Performance is cool: Exploring processors and energy efficiency

Virtualization is a nice place to start.

Terminal services or Citrix is another place. If developers can make their code only require one instance of common binaries in terminal services, that would make Terminal services far more scalable.

Intel Corporation is challenging PC designers and manufacturers to think sexy, stylish and small as they design the next generation of home PCs optimized for multimedia entertainment and powered by Intel? Core?2 Duo processors.

The Intel Core Processor Challenge is open to PC designers and manufacturers worldwide. Intel will award the grand prize winner with up to $300,000 to help enable the mass production of the winning system and $400,000 for co-marketing activities with Intel around the winning system. The first-place winner will receive up to $300,000 to help enable the mass production of their winning system.

See details at http://www.intel.com/pressroom/archive/releases/20060926corp_a.htm

The fan is a necessary evil, but boy would I like to get rid of it!

HTPC's are the biggest pain point, IMHO. The noise issue is huge when you're trying to get the PC into the living room. Heat is really secondary to noise though.

But I really would prefer cooler machines in corporate desktops and laptops. I think we all know uncomfortable feeling of using a laptop in your actual lap for too long.

That's really Apple's problem (as well as their ISVs) to deal with, not Intel's, to be honest.

J.Ja

How much slower are they? Do you have specific numbers for specific applications?

It is not unusual to see 50% of native performance with translation. If the new processors are 2x the native performance of the old ones, then the translated binaries will run at the same speed as they did on the old processor.

Of course, the performance varies from application to application and the amount of OS and I/O activity will have an impact on performance.

First of all, it could be a bogus device driver doing this and not a problem with the actual hardware itself.

Second, it could be a hardware prolem but it may not be related to heat. But to check your system temperature while in Windows, try the following tool.

http://fileforum.betanews.com/detail/MotherBoard_Monitor/929820238/1

That is a very good question. If the energy costs are borne by facilities, there is very little motivation for the data center manager to procure energy efficient systems if the cost more to buy.

For the overall organization this would be short sighted. In fact, by 2010 IDC expects that for every dollar of new server spending, an additional $0.70 will be needed for power and cooling. This is a dramatic rise from the rate of $0.50 in 2005 and $0.21 in 2000.

The group I work in manages the reporting servers (Oracle Reports, Oracle Discoverer, and SAS) and heating/cooling are nowhere on the checklist for ordering servers.

I assume somebody somewhere in IT probably does worry about that issue, I'm sure it gets factored somehow into the ordering process. But when we as the front-line organization request upgraded or replacement hardware, all we're concerned about are the computer specs (RAM, disk space, etc). We're not at all involved in determining the appropriate power-management specs for our servers. We worry about meeting the data needs of our reporting customers, and let the tech services team worry about the power consumption issues.

Thanks for a very interesting post!

There is no standard way of measuring energy efficiency. Mobile clients already deliver the lowest PC energy costs, so users with a great focus on lowering energy costs will identify mobile PCs as the preferred client solution. For the desktop PC, energy efficiency mainly translates into the energy cost required to deliver a required level of performance and capability.

As energy efficiency becomes an interesting element when comparing PCs, it is important that the industry develop a meaningful and broadly adoptable metric that enables a person to make an informed buying decision. There are, however, some challenges with measuring energy efficiency and energy efficient performance. Although, historically, the industry has developed a number of standard practices for gauging performance like SYSmark*, SPEC* CPU as well as battery life for the notebook PC using MobileMark*, there is currently no standard methodology for evaluating client performance coupled with its associated energy cost. The whitepaper at http://www.intelcapabilitiesforum.net/EEP outlines a proposed methodology for evaluating both the performance and AC energy efficiency of the client platform that may serve as an interim solution until an industry developed solution becomes available. In addition to this proposed methodology, Intel will fully participate with industry standards organizations to develop a consortium-based approach to this area of platform evaluation.

For example, Intel has measured the following performance and energy cost for the Intel? Core? 2 Duo E6700 processor with integrated graphics on the Intel? Express Intel? DG965SS motherboard. For full system configuration details, please see http://www.intelcapabilitiesforum.net/EEP?s=12

Note that both Enhanced Intel? SpeedStep? Technology and Minimal Power Management are enabled.

Energy Cost per KWh $0.10

Avg. System Power Idle (watts) 62.4

Avg System Power Standby (watts) 3.4

Avg. System Power OP (watts) 69.4

Avg. System Power ICC (watts) 75.5

Time to Complete - OP (minutes) 47.316

Time to Complete - ICC (minutes) 39.433

Time to Complete - Total (minutes) 86.75

Idle Time After Run (minutes) 15

Standby Time After Run (minutes) 18.25

Performance tests and ratings are measured using specific computer systems and/or components. Actual results may vary.

+++++++++++++++++++++++++++++++++

The previous post pointed to

http://www.tomshardware.com/2006/09/25/green_machine/page4.html which also says:

Author's Opinion

Were Athlon 64 X2 Energy Efficient processors to reduce average power use by 10 W, the average 10 hour per day office system would cost only few dollars less per year to operate.

