backtop

Print 72 comment(s) - last by The0ne.. on Jul 6 at 2:11 PM

Apple is rumored to be packing 19 nm, 400 MBps, ONFI 3.0 compliant flash memory onto the motherboard of the new MacBook Air.  (Source: 9 to 5 Mac)
Integrating chips into the motherboard could allow the MacBook Air to offer high performance affordably

With a refresh of the MacBook Air seemingly impending the rumor mill surrounding Apple Inc. (AAPL) is once again heating up.  

According to Macotakara.jp, a Japanese Apple fan site, the latest version of the ultraportable will use cutting edge flash memory technology.  The site cites sources at unnamed Asian chipmakers as claiming that the latest Airs will contain 19nm flash memory chips soldered directly onto the motherboard for blazing speeds of up to 400 megabits per second.

Current generation MacBook Airs only have RAM chips soldered on.  The SSD for the laptops are connected using mSATA connectors.

Apple reportedly calls the new technology Toggle DDR2.0.  Aside from the rather curious title (ostensibly it involves NAND flash, not DDR memory), the technology is expected to be Apple's proprietary implementation of the Open NAND Flash Interface (ONFI) Working Group's ONFI 3.0 standard.

ONFI 3.0 promises faster speeds and reduces the number of pins on flash memory chips.  These factors add up to nearly "instant" boot times and fast file copies.

If the rumor about the new flash is true, it likely comes from Micron Technology, Inc. (MU), Intel Corp. (INTC), or Spansion Inc. (CODE) -- all members of the ONFI 3.0 coalition.  Apple's current NAND supplier, South Korea's Samsung Electronics (SEO:005930), and the world's second biggest producer, Toshiba Corp. (TYO:6502) are not currently members.

A split with Samsung would make sense, given Apple's ongoing legal war with the gadget and component maker.

At times Apple has been on the bleeding edge of introducing new standards.  It was the first major manufacturer to push Intel's proprietary LightPeak interface, which it renamed Thunderbolt.  The refresh of the MacBook Air is expected to come packing Thunderbolt as well, which requires fancy chip-laden $50 USD cables to work.

OS X Lion and the refreshed MacBook Air are expected to launch on July 14.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
new standards?
By zorxd on 7/5/2011 5:01:54 PM , Rating: 2
Apple was last to get PCIe, last to get DDR2, etc.
Apple pushing new standards is the exception, not the norm. It remains to be seen if Thunderbolt will become a standard or not. It might just fail or remains niche like Firewire.




RE: new standards?
By Reclaimer77 on 7/5/2011 5:55:54 PM , Rating: 1
I don't know a single standard with active cables that has seen widespread consumer support. I think the question of Thunderbolt failing or not has been already been declared because of that decision alone.


RE: new standards?
By Pirks on 7/5/11, Rating: -1
RE: new standards?
By Reclaimer77 on 7/5/2011 6:17:42 PM , Rating: 1
Sorry but Thunderbolt will lose and lose horribly to USB 3.0. Cable prices fall, but active cable prices don't fall that much. Also Thunderbolt drives the cost up for OEM and the Enterprise segment as well because Thunderbolt takes you from having 1 central controller for all your devices like USB to having a controller for each one. One USB controller can handle up to 127 individual devices! With Thunderbolt each device has a separate controller.

Oh and "daisychaining" is supposed to be a great feature? Yeah good luck diagnosing a bad device when you have a ton of devices daisychained together. No IT professional in the world would rather put up with that than USB 3.0. Ironically "daisychaining" will give OEM's less incentive to manufacture Thunderbolt hubs, so that takes us back to my first point.

Economics, bugs, and poor implementation will bury Thunderbolt. Mark my words.


RE: new standards?
By Pirks on 7/5/2011 6:24:54 PM , Rating: 1
Well, daisy chaining SCSI devices did not bury SCSI at all, so excuse me for NOT marking your words.

Besides who cares if USB is mass consumer cheapo standard and Thunderbolt is more expensive fast prosumer standard? These two standards can coexist forever pretty well since they are targeted at different markets.


RE: new standards?
By Reclaimer77 on 7/5/2011 6:34:45 PM , Rating: 1
Did you really stop and think before you used SCIS as your example? I'm going to give you time to think about it and retract that.


