Hacker News new | past | comments | ask | show | jobs | submit login
Samsung Demos Its First BGA SSD: 1500 MB/s Read Speed and Tiny Package (anandtech.com)
144 points by desdiv on June 1, 2016 | hide | past | favorite | 131 comments



I have just built a PC with a M.2 Samsung 950 Pro on a Z170 motherboard (make sure you get the Z170 so the M.2 slot can use 4 PCI-E channels).

The performance of this thing is insane. Ars benchmarked it at 2600MB/s read, 1500MB/s write for sequential data[1]. Most SSDs benchmark at less than 600 MB/s[2].

By comparison, a RAM Disk gets ~6000MB/s on the https://hardforum.com/threads/quad-channel-ram-disk-benchmar... benchmark.

[1] http://arstechnica.com/gadgets/2015/10/950-pro-review-samsun...

[2] http://www.guru3d.com/articles-pages/samsung-850-evo-ssd-rev...

[3] https://hardforum.com/threads/quad-channel-ram-disk-benchmar...


How many iops do you get when you test it with fio (http://linux.die.net/man/1/fio) ?

Example: sudo fio --filename=/home/olav/testfile --direct=1 --sync=1 --rw=read --bs=4k --numjobs=8 --iodepth=32 --runtime=60 --time_based --group_reporting --size=1G --name=journal-test


I have two of these in raid 0 in a Lenovo P70.

  fahad@heffalump:~/fiotest$  sudo fio --filename=$PWD/testfile --direct=1 --sync=1 --rw=read --bs=4k --numjobs=8 --iodepth=32 --runtime=60 --time_based --group_reporting --size=1G --name=journal-test
  journal-test: (g=0): rw=read, bs=4K-4K/4K-4K/4K-4K, ioengine=sync, iodepth=32
  ...
  fio-2.2.10
  Starting 8 processes
  journal-test: Laying out IO file(s) (1 file(s) / 1024MB)
  Jobs: 8 (f=8): [R(8)] [100.0% done] [367.8MB/0KB/0KB /s] [94.2K/0/0 iops] [eta 00m:00s]
  journal-test: (groupid=0, jobs=8): err= 0: pid=25184: Wed Jun  1 17:32:44 2016
    read : io=22265MB, bw=379986KB/s, iops=94996, runt= 60001msec
      clat (usec): min=15, max=19595, avg=83.59, stdev=102.01
       lat (usec): min=15, max=19595, avg=83.65, stdev=102.02
      clat percentiles (usec):
     |  1.00th=[   20],  5.00th=[   65], 10.00th=[   66], 20.00th=[   68],
       | 30.00th=[   75], 40.00th=[   79], 50.00th=[   80], 60.00th=[   82],
       | 70.00th=[   84], 80.00th=[   90], 90.00th=[   97], 95.00th=[  108],
       | 99.00th=[  141], 99.50th=[  173], 99.90th=[  900], 99.95th=[ 1704],
       | 99.99th=[ 4768]
      bw (KB  /s): min=43704, max=53976, per=12.51%, avg=47524.23, stdev=1225.35
      lat (usec) : 20=0.74%, 50=1.42%, 100=89.28%, 250=8.32%, 500=0.09%
      lat (usec) : 750=0.03%, 1000=0.03%
      lat (msec) : 2=0.06%, 4=0.02%, 10=0.01%, 20=0.01%
    cpu          : usr=1.38%, sys=7.92%, ctx=5737418, majf=0, minf=89
    IO depths    : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
       submit    : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
       complete  : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
       issued    : total=r=5699889/w=0/d=0, short=r=0/w=0/d=0, drop=r=0/w=0/d=0
       latency   : target=0, window=0, percentile=100.00%, depth=32

  Run status group 0 (all jobs):
     READ: io=22265MB, aggrb=379986KB/s, minb=379986KB/s, maxb=379986KB/s, mint=60001msec, maxt=60001msec


That is awesome, what do you get with writes? (rw=write)


I ran both: https://gist.github.com/nlothian/e22ee30c6ccb8d473058d9e0864...

