Power Profiles May Help Lower Power Consumption of Chips
Scientists at the University of Texas at Austin and the Australian National University claim to have conducted the first systematic power profiles of microprocessors
There is a chance that such profiles could lead to a decrease of the energy consumption of small devices such as cell phones both small as well as large computing installations such as data centers.
The basic idea isn't exactly new, as the research apparently focused on figuring out how much power processors draw when running different applications. However, Kathryn McKinley, professor of computer science at the University of Texas at Austin, said the scientists "did some measurements that no one else had done before."
"We showed that different software, and different classes of software, have really different power usage," McKinley said. An example would be that an application that leverages the GPS chip will cause a battery to run out of power faster than an application that does not. By itself, that is not a surprise, but McKinley, whose work was recently selected as one of this year's "most significant research papers in computer architecture based on novelty and long-term impact" by the journal IEEE Micro, said that power profiles could become a standard part of software to tell the consumer how much power a certain app will draw - which would influence the decision process whether an app will be installed or not.
- High-Speed CMOS Sensors Improve DSC Low-Light Capability
- Imperva: Anonymous Used LOIC to Attack DoJ, MPAA
- 687,000 U.S. High-Tech Jobs Lost in a Decade
- 1TB SSD Crammed Into Swiss Army Knife, Costs $3000
- Apple Patents the Ultra Wide Touchpad for Notebooks
- Windows Server 8 to Get a New File System: ReFS
- Microsoft Reports $20.9B Revenue, 525M Windows 7 Sold
- Genius Intros Wireless Mouse Without a Battery
- Ultrabooks Will Be Zapped By Thunderbolt in 2Q12
- Sony Refereshes Vaio Z Series With Color, New Hardware
- Kingston Makes Red HyperX Memory for Fashionable Mobos
- ARM Doesn't See Intel as a Competitive Threat
- Intel CEO Says That 4G Will Be Integrated into SoCs
- Diablo 3's Senior Producer Quits, Leaves Blizzard
- Guild Wars 2 Public Beta Dated, Coming Soon
- Blizzard Incorporating In-Game Ads Scheme for "Titan"
- Crucial Offers Firmware Update For Crucial m4 SSD BSOD
- MIT's 100-core CPU Will Be Ready This Year






Captain Obvious much?
oh wait, I just made another discovery, a person's daily activity wll affect his/her dietary needs...
Great, another common sense "feature" published in a "research" paper... its only a matter of time till someone at Apple sees the paper and starts drafting a patent...
[Install Free!]
*
Energy consumption
5Wh average
Do you STILL want to continue to installation?
These scientists must've been bored...........
Imagine the combined power consumption of all those stupid fart apps..
I am working on a research paper for the scientific measurement of the lameness of an application. If it is totally lame that will help someone's decision to buy the app.
This is likely the most unintelligent result of the piss poor education that I've seen in recent months. Really power profiles lol wow haven't this already been thought of before. I guess they are now only discovering what Intel, AMD, Texas Instruments, and other companies have been working on for years. As for the power consumption is dependent on load imposed by the app not "profiles" >.<
Reducing clocks in real time on a modular architecture can reduce power consumption rather nicely.
You don't say?
How about a device that tells me how lame my comment is going to be before I make it?
i hear closing the light dramastic lower your power consumption.
So..... Her method defines the needs of processing power as a preset per application, requiring a substantial amount of work to configure, whilst the big companies (intel, AMD, TI, Qualcomm etc.) use advanced algorithms to accomplish the same task on the fly, fully automated.
What is the point of her method? Cure boredom?
Simplify a brilliant piece of research to the point of making it sound absurd much?
Congratulations this is why you're a journalist, the people who most reliably let ignorance and/or incompetence ruin a good story.
To everybody who says this sounds like pointless research, it would be if it was anything like what was described in the article.
How about a device that tells me how lame my comment is going to be before I make it?
That would be quite something... However, not worth buying unless it's dirt cheap.
Really if power consumption is the problem we want to discuss then how about shrinking the space between the transistors? Wait, what do you mean we know about that already? No way we did since we didn't know about this awesome new concept of "power profiles" to reduce power usage even more... What will they come up with next!?!?
While we're on the topic of power usage why don't we talk about increasing power usage? That's more fun anyway. Lets see, we can increase clock rate, increase voltage, ignore the years (maybe decades? I'm not sure here) of power profile research and ACTIVE USE, increase cache to server levels even though little to nothing we do uses huge caches well, ask AMD to make it, get worse cooling, the list of fun goes on.
Since these people shouldn't be stupid I'll give them the benefit of the doubt and assume they did something different from the current power profile technologies, maybe they implemented it much more completely than contemporary designs.
Maybe she managed to figure it out at the chip level?? Hard coding it in the Processors would be a great option.... especially since everyone is not aware on how to setup a power saving profile for their unique rigs.... So it does make sense to have it implemented at the chip level and help save a lot of power, which is lost by sheer lack of awareness.
really....common sense would tell you that hardware profiles would either save or consume more battery life...and to think they have a degree for this