RE: new standards?
By Pirks on 7/5/11, Rating: 0
RE: new standards?
By Reclaimer77 on 7/5/2011 6:39:15 PM , Rating: 2
Now you're being an idiot. I don't know why I even bother with you.


RE: new standards?
By Pirks on 7/5/11, Rating: 0
RE: new standards?
By Reclaimer77 on 7/5/2011 6:53:41 PM , Rating: 2
Yeah cause with no edit feature, nobody could possibly make a typo...

Stop being a kid.


RE: new standards?
By Pirks on 7/5/2011 7:00:19 PM , Rating: 2
Well, if you wanna talk serious - start talking serious. Tell me what didn't you like about SCSI daisy chaining, for instance.


RE: new standards?
By Reclaimer77 on 7/5/2011 7:05:44 PM , Rating: 2
The two aren't comparable, and if you thought for even a second about it, you would have seen that.


RE: new standards?
By Pirks on 7/5/2011 7:14:47 PM , Rating: 2
Well, you said daisy chaining wouldn't work for Thunderbolt while it was working just fine for SCSI. So you don't want to explain why?


RE: new standards?
By Reclaimer77 on 7/5/11, Rating: 0
RE: new standards?
By quiksilvr on 7/5/2011 8:06:03 PM , Rating: 4
Judging from the long list of threads here I think you have plenty of time to "waste". And it isn't waste if you are informing people the downfalls of SCSI.

Pirks, despite your uh...controversial views, I would be happy to explain the downfalls of this system.

As explained earlier, each device has to be configured to run Thunderbolt. This means that a bit more complex and expensive hardware has to be in place. This is much like SCSI. Yes it was parallel and allowed simultaneous data, but the format was also confusing and had multiple types of I/O. Plus, at the time it had a very poor BiOS support.

Obviously Thunderbolt will not have the same downfalls as SCSI, but Thunderbolt is doomed to fail.

Frankly, in my opinion at least, I wasn't quite sure why Thunderbolt and USB 3.0 weren't one in the same thing. I don't see the reason why the Light Peak technology could not have integrated with the USB port design. That way, you can use active and passive cables on the same port, thus opening the market for multiple formats.

Although I am always for the more superior technology, I don't see Thunderbolt winning against USB 3.0. You can already use the millions (if not billions) of devices that connect via USB on it without having to get a new port and still get more than fast enough speeds for consumer and business markets. Furthermore, the cables are incredibly expensive.

I actually thing $50 is relatively cheap for the performance and reliability it gives and that price will take years to drop. But look at what USB 3.0 already has in the market. Right now you can buy a 32GB USB 3.0 drive for $50 and any computer today can get a USB 3.0 PCI slot for a mere $20, if not less.

Thunderbolt took too long to release and was released on one of the lowest selling computers in the market. If it came out on HP or Dell this would have been a different story. But on laptops that cost over $1500? On desktops that cost over $1000? There's no chance for this to become widespread.


RE: new standards?
By Pirks on 7/5/11, Rating: 0
RE: new standards?
By quiksilvr on 7/6/2011 12:47:46 AM , Rating: 2
If a product doesn't become widespread it fails. Fail doesn't mean wiped from existence it just means that it becomes a very small, niche product. Stop splitting hairs over sentences that are essentially saying this will not be successful.


RE: new standards?
By Pirks on 7/6/2011 1:15:59 AM , Rating: 2
quote:
If a product doesn't become widespread it fails
Ferrari, Bugatti and Aston Martin are not widespread, so they fail. Say, uh, did you hit your head with any heavy objects lately?


RE: new standards?
By dark matter on 7/6/2011 2:43:07 AM , Rating: 2
Pirks.

You have to be an employee of DT, no-one could keep up this bullshit of yours for this length of time without getting paid for it.

Troll's get bored easily. You have TOO much consistency to be just a Troll. You're a shill from DT.

Fuckoff, it really bores me.