Read is a lot (LOT) faster than writes. Enough to make me thing I might have something misconfigurated there(?), although I'm not familiar enough with this benchmark to know for sure.


It's not unusual for write performance to be staggeringly worse than read on SSDs. It all depends how the controller is setup and how the manufacturer wants it to benchmark. Most sites emphasize read performance so the companies tune for it.


For 4K random reads, you'll probably want more than 32 iodepth, even on a consumer drive like this. And you'll want to try to write to each offset in the drive beforehand in case they have a "return 0" shortcut (like many enterprise drives).


Anandtech's results may be applicable here too: http://www.anandtech.com/show/9702/samsung-950-pro-ssd-revie...


Not in-front of it now. Will try to remember to do this though.


Please do this, the IOPS are very important!


> I have just built a PC with a M.2 Samsung 950 Pro on a Z170 motherboard

Ditto - I built mine at the beginning of the year. It's very fast, but I learned two things that weren't widely reported at the time.

1) The chips on the drive get _very_ hot under use (thermal throttling kicks in at 75C). Even with minimal use they seem to have a running temp > 50C. I'm a little nervous about the long-term reliability of the drive based on the heat alone.

2) Boot speed with Windows is still fairly slow unless the drive is formatted to GPT (GUID Partitiion Table) instead of the default MBR. It's possible to do this when installing Windows, but there's no easy way to convert after installing


> I learned two things that weren't widely reported at the time.

For what it's worth, the Anandtech review [0] mentioned your first issue in their review back in October. Anandtech has missed things in the past, but I trust them a lot to find and point out possible issues with hardware.

[0] Linking to the last page 'Final Words' section as it is mentioned up front and centre there. A lot of times, I'll read or skim the first page, and jump directly to the final page conclusion for hardware I just want a quick overview of - http://www.anandtech.com/show/9702/samsung-950-pro-ssd-revie...


What operating system do you use? I am trying to install windows 7 on those exact specs (M.2 Samsung 950 Pro on Z170 motherboard, ASRock Extreme6+), and I'm having trouble getting it to either recognize the M.2 slot during install, or partition an alternate SATA Samsung 850 EVO SSD to clone from. Did you have similar difficulties?


Windows 7 can be a PITA in such cases. From what I've seen, Windows 7 has started to show its age when it comes to hardware. Quite a bit of modern hardware like NVMe SSDs and USB 3 are not recognized out of the box. You're most stable bet would be to go with Windows 10 if you can. There'll be tradeoffs which I guess you are already aware of.


I attempted to install the exact same card (M.2 Samsung 950 Pro) into a NUC5I7RYH, as well as a backup SATA (Seagate 1TB Laptop Gaming SSHD (Solid State Hybrid Drive) SATA 6Gb/s 64MB Cache 2.5-Inch Internal Bare Drive (ST1000LM014)) - Neither one of them worked with Windows 7 out of the gate, and I ended spending a 4-6 hour odyssey of googling BIOS upgrades, Firmware Upgrades, and downloading various and sundry windows drivers (my final count was around 8-10 before all was said and done).

Ironically, we thought that Windows 7 would have had better support - but, as it turns out, for Circa 2H2015+ hardware, it's probably not a great platform. Here's hoping that Windows 10 stabilizes for a few years, and all the vendors agree to support that platform....


There is Windows update for 7/2008r2 which adds NVME support you need to cook it into the image.

Use this guide http://forum.notebookreview.com/threads/guide-installing-win... It worked for me. Also this :) http://www.samsung.com/semiconductor/global/file/insight/201...

I have a windows 2008R2 box at home running on a 950pro.


I used Ubuntu 16.04.

Windows 7 doesn't have M.2 drivers by default. See http://www.tomshardware.com/answers/id-2817439/install-windo... and good luck! I believe Windows 10 is fine though.


Windows 10 works well (M.2 Samsung 950 Pro on a Gigabyte GA-Z170X-UD5), but don't make the mistake I did of doing a default MBR-style format on the drive. Put the BIOS in UEFI mode and make sure the drive gets formatted with a GUID Partition Table (GPT).

With MBR the system would still take 30-60s to boot, but GPT reduced that to less than 10s.

