I’m constantly asked: "Will Thunderbolt replace USB 3.0?"

Thunderbolt and USB 3.0 will co-exist & live long, fruitful lives

As reported previously, Apple filed and recieved a patent for a new iPhone or iPad connector which looks  like a Thunderbolt connection. You can see in the picture that it includes connections for USB 3.0 and dual port Displayport.

Apple received the patent in April 2011.  Since it takes at least 3 years to get a patent, we can guess that Apple has known since at least 2008 (probably 2007), it would support USB 3.0.

Apple may or may not have had knowledge of Thunderbolt in 2007.  I can see a world (in 2007) where Apple knew it would want it’s iPad tablet to drive a bigger monitor through DisplayPort.  That could be the original reason for dual DisplayPort and not Thunderbolt.  I can only guess.

Where am I going with this?

1. Apple uses the PCs chips found in most PCs
2. Most PCs will move to Ivy Bridge
3. Ivy Bridge will be found in most PCs
4. Future Apple PCs will likely use Ivy Bridge
5. Ivy Bridge has USB 3.0
6. It’s highly likely Apple will support USB 3.0
7. Apple patents indicate they’ve planned all along to support USB 3.0

So if you are still with me, Apple will definitely support USB 3.0

Back to the question

Will Thunderbolt overtake USB 3.0?

Pros:

1. Faster than USB 3.0 today
2. Proven
3. On every Apple mobile PC

Cons:

1. Closed Standard restricts the ability to integrate.
2. Only one company provides discrete chips

The one way Thunderbolt could overtake USB 3.0 is if Thunderbolt opens up as a standard.

USB 3.0 will have more than 100 design starts for SoC integration

The USB-IF has something like 4000 members. By the end of this year, there will be about 200 different USB 3.0 Design starts.  Thats lots of people innovating products around USB.

Thunderbolt is closed.  It will stay closed.  This video explains why:

So the Innovator that designed Thunderbolt makes money from customers like Apple, Acer, and Lenovo that ship Thunderbolt.  Big piles of money.  People that need/want Thunderbolt pay for Thunderbolt.

The innovator promotes USB 3.0 for mainstream ease of use.  Consumers are happy.

Thunderbolt closed.  USB open.

Everyone is happy.

Subscribe

http://blogs.synopsys.com/tousbornottousb/

• To subscribe, click on this link: http://feeds.feedburner.com/synopsysoc/ToUSB

Try and Subscribe.

I dare you.

Apparently we’ve sold a bunch of DDR.

There was cake.

I ate 2 pieces.

PC Chipsets with integrated USB 3.0 (from the Innovator/inventor of USB) started shipping at the beginning of April for reviews.

Here’s a table from Anandtech clearly showing there will be 4 USB 3.0 ports!

They’ve added 4 USB 3.0 ports and to the existing 14 USB 2.0 ports.

If you look at your laptop or desktop, you won’t see that many USB 2.0 ports on the outside because some are used internally.  They connect to a 3G modem or a card reader or ExpressCard slot.

ASUS, HP, Samsung, Toshiba, and others have already announced they will have laptops based on the Ivy Bridge chipset, just Google Ivy Bridge PC and you’ll get all the models

It turns out that AMD has been shipping a motherboard also.

I found this article comparing the performance of the Ivy Bridge and AMD integrated chipsets against NEC, VIA, and ASMedia USB 3.0 Host controllers at http://techreport.com/discussions.x/22775.

Here’s part of the Graphs from that Report.

  From http://techreport.com/discussions.x/22775

As expected, the Ivy Bridge Chipset performs faster, it’s the Blue bar.

I’m going to make you go to the actual article at techreport.com to see what the other chips and integrated chips sets are on the graph because I think the website deserve the hits.

Integrated is Faster

Ivy Bridge is faster because it’s fully integrated:

The stand alone chips used in the NEC, VIA, and ASMedia Hosts can NOT achieve faster speeds because they are limited by their PCIe Gen 1 x1 connection to the motherboard.   PCIe Gen 1 x1 can only go up to 2.5 Gbps, and it’s less in a system where many PCIe peripherals are using the PCIe bus.  Ivy Bridge shouldn’t suffer from this because it is fully integrated into Ivy Bridge and probably has at least a PCIe Gen 1 x2, x4, or even x16 to make sure there is plenty of bandwidth to move the USB 3.0 data in and out of the system. 

For detail on factors that reduce or increase USB 3.0 performance read this blog entry.

I’m going to make you go to the actual article at techreport.com to see what the other chips and integrated chips sets are on the graph because I think the website deserve the hits.