RE: new standards?
By Pirks on 7/6/11, Rating: -1
RE: new standards?
By JAB on 7/5/2011 9:13:35 PM , Rating: 3
You can not share the USB connector because the USB consortium doesn't allow it.
USB 3 is only a small fraction of the thunderbolt
The 50 dollar cable with a cip not needed if you don't mind the performance hit and or use a short cable.
You can use a hub just like USB don't know wy you are trying to make it sound like rocket science.
Hmmm not sure how you missed Apples growth it is one of the top 3 PC makers if memory serves.
The total cost of ownership on an apple is usually lower than a windows PC make that as you will but google the research.
It powered eSATA had better implementation and adoption then yes USB 3 would be enough but you would still be talking about a minimum of 3 connectors for the same use as one thunderbolt with lower bandwidth higher latency higher CPU load and reduced battery life. SATA can AMD USB can be real power hogs for laptops in comparison to certan alternatives. Outside of people on forums like this AMD old fashioned IT guys people don't get sentimental about connectors.

Don't be surprised if it gets widely adopted on the high end if you need the function it will save money and power though only after redesigning the laptop down to the chipset level.


RE: new standards?
By quiksilvr on 7/6/2011 1:03:50 AM , Rating: 2
1) What are you talking about? Both were developed by Intel (which makes it even more confusing).

2) How much of a hit is there? My guess is its quite a significant hit going passive to make it get outpaced by 6 Gbp/s SATA and 4.8 Gbp/s USB 3.0, otherwise they would have launched it with passive and say it beats those two standards.

3) Agreed its not complicated for one device but if you want to make thousands if not tens of thousands, the money adds up for manufacturers.

4) In America it's number 3 but globally its number 5. Why wouldn't you at least put it in with the top two competitors? Who pegs a golden standard on 3rd or 5th place? Furthermore, why wasn't it just spread throughout ALL top manufacturers? Why the limitations?

5) A minimum of three connectors for the same use as one Thunderbolt? For what purpose? And what are these claims about latency? That depends on what's using it. A hard drive will give you latency (sometimes as high as a whopper of 4 milliseconds) but that's through no fault of SATA or USB. It is limited by its rotational speed. Solid State negates this latency argument.

And bandwidths on even the fastest Solid State Drives have yet to push 6 Gbp/s SATA to its limit yet (unless its PCI...but its connected via PCI and reaching its potential so who cares?). And when that day comes when it surpasses SATA, they will just upgrade SATA along with it.

6) Agreed, but it has a lot of competition and its introduction was not nearly aggressive enough for this to happen.


RE: new standards?
By Bad-Karma on 7/6/2011 1:18:14 AM , Rating: 2
Quote from Anand "Light Peak isn't a protocol, but instead simply provides a new physical layer for existing connector protocols. Traffic like DVI-D, HDMI, USB, and audio can all coexist across Light Peak." http://www.anandtech.com/show/3930/intel-light-pea...

The article even mentions that the goal is to include display port, Ethernet, firewire,PCIe. Think of it as an opportunity to get everything down to one cable. Right now it's at 10Gbps but it has the promise of scaling to 100Gbs. USB just can't go there, at least not for a few more renditions.

Article: http://www.anandtech.com/show/3930/intel-light-pea...


RE: new standards?
By CharonPDX on 7/5/2011 8:30:52 PM , Rating: 2
SCSI is alive and well, thankyouverymuch.

It is now "SAS", which is still very much in use. (As is Fibre Channel, as is FireWire in some circles, both loosely based on SCSI; although not as much as SAS - which stands for, wait for it... Serial Attach SCSI.)

Yes, USB replaced it for the more 'consumer' focused uses, as it should have. But SCSI did not die. Likewise, I imagine Thunderbolt will live on as a more focused connector than Apple intends it to be. It will live on as a high-speed interconnect between systems and peripherals, possibly to replace Fibre Channel. (Have you priced a Fibre Channel cable recently?)


RE: new standards?
By Pirks on 7/5/2011 9:17:37 PM , Rating: 1
Oh, poor Reclaimer :))) LOL


RE: new standards?
By Reclaimer77 on 7/5/2011 9:49:12 PM , Rating: 2
quote:
SCSI is alive and well, thankyouverymuch.


Where in the hell do you people get out of my posts that I was bashing SCSI? I was just trying to get Pirks to understand he's comparing apples to oranges.


RE: new standards?
By retrospooty on 7/6/2011 1:27:58 AM , Rating: 3
"I was just trying to get Pirks to understand"

Mistake #1 . LOL . It will never happen.


