Interactive Brokers Advertisement
Home > Boards > US Listed > Computers - Semiconductors > Intel (INTC)

Interesting AMD 65nm Process Variation Analysis

Public Reply | Private Reply | Keep | Last ReadPost New MsgReplies (3) | Next 10 | Previous | Next
wbmw Member Profile
Member Level 
Followed By 28
Posts 26,118
Boards Moderated 0
Alias Born 09/11/02
160x600 placeholder
ViewSonic ViewBoard® with the Intel Unite® Solution Modernizes Meeting Rooms "Business Wire" - 10/10/2017 8:00:00 AM
Chip Makers Are Adding 'Brains' Alongside Cameras' Eyes "Dow Jones News" - 10/4/2017 7:29:00 AM
Statement of Changes in Beneficial Ownership (4) "Edgar (US Regulatory)" - 10/3/2017 5:04:21 PM
Statement of Changes in Beneficial Ownership (4) "Edgar (US Regulatory)" - 10/3/2017 5:03:05 PM
Statement of Changes in Beneficial Ownership (4) "Edgar (US Regulatory)" - 10/3/2017 4:59:36 PM
Former Intel CEO Paul S. Otellini Dies at Age 66 "Business Wire" - 10/3/2017 12:03:00 PM
Intel Offers Innovative Approach to IoT Scaling and Security "Business Wire" - 10/3/2017 4:30:00 AM
Intel Commences Compulsory Acquisition for Remaining Mobileye Shares "PR Newswire (US)" - 9/28/2017 4:11:00 PM
Intel Commences Compulsory Acquisition for Remaining Mobileye Shares "Business Wire" - 9/28/2017 4:10:00 PM
Intel Editorial: Intel’s New Self-Learning Chip Promises to Accelerate Artificial Intelligence "Business Wire" - 9/25/2017 1:00:00 PM
Intel Unveils the 8th Gen Intel® Core™ Processor Family for Desktop, Featuring Intel’s Best Gaming Processor Ever "Business Wire" - 9/24/2017 11:00:00 PM
Initial Statement of Beneficial Ownership (3) "Edgar (US Regulatory)" - 9/19/2017 7:41:58 PM
Intel Invests $1 Billion in the AI Ecosystem to Fuel Adoption and Product Innovation "Business Wire" - 9/18/2017 4:00:00 PM
Waymo and Intel Collaborate on Self-Driving Car Technology "Business Wire" - 9/18/2017 12:00:00 PM
Q2 2017 S&P 500 Buybacks Fall 9.8% from Q1, to $120.1 Billion "PR Newswire (US)" - 9/18/2017 10:00:00 AM
Current Report Filing (8-k) "Edgar (US Regulatory)" - 9/18/2017 9:14:23 AM
Andrew Wilson Elected to Intel Board of Directors "Business Wire" - 9/18/2017 9:00:00 AM
Intel Declares Quarterly Cash Dividend "Business Wire" - 9/14/2017 9:00:00 AM
Mobileye and Munich Re, US Announce Collaboration to Reduce Automotive Collisions "Business Wire" - 9/12/2017 10:00:00 AM
Google Appeals EU Antitrust Penalty -- WSJ "Dow Jones News" - 9/12/2017 3:02:00 AM
Intel Wins Reprieve on EU Antitrust Fine -- WSJ "Dow Jones News" - 9/7/2017 3:02:00 AM
EU Court Backs Intel's Appeal of 2009 Fine, in Blow to Regulator -- 2nd Update "Dow Jones News" - 9/6/2017 10:24:00 AM
EU Court Backs Intel's Appeal of 2009 Fine, in Blow to Regulator -- Update "Dow Jones News" - 9/6/2017 7:20:00 AM
EU's Top Court Backs Intel's Appeal of 2009 Antitrust Fine "Dow Jones News" - 9/6/2017 4:44:00 AM
Intel Editorial: When Will You Be Ready to Get in a Driverless Car? "Business Wire" - 8/24/2017 11:00:00 AM
wbmw Member Level  Wednesday, 07/18/07 02:46:14 PM
Re: chipguy post# 45786
Post # of 149476 
Interesting AMD 65nm Process Variation Analysis

I was looking through AMD's thermal design guide, and found some interesting results:

If you look at the IDD current of their C1 states, you get an idea of the leakage at various voltages. I looked at their current CZ (F3 stepping, 90nm) and their DD (G0 stepping, 65nm) parts at the max P-state (1.2-1.3V for 65W TDP parts and 1.3-1.35V for 89W TDP parts) and min P-state (1.1V for all products).