+++++++++++++++++++++++++++++++++

Energy-efficient performance (EEP) is the intersection of great performance, expanded capabilities and energy efficiency. One needs to compare the energy efficiency of similarly configured systems that deliver similar performance.

You may want to look at reports at other sites.

http://techreport.com/reviews/2006q3/core2/index.x?pg=2 says:

AMD answers with Energy Efficient Athlons:

Anticipating better power efficiency from Intel's new desktop processors, AMD has begun offering Energy Efficient versions of many of its CPUs for the new Socket AM2 infrastructure. Much like the Turion 64 mobile processor and the HE versions of the Opteron server chips, these Energy Efficient Athlon 64s have been manufactured using a tweaked fabrication process intended to produce chips capable of operating at lower voltages. Making these more efficient chips isn't easy, so AMD charges a price premium for the Energy Efficient models that averages about 40 bucks over the non-EE versions.

http://techreport.com/reviews/2006q3/core2/index.x?pg=16 says:

AMD has made substantial progress on this front with its new Energy Efficient processors. Under load, the Athlon 64 X2 4600+ EE system pulls about 20W less at the wall socket than the stock X2 4600+ system, and the 35W-rated X2 3800+ system draws less power than anything else we tested. Still, even these new CPUs can't match the performance of the Core 2 processors that are in the same neighborhood in terms of power draw.

Whoa. Performance is way up with the Core 2 processors, and power draw is way down. The Core 2's mythical "performance per watt"?which is actually a rather slippery thing to quantify?has gotta be the best on the market. The Core 2 Duo E6700 outperforms the Athlon 64 FX-62 more often than not, yet the E6700-based system draws 74 fewer Watts under load.

http://techreport.com/reviews/2006q3/core2/index.x?pg=18 says:

The PC industry can also breathe a collective sigh of relief about power and thermal issues now that Core 2 has arrived. Intel finally has a firm handle on those problems. These processors consume less power?and thus produce less heat?than desktop Pentiums have for quite a while. The E6700 system's total power draw when fully loaded was 156 W, only 14W more than the Pentium Extreme Edition system drew while sitting idle. What's more, even the high-end Core 2 processors' power use was in line with that of the Energy Efficient versions of the Athlon 64 X2. That leaves room for many good things to happen, from less expensive cooling systems to quieter, smaller enclosures and even some righteous overclocking. Combine the low power draw with the performance we've seen, and the Core 2 is clearly the most energy-efficient desktop processor around.

You also asked about server performance and power. We have worked with independent thrird parties on measuring actual energy consumption under several workload coditions compared to the competetion.

You can find the results at:
http://www.principledtechnologies.com/clients/reports/Intel/Intel.htm

SPECjbb2005 performance and power consumption on Intel- and AMD-processor-based servers (May, 2006; 498KB)

64-bit SunGard ACR financial workload performance and power consumption on Intel- and AMD-processor based servers (May, 2006; 350KB)

64-bit Black-Scholes financial workload performance and power consumption on Intel- and AMD-processor based servers (May, 2006; 290KB)

SPEC CPU2000 SPECint_rate_base performance and power consumption on Intel- and AMD-processor based servers (May, 2006; 350KB)

WebBench performance and power consumption on Intel- and AMD-processor based servers (May, 2006; 203KB)

Windows Media Load Simulator performance and power consumption on Intel- and AMD-processor based servers (May, 2006; 726KB)

64-bit financial application-based workload performance on Intel- and AMD-processor based servers (May, 2006; 337KB)

Core 2 Duo and Pentium D are both dual core and they support 64 bit capability. Hyperthreading is only turned on in the Pentium Processor Extreme Edition product family.

Pentium D is based on the older Netburst architecture, while the Core 2 Duo is based on the new Intel Core Microarchitecture.

In general, the Core Microarchitecture delivers higher performance at lower power compared to the Pentium D processor. The exact numbers will vary with the specific model numbers and workload.

For a feature comparison try http://indigo.intel.com/compare_cpu/default.aspx?familyID=1&culture=en-US

However, if you already have an existing PC with a slower Pentium D or Pentium 4, you might be able to get higher performance by upgrading to the fastest available Pentium D processor. Check with your system vendor and/or motherboard supplier to see if your system can be upgraded.

Let me know me if you have any follow up questions.

BTW, what is a 478?

"Everyone else has zero incentive to turn the lights off. Power for servers doesn't come out of IT. They purchase UPS's and power systems based on what hardware they have, not the other way around. Facilities pays the power bill, and they have zero control over it."

You nailed the problem! We even had people balk that Microsoft is adding group policy control of power savings in Vista. People freak out if you try to shut down their machines. A lot of people still believe in the myth that turning off the computer is somehow bad for it. Human behavior and ignorance is the biggest problem.