RE: new standards?
By Pirks on 7/6/11, Rating: 0
RE: new standards?
By The0ne on 7/6/2011 2:11:10 PM , Rating: 2
SCSI devices are still used for high speed data acquisitions (100+ data points). Granted other techs are catching up in speed but SCSI still is useful in the right applications and are usually a lot more reliable, hence the use in critical applications and fields.


RE: new standards?
By 8steve8 on 7/5/2011 6:41:16 PM , Rating: 2
agreed.

and even if Light Peak cables eventually become passive (not sure what their plan is... but still...), it'll be the firewire of the last generation of interconnects.

usb2 won because it was built into all chipsets.
usb3 will be built into all chipsets next year, light peak will not. It will require extra chips/money/space/energy on top of the chipset. This is why it will be a niche product.

That said, niche products can still be awesome, but I'm not sure why apple is pushing it, who benefits? i suppose its nice to use one port for display and possible expansion out of a macbook air, as apposed to a separate minidisplayport and esata port... still...

next year when ivy bridge chipsets have built in/free usb3, it'll be sad that the macbook air will require the extra space / chip/ energy to power a chip for the thunderbolt functionality in the minidisplayport.

seems like a real miscalculation on their part.


RE: new standards?
By 8steve8 on 7/5/2011 6:45:26 PM , Rating: 2
backwards compatibility also sealed the deal for usb2


RE: new standards?
By Pirks on 7/5/2011 6:53:53 PM , Rating: 2
quote:
it will be a niche product
Is this equal to Reclaimer's "lose and lose horribly" above? If yes then I agree with you.
quote:
why apple is pushing it, who benefits?
Apple customers of course, who else would that be?
quote:
it'll be sad that the macbook air will require the extra space / chip/ energy to power a chip for the thunderbolt functionality in the minidisplayport
Sad because of extra speed gained by using this power? Since when did gaining extra speed become "sad" huh?
quote:
seems like a real miscalculation on their part
Yeah, being too fast is miscalculation, better being slow eh? :)))


RE: new standards?
By 8steve8 on 7/5/2011 7:19:33 PM , Rating: 2
eh i don't believe many consumers would notice the difference between a device on a usb3 and a thunderbolt port.

could you give me an example?


RE: new standards?
By Pirks on 7/5/2011 7:33:24 PM , Rating: 2
SSD RAID


RE: new standards?
By 8steve8 on 7/5/2011 8:17:01 PM , Rating: 3
yeah definitely something the typical apple consumer will use on their laptops. //


RE: new standards?
By 8steve8 on 7/5/2011 8:20:07 PM , Rating: 2
- externally -

: )


RE: new standards?
By yomamafor1 on 7/5/2011 8:51:24 PM , Rating: 2
It really shows how he's in tune with actual Apple buyers... LOL


RE: new standards?
By Pirks on 7/5/2011 9:21:41 PM , Rating: 2
this is for typical Apple prosumer, these guys kill for speed


RE: new standards?
By seamonkey79 on 7/6/2011 12:39:04 AM , Rating: 3
Then why do they buy Apple? For the blazing last generation video, or the blazing 3rd and 4th tier CPUs? It must be for the blazing speed of the whites from the LCD that get to your eye faster than any other computer's whites...


RE: new standards?
By Pirks on 7/6/2011 1:29:54 AM , Rating: 2
quote:
For the blazing last generation video
Macs are not the best gaming machines, so what?
quote:
blazing 3rd and 4th tier CPUs?
So Xeon E5620 is "3rd and 4th tier" CPU? I smell some envy here! Looks like you can't afford those, hence you call them "3rd and 4th tier", hahaha :)) I can see through ya! No need to try to BS me.
quote:
It must be for the blazing speed of the whites from the LCD
Don't envy the Thunderbolt bandwidth. You work some more hours and you could then afford at least one Thunderbolt cable. Maybe. Such is life, not everyone can get the fastest stuff. Deal with it.