I pay special attention to their 3800+ part, which was their downbinned part for 90nm (therefore, highest leakage), and their higher bin parts, which presumably require binning for low leakage in order to fit in the power envelopes. Note that with the new price cuts, AMD's new 90nm downbin is a 4200+ part, which is not listed in this already outdated thermal guide. For 65nm, their downbinned part was the 3600+ part. Today, it is 4000+. Note that you can see the part name in the ID Tag.
ID Tag Process IDDC1@HighV IDDC1@LowV
ADO3800IAA5CZ 90nm 23.4A @ 1.25V 7.6A @ 1.1V
ADO5000IAA5CZ 90nm 14.8A @ 1.25V 4.8A @ 1.1V

ADA5000IAA5CZ 90nm 24.8A @ 1.35V 5.7A @ 1.1V
ADA5600IAA6CZ 90nm 21.5A @ 1.35V 5.0A @ 1.1V

ADO3600IAA5DD 65nm 32.2A @ 1.30V 11.2A @ 1.1V
ADO4000IAA5DD 65nm 26.5A @ 1.325V 7.8A @ 1.1V
ADO4400IAA5DD 65nm 24.1A @ 1.325V 7.1A @ 1.1V
ADO4800IAA5DD 65nm 17.9A @ 1.35V 4.6A @ 1.1V
ADO5000IAA5DD 65nm 16.6A @ 1.35V 4.4A @ 1.1V

I looks to me like AMD splits 65nm parts into three buckets: let's call them low, medium, and high leakage.

1. The low leakage parts are used for the 4800+ and 5000+ products. They are lower leakage than AMD's best 90nm parts. This is good news. They draw about 8% less current at 1.1V, and about 23-28% less current at 1.35V.

2. The medium leakage parts are used for the 4000+ and 4400+ products. They are somewhat as high in leakage than AMD's leakiest parts on 90nm, and definitely leakier than their 90nm median parts. The 4000+ part, for example, draws more current at 1.1V than AMD's downbinned 3800+ part on 90nm. At 1.325V, they are drawing more current than AMD's high end 90nm parts at 1.35V. This is certainly not good news, and suggest that the median of AMD's 65nm process leakage is worse off than at 90nm.

3. The high leakage parts are downbinned to the 3600+ chip. Although this part has been removed from the current lineup, it's not clear whether they are still producing these and selling them as 4000+ parts, or whether their process has improved. At any rate, these parts are insanely leaky. A 1.1V, they are drawing almost 50% more current than AMD's worst 90nm part. And good thing AMD restricted the voltage to 1.3V, because even at this voltage, the leakage towers over the entire 90nm product line.

I think these results are pretty interesting, and may explain why AMD has not been able to ramp 65nm. The leakage is killing them, and only their lowest leaking parts are able to hit 2.6GHz at 1.35V, and still maintain a reasonable power envelope.

Also, I found something interesting in one of the other AMD datasheets I was looking at. Ever wonder why AMD's Brisbane chips do so much better in idle power dissipation tests done by reviews...? It's because they enabled a new mobile sleep state on it.

Check out Table 64 on page 278. Previously, desktop chips supported no better than C1 Halt state. Starting with G-step (Brisbane), they now support C3.

Interesting that AMD has needed to start enabling mobile sleep states on their desktop parts. Intel's mobile parts support all the way down to C4E, while their desktop parts only support C1E. C4E flushes the cache and goes to a lower voltage than regular C4. C1E is also lower voltage than regular C1. Note that Brisbane also supports C1E and can hit lower idle power in that state. AMD doesn't state the voltage of that state, so I didn't bother listing it in the above tables. At any rate, their C1 is still at 1.1V, which is nice, because it lines up with current 90nm parts.

Anyway, I thought people would find this interesting.

P.S. (to Chipguy) I posted your other response on SI. Let's see if Petz responds.

Public Reply | Private Reply | Keep | Last ReadPost New MsgReplies (3) | Next 10 | Previous | Next
Follow Board Follow Board Keyboard Shortcuts Report TOS Violation
Current Price
Detailed Quote - Discussion Board
Intraday Chart
+/- to Watchlist