You may need to copy the install data to a USB drive to get it to work. Initially I burned the iso image to a DVD and but then that DVD drive wasn't bootable under UEFI.

https://technet.microsoft.com/en-us/library/dn336946.aspx


Have you tried Windows 10 on it? Windows 7 is pretty old now and doesn't have much support for newer hardware. Your performance would be much better on Windows 10.


My bet is 64bit Linux


Dis you change bios to ahci instead of ide for the disk?


I am guessing a flavor of Linux.


ALL sata 3 SSDs should benchmark at 600MB/s or less, because 600MB/s is the maximum supported bandwidth throughput of SATA 6Gb/s


The 950 Pro (NVMe) uses PCIe x4, not SATA.


correct, i was replying to the sentence preceding source 2.

"Most SSDs benchmark at less than 600 MB/s[2]."


Which is why NVMe and M.2 are a thing now. SSDs have surpassed SATA (and in some cases, even SAS)


If you look at the post i was replying to, source two was a sata drive, and the claim was "most ssds cap out at 600MB/s"

of course they do, that is the theoretical limit of non pcie drives.


I used to repair seismic nodes as part of my job. BGAs had the highest failure rate in high shock and temperature cycling environments by far. I find it worrying that so many chips are being designed with BGA. All it takes is one ball disconnect and you have a useless device without an expensive solder reflow. It seems like a really good way to design a device to have to be replaced after 5-10 years...


BGAs can be underfilled [1] to greatly increase mechanical reliability and shock resistance.

1. http://m.eet.com/media/1122054/0112esdnexlogic02.png


I have not knowingly seen this used in the wild but it looks intriguing. I've seen a lot of conformal coatings used to increase shock resistance but from my experience that is only half the battle. The thermal expansion and contraction can be the real slow silent killer. The caveat being that I was dealing with devices rated (and used) from -30°C to 85°C.


It gets used quiet often, even Nest uses it [1]. Obviously proper heat management is important.

1. https://cdn.sparkfun.com/assets/8/5/c/7/4/52e157adce395f4759...


Anybody can explain what are "seismic nodes"? I tried googling it, as well as "repair seismic nodes", but to no apparent success


I was referring to a seismic acquisition node used in geophysical exploration. Here is an example of a seismic node made by Sercel, one of the largest players in that market: http://www.sercel.com/products/Pages/unite.aspx


Probably earthquake monitoring equipment


Petroleum companies search for oil via seismic analysis.


If MTBF is 5-10 years, that's comparable to existing HDDs/SSDs, no?


People are saying it being soldered down is a big disadvantage but I'm not following. The article says that M.2 spec will be available with PCIe or SATA interfaces and there is a photo of a Toshiba version on what looks like a small PCIe card. Aren't these the same as buying an existing card-based SSD?

If not, what is the relationship between all these interface types? Are they intended to be soldered onto the motherboard?


It mostly applies for laptops and other portable devices.

Manufacturers may decide to solder the SSD directly to the motherboard, like Apple currently does with RAM on their laptop computers. This is beneficial to the manufacturer because it's more compact, less chance of mechanical failure, easier to assemble (pick&place instead of humans), lower price etc. In short: it's very attractive for manufacturers.

For consumers this is worrying though, as with the RAM on Apple laptops:

- You cannot replace a failed SSD, you need to replace the entire motherboard

- You cannot upgrade the SSD, so you have to accept the ridiculous price stepup when ordering the laptop (i.e. the 200 dollar price bump to get 8GB extra RAM on a macbook pro)

- Laptop is less valuable when trying to sell it on after a couple of years, because the storage size will be low compared to new SSD's.


One major disadvantage that you can't easily remove the storage media for secure erase. Instead of removing the hard drive and destroying it, then selling/giving away the rest of the PC, companies would probably opt for destroying the entire PC.

If the controller fails, with existing "discrete" SSDs you can still desolder the flash chips directly, put them in a reader, and recover the data that way (e.g. https://news.ycombinator.com/item?id=8133450 ) With this integrated "SSD", it could be nearly impossible to read the raw flash, meaning no more data recovery beyond the logical level. It's basically like an SD card.

less chance of mechanical failure

The widespread problem with RoHS lead-free solder and cracking balls disagrees, as also evidenced by the large amount of reflow/reballing services and the fact that reliability-critical applications like aerospace still use leaded solder and more expensive solder columns instead of balls.