RE: new standards?
By Bad-Karma on 7/6/2011 2:04:17 AM , Rating: 2
quote:
Xeon E5620 is "3rd and 4th tier"


Itanium -Tukwila (soon to be eclipsed by Poulson)

E3-12xx-series "Sandy Bridge"

E7-x8xx-series "Westmere-EX"

6500/7500-series "Beckton"

3600/5600-series "Gulftown & Westmere"
X5698
X5690
X5680
X5677
X5675
X5670
X5660
X5650
L5640
L5638
E5640
E5630
E5620
L5630
L5609
W3670
W3680

C3500/C5500-series "Jasper Forest"

So your E5620 is actually at the lower third of the 5th tier (not counting Poulson). Sells on Newegg right now for $389, just $75 more than a I7-2600K. http://www.newegg.com/Product/Product.aspx?Item=N8... Which is 4 times cheaper than my last Raid card......


RE: new standards?
By Pirks on 7/6/2011 3:05:50 AM , Rating: 2
Decomposing corpse of Itanic has no place here, and if seamonkey meant "3rd of 4th place in the global speed hierarchy of all CPUs out there" then yeah, I kinda agree with that. I know Macs don't have the fastest CPUs on Earth right now. Just wasn't sure about what seamonkey meant exactly.


RE: new standards?
By yomamafor1 on 7/5/2011 9:10:28 PM , Rating: 2
You really have no clue do you?

First, let's talk about the technical side. Currently the Thunderbolt has the most bandwidth @ 10Gbit/s, or approximately 1250MB/s, while USB 3.0 maximum throughput is 5Gbit/s, or 625MB/s. So USB 3.0 has about half the bandwidth of Thunderbolt. We also know that Thunderbolt will be used to transmit large video/audio files to an external storage. However, the fastest 7200RPM HDD, what will most likely be used as the storage medium, is 100MB/s. Even the fastest 3TB drive at this time is 170MB/s, which only consumes less than 30% of the total bandwidth of USB 3.0, and less than 15% of the total bandwidth of Thunderbolt. So why should I go out and get a $50 cable, just have a bunch of unused bandwidth lying around, when I can get a $.50 USB cable that get the same performance?

Next, the practicality side. For desktop users, if they want to use SSD RAID, they put them in the case via SATA, not externally. For laptop users, they won't use an external SSD to RAID with the internal drive because the moment you break the RAID, you have to rebuild it. You can RAID two SSDs externally, but what's the point of RAIDing two SSDs for performance, just to use as a mass storage device? So both of your scenarios make absolutely no sense at all.


RE: new standards?
By ph0tek on 7/5/2011 9:24:06 PM , Rating: 2
Theres very little use for LighrPeak on Apples laptops, it was a stupid decision considering the type of people (idiots) that buy them. But it has many other uses for professional video editors for example, many do not use HDD's, they're not fast enough. And Sony's new Z laptops use LightPeak (and they call it LightPeak too, not Thunderbolt) to power an external GPU. USB3 has nowhere near the bandwidth for that. LightPeak is basically PCIE in cable form and that has many uses, just not so much for average consumers yet.


RE: new standards?
By Pirks on 7/5/2011 9:41:57 PM , Rating: 1
quote:
it was a stupid decision considering the type of people (idiots) that buy them
So the guys who do professional HD NLE on these laptops are idiots? More likely the idiot is you then. Much more likely.


RE: new standards?
By Pirks on 7/5/2011 9:38:06 PM , Rating: 2
quote:
You can RAID two SSDs externally, but what's the point of RAIDing two SSDs for performance, just to use as a mass storage device? So both of your scenarios make absolutely no sense at all.
Learn what the HD NLE is son :))) You just have no idea about some very interesting things that exist outside your little box.


RE: new standards?
By ph0tek on 7/5/2011 9:14:07 PM , Rating: 2
You can connect external GPU's with LightPeak (i'm never calling it thunderbolt). As with Sony's new laptops. I dont think USB3 and LightPeak are meant to compete. Light Peak is considerably superior though and will last a lot longer. Already USB3 barely cuts it for SSD's.


RE: new standards?
By yomamafor1 on 7/5/2011 9:17:51 PM , Rating: 2
Barely?

http://www.anandtech.com/show/4256/the-ocz-vertex-...

Assuming best case scenario for sequential write, OCZ Vertex 3 can hit 370MB/s.

The maximum bandwidth for USB 3.0 is 625MB/s. Barely indeed.


