Subscribe via E-mail

Your email:

Download our latest technical papers


cta-noc-power-benefits-paper-pink-frontpage

Kurt Shuler bio

Kurt Shuler Arteris Intel TI MIT USAFAKurt Shuler is the VP of marketing at Arteris. 

He has held senior roles at Intel, Texas Instruments, ARC International and two startups, Virtio and Tenison. Before working in high technology, Kurt flew as an air commando in the U.S. Air Force Special Operations Forces.

Kurt earned a B.S. in Aeronautical Engineering from the U.S. Air Force Academy and an MBA from the MIT Sloan School of Management.

Latest Posts

Posts by category

Follow Me

Arteris Connected Blog

Current Articles | RSS Feed RSS Feed

Updated Interface Standards Table - MIPI HSI, HSIC, UniPro, UniPort, LLI, C2C

Posted by Kurt Shuler on Thu, May 17, 2012 @ 11:07 AM
  
  
  
Add to delicious  delicious  
  

I've received a lot of feedback on the "Interchip Connectivity: HSIC, UniPro, HSI, C2C, LLI...oh my!" article I wrote last year and I thank everyone for their feedback!

[Note: I updated the table below based on Ariel Lasry's feedback. Thanks Ariel!]

I have updated the table below to list the PHY required for each standard. Hope this helps!

LLI C2C standards table updated 20120517 resized 600

I also have it below as a text table for your cut-and-pasting pleasure:

Standard

Release Date

Unidirectional Throughput

Pins/Signals/ Channels

Low Latency

PHY required

MIPI HSI

2003

200 Mbit/sec

8

No

MIPI HSI PHY

USB HSIC

2007

480 Mbit/sec

2

No

USB 2.0 HSIC PHY

MIPI UniPro, UniPort-D, and UniPort-M*

2007

2,900 Mbit/sec per HS-G2 lane

5,800 Mbit/sec per HS-G3 lane

2 pins per unidirectional lane

No

MIPI D-PHY or M-PHY

MIPI LLI*

2011

2,900 Mbit/sec per HS-G2 lane

5,800 Mbit/sec per HS-G3 lane

2 pins per unidirectional lane

Yes ~80 ns

MIPI M-PHY

C2C

2010

6,400 Mbit/sec

16

Yes ~100ns

No

*     Although UniPort-M and LLI both use the MIPI M-PHY, the UniPort standard has higher protocol efficiency than LLI (97% versus 66%). This is because the primary purpose of UniPort is high throughput connectivity while the primary purpose of LLI is low latency.

You must be a MIPI Alliance member to access the MIPI standards at www.mipi,org.

Tags: , , , ,

Comments

How can you state that UniPro has a different throughput than LLI, while using the same M-PHY ? 
 
For the same number of LANEs you will have exactly the same throughput on UniPort-M and LLI. 
 
Moreover, UniPro offers today a much better protocol efficiency than LLI (97% vs. 66%).
Posted @ Sunday, April 15, 2012 5:09 AM by Ariel Lasry
Ariel, how are you doing? If you are in Japan this week, let me know. I am here and we can have some beer. 
 
 
 
You are correct on the throughput measurements. I think I may have based the UniPro numbers on older D-PHY numbers? (I can't remember.) I'll fix the table as soon as I get back from this trip (or even sooner). 
 
 
 
The protocol efficiency numbers you state are also correct. I neglected to discuss efficiency in this article because the point of this article was to highlight the new low latency options, and I didn't think about it at the time. I'll discuss it in future versions. 
 
 
 
The key point I want everyone to know is to "use the right tool for the job at hand."  
 
1. If throughput is the requirement, then an engineer has many choices including UniPro/UniPort. 
 
2. If one wants low latency for memory sharing, then the choices are more limited(LLI or C2C). 
 
3. The same two choices are available if one wants to avoid a runtime software stack. 
 
 
 
In the next version of the table I'll include more detailed examples of the throughput=lanes+gears options. 
 
Posted @ Tuesday, April 17, 2012 7:10 AM by Kurt Shuler
Hi Kurt, 
 
Yes, I think you are right. The numbers were old D-PHY numbers and you summarized nicely the key points. 
 
Looking forward to see the updates. 
 
Ariel
Posted @ Sunday, April 29, 2012 4:08 PM by Ariel Lasry
Hi everyone! I updated the table based on Ariel's feedback. When I have some time I was thinking I would post some links to some good articles that explain in more detail the M-PHY. What do you think?
Posted @ Thursday, May 17, 2012 11:11 AM by Kurt Shuler
Shouldn't SSIC appear in your table too?
Posted @ Wednesday, June 27, 2012 4:04 PM by yasser m. ibrahim
Post Comment
Name
 *
Email
 *
Website (optional)
Comment
 *

Allowed tags: <a> link, <b> bold, <i> italics