Here's some info on solder columns for everyone else: http://www.topline.tv/CCGA_column.html

While reading that myself, I also happened upon solder springs, which seem like an even niftier idea: http://www.topline.tv/CCGA_MCS_Micro_Coil_Spring.html

Apparently that company makes the solder springs available also has RoHS safe versions of them, perhaps this could solve the cracking problem?

Another company has a paper out which says that you can use their solder column attach process with existing BGA chips: http://www.solderability.com/newsletter082005/ConvertingBGAC...

Could some manufacturer (while unlikely, curious if it's possible) choose to use solder springs with this SSD's BGA package to make it more mechanically reliable when installed?


There's already a Flash upgrade service for iPhones available in China (similar kind of BGA chip). Which is kind of terrifying, the Flash on the iPhone has undersell so it has to be milled off on a CNC router, but apparently the process works quite reliably.

Where there's a will, there is a way, and all that...


There was RAM upgrade service (from 8MB to 16MB) for Psion Revo 15 years ago. RAM was also soldered on board.


I don't remember if the Psion was BGA or not; I'm guessing not which makes soldering it substantially easier.

In a similar vein, you could double the RAM on the original Xbox by soldering a second memory chip to the unpopulated pads on the underside of the board.


Apple caught onto people getting storage upgrade for a reasonable price and made it so those phones will be bricked if you update them to the latest iOS. Scumbag Apple as usual :(


Are you referring to the touch ID replacement, or storage upgrade?


It's not just a little less valuable two years later -- it's a lot less valuable, because the SSD lifetime is unknown and unreplaceable.

This makes manufacturers happy, because more end-users will buy new devices rather than get a used one.


> because the SSD lifetime is unknown and unreplaceable.

I just discovered this concern last weekend when a less than 3 year old Samsung SSD 840 died. Since this is not mechanical I don't know why they don't include extra pins (e.g. address, data) to read the memory in raw from the SSD chips if the failure is in the controller or firmware.


Place whatever value on this you like but I also had a Samsung 840 500GB SSD fail just last week that was almost exactly three years old. It came with a three year warranty so I guess that's the only upside.

It's maddening that not only can I not get back whatever data wasn't backed up (thankfully not much) but I have no way of wiping it before returning it under warranty.

The people who handle the returns for Samsung say returned drives will be "formatted". Given it's uunresponsive, this seems unlikely

It puzzles me why drive manufacturers don't offer data recovery for their own products as a value add service. Nobody could get the data off this drive as cheaply as Samsung themselves.


You (read: everyone) should be using FDE. Every OS worth running supports it, and it provides a real layer of protection in the event of any device failure.

Broken or failed screen, input device, storage controller, yada, -- doesn't matter. FDE means your data is that much more safer than without (and makes it safer provided you use a strong key).


> It puzzles me why drive manufacturers don't offer data recovery for their own products as a value add service. Nobody could get the data off this drive as cheaply as Samsung themselves.

Obviously we have our minds completely synchronized. At this point I don't care about the warranty but about the safety of the information in the disk. I contacted datarecovery.com and they gave me a rough range between $ 300 and 2k to recover the SSD. I don't think the price is crazy but I imagine that is a simple task to achieve if you have the right equipment.


On the last point: storage sizes on laptops have plateaued. I've got a 256gb SSD in my three year old rMBP, and can't see any reason why I'd want more in my new one.


Video. Particularly if you do any HD editing, let alone 4K, you can burn through 100s of gigabytes very quickly.


And games. Just my 280GiB Windows partition on my laptop is hovering around 90% full constantly.

Unless you only use the computer for editing text then storage requirements are growing quicker than ever.


Or Photography. My Lightroom catalogue alone is 60GB. My photos several hundred GB.


I'm shopping for a rMBP and the 256GB flash thing is a big problem. I need OSX + Xcode for the occasional iOS app development, but mostly use Windows. OSX + Xcode and a few iOS emulator images is easily 100GB, so that leaves me with only 156GB on the Windows partition.