RE: new standards?
By Pirks on 7/5/2011 9:33:29 PM , Rating: 2
quote:
OCZ Vertex 3 can hit 370MB/s
Four of these in RAID 0 will pwn ya puny little USB like there's no tomorrow :))) Are you even serious? I see you just have no idea about what SSD RAID is. Go do some basic reading about SSD RAIDs before replying please

P.S. Your babbling about HDD RAID is pretty funny too. Babble all ya want, I said NOTHING about HDDs, so I don't care what ya babble about 'em.


RE: new standards?
By kitonne on 7/5/2011 10:01:27 PM , Rating: 2
Copper is reaching its limits, time to go to fiber optic :) If you ever played with fiber data connections, it is not too big of a deal; true, you cannot make your own patch cables, and welding fiber optic strands require fairly pricey equipment, but fiber patch cables are way cheaper then Thunderbolt cables, and they are already made in very large numbers by many players around the world.

Number of devices or what kind of data is being moved / supported does not come into play - that layer is in software, way above the physical link represented by the Thunderbolt hardware.

BTW, SCSI was easy to troubleshoot because was daisy-chained - you could start with just ONE device, and kept of adding until you found the issue. Its bus arbitration just works, you can talk to slow devices and fast devices on the same chain with no issues. USB may be more difficult to troubleshoot if you have multiple devices going to a hub, specially when some devices are streaming, as its arbitration model is much weaker.

I do agree that USB 3.0 is likely to win, as most new motherboards / computers have a couple of ports built-in, and for most applications, cheap and good enough is THE winning combination (as compared to best and not-so-cheap)....


RE: new standards?
By Pirks on 7/5/2011 11:33:30 PM , Rating: 2
quote:
SCSI was easy to troubleshoot because was daisy-chained - you could start with just ONE device, and kept of adding until you found the issue. Its bus arbitration just works, you can talk to slow devices and fast devices on the same chain with no issues. USB may be more difficult to troubleshoot if you have multiple devices going to a hub, specially when some devices are streaming, as its arbitration model is much weaker
Gah, Reclaimer is gonna pop right now after reading this

Reclaimer, do something terrible please :)


RE: new standards?
By Bad-Karma on 7/6/2011 12:32:40 AM , Rating: 2
quote:
Copper is reaching its limits, time to go to fiber optic :)


Lightpeak with copper is just the 1.0 edition. Lightpeak was built around a fiber line but reverted to copper a couple months before Apple made the announcement about "Thunder". I don't remember if it was here or "Tom's" but there was a decent article that quoted the Intel team. They went with Fiber in this first release do to manufacturing costs.

Maybe in 2.0 we'll see the fiber return.


RE: new standards?
By Bad-Karma on 7/6/2011 12:46:44 AM , Rating: 2
quote:
BTW, SCSI was easy to troubleshoot because was daisy-chained - you could start with just ONE device, and kept of adding until you found the issue. Its bus arbitration just works, you can talk to slow devices and fast devices on the same chain with no issues.


Yeah, but do you remember all the different connectors and adapter cables you'd have to mess with if you had to intermingle different SCSI 1/2/3 systems. Then you'd get into LVDs, or external drives with no pass through (thank you so much for that headache Iomega!)

Fiber Channel and Infiniband sure make things allot easier now.


RE: new standards?
By Gungel on 7/6/2011 8:09:01 AM , Rating: 2
And don't forget that LightPeak will soon have to fight with another open standard from the PCI consortium, the 32Gbps PCIe Express Cable. A copper connect open standard available to anyone not just a few that cough up licensing fees to Intel.


RE: new standards?
By TakinYourPoints on 7/5/2011 6:11:28 PM , Rating: 2
It really depends. I can't speak to PCI-E since I don't pay attention to Mac Pros, but DDR2 and DDR3 were in their machines at about the same time as everyone else. Apple also got mobile Penryn CPUs in their Macbook Pros first, the custom packages in the original Macbook Air first, and the Harpertown and Nehalem Xeons in the Mac Pros first. They had all of these things weeks or months ahead of other PC builders. I mostly own PCs, but I recognize that as much as they may lag sometimes, they certainly lead at others.

Pushing new standards in particular can be a headache for users when it means that what they own is suddenly getting depreciated (ie - Thunderbolt, dropping old connectors for USB), but it also means adoption of better interfaces and standards much faster in the long run.