Why the performance isn’t even faster (maybe)

Performance Note: the report does NOT tell us what kind of USB 3.0 Drive they used for testing. I can actually guess which Flash Drive they are using based on the max read speeds that I see, but that will be for a later blog entry.

I think the throughput is limited by the speed of the actual USB 3.0 Hard Drive or USB 3.0 Flash drive being used.

1. If it’s a USB 3.0 Hard Drive, it probably uses a bridge chip from USB 3.0 to a SATA 3 Gigabit/second (Gb/s). This means the maximum Read speed would be near 300 MegaBytes per second (MB/s) or about 3 Gb/s. 
2. If it’s a USB 3.0 Flash Drive, it’s speed will be limited by the quality of the Flash inside the drive.  The fastest flash speeds we’ve ever seen is about 300 Mb/s with a $600 SuperTalent flash drive that arranged 2 banks of flash in a RAID configuration.   At $600, this isn’t a really a consumer product.

   Basically, it was fast because it used the fastest Flash memory, lot of it (128GB), and arranged it in two pieces for simultaneous access to both pieces.  I don’t consider this to be a commercial product, and it didn’t sell that many because it cost more than a Hard Drive, but it gives us an idea of what performance is possible with Flash Memory.

USB 3.0 in Tablets

Apparently Ivy Bridge is also targeted at Tablet PCs with Win 8.   This makes Ivy Bridge the second chipset with integrated support along with TI’s OMAP 5 demonstrated at CES 2012 in January.

This confirms Rahman Ismail’s comments (USB-IF CTO) correctly stated USB 3.0 in tablets and smart phones will be appearing in 2012.

Subscribe

This Blog Address moved to:

• http://blogs.synopsys.com/tousbornottousb/
• To subscribe, click on this link: http://feeds.feedburner.com/synopsysoc/ToUSB

Reader Mail

Thanks for reading this blog.

Comment below or send me an e-mail. Maybe if it’s insightful or insulting, I’ll post it here.  Either one works.

Reader Mail Below

> From: Name Withheld
> Sent: Thursday, April 19, 2012 5:35 PM
> To: Eric Huang

> Subject: Love your blooper on youtube!

> I just KNEW you were human!

My response:
> Was I a nematode before?

His Response:
> Correct

(Nematode = Worm)

Let me know if you know any spam bots.  I’ll send you his e-mail address…

It’s finally official, Windows 8 will support SuperSpeed USB 3.0.  Microsoft’s blog really lays out all the virtual and real hardware testing that Microsoft has to go through to make sure the drivers work, and work well. with billions of USB products.

Microsoft hasn’t announced a Windows 8 availability date, but we know from IDF Beijing that the PC Chipset makers will release USB 3.0 support in 2012.  So we can speculate that Windows 8 and these PC Chipsets will ship at the same time.  Links to the IDF Beijing reports that April Blog entry.

Interestingly, Microsoft uses a combination of a virtual approach and hardware testing.

The Virtual Approach looks similar to our Virtual Platforms.

The ASSP looks like a flash/thumb drive.  It has a USB 3.0 Device Chip in it.  It basically exercises many different USB transfer types varying parameters like packet sizes.  You can see the Microsoft USB Test Tool (MUTT) with it’s creator David Hargrove.  If I can find purchasing info, I’ll post it below.

David Hargrove and Microsoft USB Test Tool – Picture from Microsoft Blog Site

USB 3.0 Host Chip ASPs

I made an error in the graph and pricing in the last entry (yesterday).  My apologies.

In fact, the January time frame I reported that while Digitimes had some source saying prices would drop to $1.70-$1.80, in fact Host ASPs remained up around $2.35 per chip.

What Digitimes is saying is that prices will drop to $1.20 next year, but now Renesas/NEC have prices up at about $2.00 per chip while ASMedia is down at about $1.50 per chip.

“Renesas reportedly has notified its partners that it plans to lower prices for its 2-port controller chips to below US$1.20 from about US$2 currently. Comparable 2-port chips from ASMedia are currently available at US$1.50-1.70.” DigiTimes August, 19, 2011

I’ve corrected the graph here.

So I’ve lost any credibility I had in the community, so clearly you want to subscribe.

SUBSCRIBE

Send this URL onto your friend. and tell them to
Subscribe to this Blog, one option to subscribe is as follows:

• Go into Outlook
• Right click on “RSS Feeds”
• Click on “Add a new RSS Feed”
• Paste in the following “http://feeds.feedburner.com/synopsysoc/ToUSB?format=xml”
• Click on “Accept” or “Yes” or whatever the dialogue box says.