I don't mind paying for the higher end 512GB rMBP, but the discreet graphics is a deal-breaker for me. Thus I'm left with configuring the baseline integrated graphics rMBP with 512GB flash.


How come the discrete graphics is a deal-breaker if not for the price? Battery life?


In the order importance:

1. Driver issues on Linux

2. Discrete graphics failing and taking the whole logic board with it. Most of my laptops have died this way.

3. Battery life

4. Heat


battery life is one, but personally I would prefer an otherwise max-spec RMBP with integrated graphics only, especially with the Skylake refresh - historically the swap between the two has been less than reliable, and the modern integrated options are adequate for non-game purposes.


I want SSDs to replace my hard drive storage and store my entire photo collection of 1TB. I want it on SSD because the performance in photo library management apps is a lot better on local SSDs. And I want it on a laptop. So a 2TB SSD being standard would of been very helpful for me.


For the virtual machines there is never enough disk space.


Doesn't Apple do it currently with their SSDs as well?


Could very well be, my MBP is from 2012 it still has a SSD module, don't know about the newer ones.


I just looked it up on ifixit and my late 2013's SSD is removeable, but it's a proprietary device so it's astronomically priced.


What's the most annoying about this is that, ok fine it's proprietary so you would expect Apple stores to be able to read from them.

NOPE. The Apple Store just sends you to a data recovery company instead, because they can't read it.

There's no better sign of bad customer service when you use a custom part but even you don't care to learn how to fix it.


I think you missed one of the big advantage/disadvantages, non-replaceable components let the manufacturer add their own premiums on higher memory/storage variants.


Will the 512GB model be on the same size package? That would be pretty dense. I am still waiting for Samsung to release the 1TB version of their 950 Pro, which comes on a much larger M.2 80mm form factor.


It's slight, but there are already possible thermal throttling issues with the 512GB 950 Pro[0]. I'm wondering if that becomes more pronounced as they try to pack 1TB onto an m.2 form factor.

[0] (search for "thermal" on this page and through the article): http://www.anandtech.com/show/9702/samsung-950-pro-ssd-revie...


For my use-case(desktop PC), it shouldnt really matter much. Access would be bursty so it shouldnt really heat up too much, plus there should be adequate ventilation. I mean I probably don’t need the performance of a NVMe SSD at all, but I like to have the best.



Well I have one of these N1 cases[0] which takes a mini-itx board. The graphics card will be taking up the only full size pci-express slot, so I am looking at M.2 ssds.

[0] https://www.ncases.com/


Is there a reason to not put this into smartphones?


I suspect data carriers would prefer you spend your dollars streaming rather than consuming stored media on the handset.


Hehe, that might well be but I wonder. At least in France I suspect that carriers would rather people did not use all of their allocated BW (you can get 50GB on LTE for 20 Euros per month).


In rural US I pay 45 Euros for 30GB/month. In the city I get about 1mb/s, but in the countryside I can get 40mb/s.


Do you mean $45 or ~$50 (45 euros)?

Also, which cellular provider here in the states is offering 30 GB / month at that low of a price (do they have special rural pricing or something)?


I'm crying, I pay the same of 1/100th the data in Canada


I pay the same for 1% of the capacity.


I doubt that. Less streaming per customer means they can support more customers with a given amount of infrastructure (I'm thinking about towers and spectrum in particular), so their infrastructure costs effectively go down on a per-customer basis.


I'd guess the power use is on the high side for a phone.


I think that's what makes most sense for it.


One of these in 128Gb (16GiB) might be cheap enough to be stacked below the RAM/CPU combo on the next raspberry Pi. That would significantly increase the performance and reliability of an rPi, I'd love to see that!


I'd hate to see that, one of the big benefits of the RPi is that everything stateful is on the memory card.


Yeah, until you actually start using the sdcard. Sd cards and usb sticks corrupt in no time somehow


Personally, I've found that external drives only tend to corrupt if you use them as rw boot drives or if you don't unmount and sync them properly (safe remove / eject) on every use.