RE: new standards?
By yuhong on 7/5/2011 10:03:54 PM , Rating: 2
I think parent was referring to the PowerPC age when things were different.


Probably not Spansion or SandForce
By MrTeal on 7/5/2011 4:41:32 PM , Rating: 3
quote:
If the rumor about the new flash is true, it likely comes from Micron Technology, Inc. (MU), Intel Corp. (INTC), Spansion Inc. (CODE), or SandForce -- all members of the ONFI 3.0 coalition.


I can't see Apple relying on Spansion, their product line does match up with high volume SSD production, and Apple might not want to deal in quantity with a company just coming out of Chapter 11. SandForce doesn't make any flash, so they would be tough to buy flash from.

Regardless of whether it's soldered onto the board or routed through an external connector, both interfaces will likely still have to go through the SATA or PCIe bus, unless Apple has convinced Intel to make them a chipset that directly talks to the NAND. Performance will still be dictated by whatever controller they use in the design.




RE: Probably not Spansion or SandForce
By Motoman on 7/5/2011 4:47:56 PM , Rating: 1
quote:
Regardless of whether it's soldered onto the board or routed through an external connector, both interfaces will likely still have to go through the SATA or PCIe bus, unless Apple has convinced Intel to make them a chipset that directly talks to the NAND. Performance will still be dictated by whatever controller they use in the design.


Exactly, ultimately being one more wildly anti-consumer "innovation" by Apple...because then what do you do when one of those chips goes bad? Plop in a new hard drive/RAM stick? Nope. Motherboard replacement. Just like not giving consumers access to replace a battery or something...they'll spin it with the usual kool-aid marinade, but in the end it will be Apple consumers that take a bath in it. Again.


By MrTeal on 7/5/2011 4:56:09 PM , Rating: 2
Why would you replace the motherboard when you could just buy a new 2013 MBA?

It makes you wonder what they plan to offer for storage sizes though, and if there will still be a port for expansion. At least now there are other vendors of mSATA hard drives that you could move to if you want a higher capacity than Apple offers. If Apple goes beyond just creating a fast boot SSD on the MB to complement the other drive and completely removes the mSATA port, you're stuck with whatever they give you. It's a step back to the bad old days.

BTW- 400Mbps refers to the interface speed for the toggle NAND. That's different than saying 400MB/s, as in the title.


RE: Probably not Spansion or SandForce
By Pirks on 7/5/2011 6:11:58 PM , Rating: 2
quote:
being one more wildly anti-consumer "innovation" by Apple
Guys, just watch Motoman rolling and changing course every day, yesterday he was frothing about how ordinary and standard Apple is, "just like the others" and shit, today he's completely rotated in the opposite direction, now Apple's innovations are "bad" and shit. Tomorrow he'll forget about these innovations and will start his old dumb song about Apple being "no innovation" and shit, then he'll rotate again. And again. And again. Just like a wind sock, hehehe :) LOL


By yomamafor1 on 7/5/2011 9:15:26 PM , Rating: 2
Only two manufacturers make 19nm NANDs, Toshiba and Sandisk.

As for SSD controller, I doubt it'll be SandForce. Apple would need to inflate the price of their MBAs even more to justify the use of SandForce. My money would be on the low cost Marvell, since most, if not all of the MBA users don't know the difference between HDD and SSD, let alone one that is powered by Marvell.


way to push better technology DT...
By Conner on 7/5/2011 5:34:25 PM , Rating: 2
quote:
which requires fancy chip-laden $50 USD cables to work.

This "fancy chip" tech will allow any future faster optical cables to work with older thunderbolt/light peak computers. It's a strategy called thinking ahead.
Its a shame that sony's pissing on the standard with their own unique interface...




By TakinYourPoints on 7/5/2011 6:23:30 PM , Rating: 2
The prices don't seem out of line either. Here are copper interconnects for servers that are the same 10 Gbps, but unlike Thunderbolt they are not bi-directional, they are only good for transmitting storage data (no displays), and they cost $70-$150 instead of $50: http://www.amazon.com/s/ref=nb_sb_noss?url=search-...

