while now. It only naturally follows that density would increase. Its good to see technology move down this path of consolidation as drive current is lowered over all, and total power consumption ends up lower. We're getting to pretty astounding levels of performance in mobile platforms.
>>54845768
It's great that something is advancing like this.
Since it seems that CPU progress has almost stagnated and GPU improvements are slowing down.
Not to mention it's about time that storage speeds started getting better.
Now it just remains to be seen which standard is going to triumph in the end.
There's the 3D Xpoint, which we still know very little about and it's supposed to be ridiculously fast and now confirmed coming out this year
http://techreport.com/news/30216/intel-computex-keynote-confirms-kaby-lake-and-optane-for-2016
NVMe drives have been advancing like crazy and they're fast as hell.
Then there was the HP Memristor thing, that I think was supposed to compete against the Xpoint, but I don't know what happened with that one.
Also SATA connector has become virtually obsolete with these new speeds and now it's going to be a showdown between m.2 and u.2 and whatever other connectors are out there.
It's going to be a bit of a mess until a new universal standard arises.
>>54845824
That's why I put a heatsink on mine.
>>54846182
Good for you, buddy.
>>54846219
Thanks, friend.
>>54845824
>>54845732
Well, I only have the slower 950 Pro 256 GB version, and after benchmarking the shit out of it I really couldn't get it to heat up, so I'm not very concerned.
All the testing I've read on the interwebs on the heat problem are from the 512 GB model.
Mine runs at a "mere" 2200/900 MB/s read/write.
When I eventually replace the PSU (because Silverstone's SFX-L series is some annoying sounding shit) I may end up epoxying the heatsinks on there anyway for shits and giggles. I've had those GDDR ramsinks from a previous build back in like... 2004. It'd be nice to use them again.
>>54846251
it's the high write speed that really heats the drive up.
>>54845934
Amazon lists some. I'd honestly recommend against it though unless you have no other choice as they're around $1 per GB.
>>54846043
I don't think it's that they couldn't do it faster I think they just try to get as much money out of an architecture before moving on as possible. There always seems to be some wall that engineers say is the absolute limit but it never ends up being true. Humans are pretty fucking inventive.
>>54839140
So its a super high speed sd card???
>>54846314
That makes total sense on why I'm not seeing it then.
>>54846794
No, different interface and everything.