So yeah, the RasPi idea of having using those to boot might corrupt quickly, but if you use them for a camera or just transferring files between computers, you should be safe.


If these could fit in a 1u chassis on a PCIe slot without a bracket, that would be awesome!


It's 16mm x 20mm, you could probably lose it inside a PCIe socket.


Could not find anything on a ball pitch. Can it be suitable for DIY projects?


If you have an X-Ray to check if the balls are soldered perfectly, then yes. Put it on the board and off with it in the reflow-oven. Keep in mind, it's only 20x16mm.

If not, then you're mostly out of luck, but you can try if you manage to source it. And find some datasheets. And have a driver for it. And...

On a more serious note: It would be very nice, if it was possible to design a board with components like this, but it is way to hard and needs way too much equipment and experience to do right.


Well, OSH Park and alike are surprisingly good now, I even had some success recently with an ice40lp WLCS16 package (by pushing well beyond the recommended design constraints), got 1 passable board out of 3. So even a very small ball pitch is somewhat suitable for DIY.

What caught my eye in this SSD package is a low ball count, so it might be potentially suitable for DIY even without an expensive hardware. If it's possible to route it on 4 layers without too many small vias it might be a game changer.


Low ball count also means that it is easier for the part to not sit flat when soldered.


I guesstimated it to 0.83 mm from the single image in the article. That's not very small by BGA standards, as far as I know.

On the other hand, 1,500 MB/s data and PCIe are quite serious levels of performance by most DIY project standards. :)


Still it might be suitable for a relatively cheap mid range FPGA, even Spartan6.

EDIT: actually, no, it's a single lane PCIe3, such clock rates are well beyond the FPGA range. Pity.


Why not just use a lower PCI Express version? You might lose some speed but it will be backwards compatible with the lower clocks.


Actually, both Xilinx and Altera offer PCIe v3.0 blocks on their high end FPGAs (Virtex 7/Kintex 7 and Arria 10/Stratix V).


Yes, of course, but this is already a little bit beyond the DIY budget I hoped for.


I made a Kintex 7 board in my homemade reflow oven and 2/2 worked. I had more trouble with the QFN I used for the core power supply than the FPGA.


I'm more concerned about a budget here than a reflow. Cheapest Kintex 7 part is around £90.


Ah I missed the part about DIY - that sucks though.

You seem to be familiar with PCIe 2 though, so now I'm wondering - is there a chip that could convert the 1x PCIe 3 lane into a 2x or 4x PCIe 2 set of lanes? Also, worst case, if you really wanted to use this chip, couldn't you just use the chip with PCIe 2 (albeit slower, but I think revisions of PCIe are supposed to be cross compatible)?


> is there a chip that could convert the 1x PCIe 3 lane into a 2x or 4x PCIe 2 set of lanes?

Never heard of such a thing, although it might have been really useful.

> couldn't you just use the chip with PCIe 2

Did not even think of this option, sounds interesting. I guess we should wait for the datasheet for this chip to be sure.


What are the very obvious downsides to using these, lifetime or cost?


It's BGA, so it's soldered onto the motherboard/logic board. Non user upgradeable, and if the SSD dies, you're in for a costly repair. Another step in the direction of (very) disposable hardware with little to no service options.


This isn't necessarily the case. The article explicitly talks about physical standards for putting them on M.2 cards, and how some are already shipping on M.2 and PCI-E form factors.


Surely someone can solder it onto an USB stick, SATA interface, etc. with the added circuitry of supporting those interfaces.


Does the other packages give you this ability? The other fine pitch parts are also very difficult to change over.


This is a chip that gets soldered directly on the board. While it is possible to solder this specific chip to a detachable interface that can be removed, I highly doubt anyone will ever do it.

This form factor has a huge advantage for OEMs - they can just slap on a single chip that has everything built in - controller, memory management and the actual storage. Less parts = less convoluted assembly process and that is cheaper. While theoretically you can put a BGA fully integrated SSD SOC like this on a PCI-E card or M2 etc., it kinda beats the purpose.


Personally, I think BGA parts are great. They are small and quite robust mechanically an highly integrated (as you have said).