BOM is a terrible way to judge value in any case. Utility and scarcity drive prices, and in this case you have incredible performance currently aimed at the niche of AV professionals, so of course it will initially cost more than USB or Firewire. Prices will drop as peripherals and TB capable PCs get more widespread, the same thing happened with USB and Firewire. It isn't like SATA, USB, Firewire, and HDMI cables always used to cost $2 at Monoprice.

That you can get this level of performance and utility (transferring data and displaying video over the same cable is pretty cool) without spending the money on fiber cards is not bad.


RE: way to push better technology DT...
By LordConrad on 7/5/2011 6:23:31 PM , Rating: 2
Sony got it right this time, not Apple. Apple changed the name to Thunderbolt because they dropped the optical and could no longer call it LightPeak


By TakinYourPoints on 7/5/2011 6:36:45 PM , Rating: 3
Except that Sony isn't using optical interconnects either. Also, the only Thunderbolt components it will work with are their own proprietary docks, unless peripheral manufacturers which are already conforming to the Intel/Apple mini-DisplayPort connectors decide to go with USB connectors as well. Problem is this statement from USB Implementers Forum last year:

quote:
"USB connectors are not general purpose connectors and are not designed to be used in support of other technology applications or standards or as combo connectors."


This is why Intel and Apple went with the open and royalty free mini-DisplayPort connector instead of the USB connectors that they first demonstrated Light Peak on. Sony must have made a deal, but in any case we'll see if this ends up being forward compatible. It certainly won't be when the cables switch from copper to optical.

Based on Sony's long and illustrious history of proprietary hardware and media formats that fall by the wayside after only a few years, I wouldn't be betting on their implementation just yet.


Remind's me of intel's Braidwood
By someguy123 on 7/5/2011 4:21:46 PM , Rating: 2
I've always thought intel's "flash-on-a-motherboard" would make sense on portables, and I'm assuming intel's the main partner on this push considering they're also the ones pushing TB and they produce their own NAND. I think it's a better design choice on portable devices to go with NAND internally with external ports for storage drives than to shove in a complete HDD inside.




By spread on 7/5/2011 6:46:27 PM , Rating: 2
It's the Macbook Air and it's designed to be as thin and light as possible. It's not a full fledged computer. It's an ultraportable.


Speed
By ipay on 7/5/2011 6:31:45 PM , Rating: 4
400 megabits? Umm no, the MB here means megabytes.

-- Brian




This edit doesn't make sense
By MrTeal on 7/5/2011 5:01:32 PM , Rating: 3
Jason, you changed the article to include this...
quote:
Apple's currently NAND supplier South Korea's Samsung Electronics (SEO:005930) and the world's second biggest producer, Toshiba Corp. (TYO:6502) are not currently members. A split with Samsung would make sense, given Apple's ongoing legal war with the gadget and component maker.


It's not Apple that calls it Toggle NAND, it's Samsung and the other manufacturers.
http://www.samsung.com/global/business/semiconduct...
Look down at the key features of Toggle NAND 2.0. IMFT, Samsung and Toshiba all make toggle flash.




czx
By gagafafa on 7/5/2011 9:17:44 PM , Rating: 2
http://www.benzlogo.com/

I tide fashion Good-looking, not expensive Free transport




By kitonne on 7/5/2011 9:40:18 PM , Rating: 2
http://www.theregister.co.uk/2011/07/05/macbook_ai...

Toshiba is putting 16 chips in one package; if the article is right and each chip is capable of 400Mbps data rate, the packaged device should do 800MBps (as in Bytes or 8 bits at a time); more likely it will be a x16 device, and do 400Mwps (words per second, with a word being 16 bits).

It looks like it will be a very dense memory, with 2 bits / cell storage (and 3 bits / cell next year); how well is it going to work in real life, we'll just have to wait and see :) The lower the geometry, the higher chances for process issues and "random defects". I would love to see PPAP data for the new chips - how many write cycles they can handle, what is the guaranteed data retention at hot, if they rely on internal ECC for data correction, what is the write and erase speed, number of blocks, demonstrated number of defects/wafer, leakage and supply current distributions, etc. All I have seen so far is the sequential READ data speed, which is not enough to get the full picture, specially for a new process and cutting-edge technology.




"If a man really wants to make a million dollars, the best way would be to start his own religion." -- Scientology founder L. Ron. Hubbard











botimage
Copyright 2013 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki