Embedded Systems Conference
Breaking News
Comments
Newest First | Oldest First | Threaded View
rbtbob
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
rbtbob   7/28/2012 11:24:29 AM
NO RATINGS
Hey! What makes you an expert on algorithms and whether or not they can be effective? :-)

David Brown
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
David Brown   7/25/2012 10:30:26 AM
NO RATINGS
The whole idea of making this data-dependent is silly and unnecessary. If it were to make sense, then you would be better dealing with the data at a higher level. For example, if you want to store ASCII data efficiently, then compress it first - saving far more than 22%. There is an extremely simple way to get a similar power saving (over 23%), albeit at the cost of extra bits. For each 8-bit byte, you have an additional "invert" bit. For each data read, you use this bit to see if the data bits should be inverted. When writing, you consider both the non-inverted and inverted versions of the data, and choose whichever uses least power. Simple, fast, low-power, and independent of the data.

resistion
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
resistion   7/25/2012 12:54:30 AM
NO RATINGS
I forgot to cite the STT MRAM energy reduction reference: here it is: P. Zhou et al. ICCAD 2009, "Energy Reduction for STT-RAM Using Early Write Termination" (U. of Pittsburgh)

Kristin Lewotsky
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
Kristin Lewotsky   7/24/2012 10:00:47 PM
NO RATINGS
Thanks for your comments. That's interesting. Based on the paper, I'd understood that having Er/Es greater than 1 was what delivered the energy savings. I'll have to reach out to the authors and see what they say.

Semyon Savransky
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
Semyon Savransky   7/24/2012 7:41:38 PM
NO RATINGS
The "read before write" approach for is well-known for PCM for a while (at least I recall it from ~ 1985) but effective algorithm can be useful.

resistion
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
resistion   7/24/2012 4:11:01 PM
NO RATINGS
For STT MRAM, Er/Es ~1, and a similar strategy has been proposed.

resistion
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
resistion   7/24/2012 2:51:15 PM
NO RATINGS
But I presume it could work for other Eres/Eset ratios. It's merely an algorithm fit.

Kristin Lewotsky
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
Kristin Lewotsky   7/24/2012 2:21:13 PM
NO RATINGS
Actually, no—the method works in no small part because the energy required to reset a bit (Er) is roughly twice that of the energy required to set a bit (Es), while the read energy is negligible. It's a nifty bit of work.

resistion
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
resistion   7/24/2012 11:37:59 AM
NO RATINGS
Looks like this approach is fairly generic rather than benefitting PCM specifically.

R G.Neale
User Rank
Author
re: Tech Beat: Codewords cut PCM energy needs
R G.Neale   7/24/2012 8:52:19 AM
NO RATINGS
Kristin- The RICE authors also raise the subject of multi-level cell (MLC-PCM) and cite the IBM-MLC. It would certainly be interesting to hear their view on how their memory overhead would be added to IBM’s drift compensating codeword memory overhead and methodology. In that respect, they might also need reminding that the write process for the multi-level cell and some earlier MLC work involves starting with a cell in the reset state and reaching the required data state resistance value by iterative steps. How many more memory codeword solutions to PCM memory problems can be added before in becomes a write never memory (WNM) because there is no room for the data?



Radio
LATEST ARCHIVED BROADCAST
As data rates begin to move beyond 25 Gbps channels, new problems arise. Getting to 50 Gbps channels might not be possible with the traditional NRZ (2-level) signaling. PAM4 lets data rates double with only a small increase in channel bandwidth by sending two bits per symbol. But, it brings new measurement and analysis problems. Signal integrity sage Ransom Stephens will explain how PAM4 differs from NRZ and what to expect in design, measurement, and signal analysis.

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Most Recent Comments
radar1093
 
Loser99
 
MeasurementBlues
 
David Ashton
 
samlebon2306
 
perl_geek
 
Clive
 
Clive
 
junko.yoshida
Like Us on Facebook
Special Video Section
The LTC®6363 is a low power, low noise, fully differential ...
Vincent Ching, applications engineer at Avago Technologies, ...
The LT®6375 is a unity-gain difference amplifier which ...
The LTC®4015 is a complete synchronous buck controller/ ...
10:35
The LTC®2983 measures a wide variety of temperature sensors ...
The LTC®3886 is a dual PolyPhase DC/DC synchronous ...
The LTC®2348-18 is an 18-bit, low noise 8-channel ...
The LT®3042 is a high performance low dropout linear ...
Chwan-Jye Foo (C.J Foo), product marketing manager for ...
The LT®3752/LT3752-1 are current mode PWM controllers ...
LED lighting is an important feature in today’s and future ...
Active balancing of series connected battery stacks exists ...
After a four-year absence, Infineon returns to Mobile World ...
A laptop’s 65-watt adapter can be made 6 times smaller and ...
An industry network should have device and data security at ...
The LTC2975 is a four-channel PMBus Power System Manager ...
In this video, a new high speed CMOS output comparator ...
The LT8640 is a 42V, 5A synchronous step-down regulator ...
The LTC2000 high-speed DAC has low noise and excellent ...
How do you protect the load and ensure output continues to ...