I did my early engineering years with DIP parts, the days of when you could just desolder the memory device and place with a higher density device are long gone.


SATA, PCIe, and M.2 drives are trivial to replace.


Apart from being soldered to the board? Not much, I suspect; the tiny form factor probably keeps costs down. It'll be several chips wirebonded together. It's essentially a microSD card with BGA pins.


I dunno, at least initially this could be the first M.2 2240 PCIe SSD. As far as I am aware currently you can only buy M.2 2280 PCIe SSDs. That's a significant savings in space. Alternatively -- for example an ATX motherboard -- could offer two sockets where previously only one would've fit. The Intel 200 series chipsets for Kaby Lake happens to raise the number of PCIe lanes by 4, how fortunate!


It's essentially a microSD card with BGA pins

Also known as eMMC.


> Apart from being soldered to the board?

That should be an absolute dealbreaker, though. Talk about holding your data hostage...


Why is that a dealbreaker? If you don't want your data held hostage to fortune you back it up.


Backups and being able to access the disk don't cover the same area. If they used the same standard connectors as everyone else ever, you could take it and stick it into another computer and be up and running again, in a total of 5 minutes. Backups are rarely that comprehensive.

Also, get real, most people don't have any backups set up at all. They still shouldn't be artificially prevented from accessing their data in case of a failure...


Probably both at first, but eventually there will be no downsides everything moves move into single ASIC/SOC eventually.

If you can get your memory density high enough there is no need for multiple chips, if there is no need for multiple memory chips then a discrete controller doesn't makes sense since it only needs to communicate with a single chip in that case the extra die costs can easily be canceled out by cheaper packaging costs and the fact that you can now sell SSD's to system integrators that could not use them before.


That page says "BGA SSD not only integrates NAND flash memory, but also a NAND controller, DRAM" - which makes sense as they need 3 different types of different silicon processing.

We're not quite at the stage of putting the controller in the chipset and just soldering Flash to the board for PCs, although that is how it's done for SoC systems.


Most dies already have to go through multiple processes even when it's not a SOC/CPU so I don't think the extra cost is that high. You are saving money on binning and packaging costs, and multi die packages are also a possibility if you want to make low cost BGA parts with less performance.


I'd love to ship my software on a little SSD like that.


There are rumours floating about that Nintendo's next game system might go back to game cartridges, which is pretty much exactly "shipping software on a little SSD".


Do ROMs come this big? SSDs make a shitty long term storage solution as they require continuous write cycles to keep the data alive which causes damage to the flash over time

Can't imagine a 512GB ROM would ever be economical given the cost to manufacture and make manufacturing changes to such a beast


Mask ROMs will always be cheaper at scale than Flash.


How in the world are you going to get the scale anywhere near the same as the flash market though?

I have been wondering about going back to carts. When I see a 64GB Flash Drive for $10 retail I have to think that that the price premium over pressed Blu-Ray disks starts to make sense. You can significantly cut down on loading times, avoid obnoxious installs to the local disk, and have potentially more flexibility in the future to release larger games that don't require multiple disks.

Plus, the optical drive is one of if not the most failure prone component on the system (next to only the fan), so eliminating it could cut down considerably on returns and warranty claims.


When I say 'at scale', I really mean ~1000 units. You can already get under your price point with Mask ROMs (and they're faster than flash!).


The problem is that your competition isn't Flash Drives at probably $5/per in bulk. It's Blu-Ray disks that cost $0.005/per in bulk. Plus you need to build up a factory to make massive ROMs, since most people doing this are targeting sizes in the handful of KB. You can't even take advantage of competition in the marketplace because your product will be so specialized. Not like the disk pressing houses that are all over China.


I really hope your software isn't even remotely this huge to justify it though.


Next Gen Games? instead of multiple blu-rays, one of these devices?


Why couldn't it be?


You mean not all pieces of software are tiny weather apps?


That could be a good idea for game consoles, especially handhelds.


That's very few pins for a BGA device, might be something for the hobbyist?


It doesn't need that many pin's it's an SSD on chip, all it needs is power and PCIe pinout.


BGA is typically not for the hobbyist market because they're so hard to solder.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: