TG Daily

Munich (Germany) and Westlake Village (CA) - Connect any USB 2.0 device to your notebook and lose more than one hour of battery time: Tom's Hardware Guide's tests of a Windows-based Intel Core Duo mobile processor platform revealed a serious power consumption issue that, according to Intel, is caused by a Microsoft driver bug - a bug that has been known by Microsoft for some time, but kept from the public eye until today.

UPDATE 14 Feb 2006: Microsoft to release patch for power drain bug

In Tom's Hardware Guide's review early last week of a Core Duo-based notebook with Intel's 945 PM chipset and ATI's dedicated X1600 GPU, our test engineers reported stumbling across a mysterious power consumption issue. The unit tested showed abnormally high power consumption, which at first, they believed could be attributable to the Core Duo processor itself, with the high-performance graphics system being an alternate candidate.

But engineers with whom we consulted later in the week indicated that the issue may lie with the chipset of Intel's new dual-core platform, which previously was code-named "Napa." We spoke with representatives of major notebook computer manufacturers, all of whom asked us not to reveal their names, but all of whom said this particular issue has prevented their systems from achieving the goal of four to five hours of battery life which Intel had led them to expect. After a series of extensive tests involving battery benchmarks, Tom's Hardware Guide's Munich Labs engineers were able to observe the precise power consumption problem involving the Core Duo 945 GM chipset.

Our assessment and observations, coupled with extensive consultation with engineers familiar with Intel dual-core architecture, plus information we received today from Microsoft, have led us to conclude that an anomaly in the way the currently available version of the ICH7-M Southbridge communicates with Microsoft's ACPI driver, is at the heart of the power drain issue.

Intel ICH7-M Southbridge

TG Daily and Tom's Hardware Guide have been in communication with Intel representatives throughout this week about our findings described in this article. Late Friday, Microsoft acknowledged to TG Daily - via the hands of Intel - that they believe the problem our engineers observed to have been caused by a misbehaving driver included in Windows XP SP2 - specifically, the Advanced Configuration and Power Interface (ACPI) driver, which is part of the operating system's power management scheme for USB 2.0.

The delicate issue about this bug, provided to Microsoft partners under confidentiality guidelines as Knowledge Base article KB899179, is the fact that it has been known to be a "problem" at least since 12 July 2005. It is unclear at this time why Microsoft has not issued a patch or an advisory the public to fix or explain this bug. So far, Microsoft has not responded to our inquiries and we are still waiting for a comment on this power drain issue.

The Benchmark: Core Duo notebook loses 76 minutes of battery time

Tom's Hardware Guide's initial comparisons used two notebook computers from the same manufacturer, using identical components but with key exceptions: The first unit used Intel's older Sonoma platform, with a single-core Pentium M 750 (1.86 GHz) processor, and a 915 GM chipset with the ICH6-M Southbridge. The second uses the new Napa platform, with a Core Duo T2400 (1.83 GHz) processor, and a 945 GM chipset with the ICH7-M Southbridge.

In our first test, Tom's engineers ran Mobilemark 2005 - Reader Benchmark on both machines, with no external USB device connected to either unit. The Sonoma unit clocked in at a running time of 3 hours, 08 minutes; the Napa registered 4 hours, 24 minutes. In ordinary circumstances, the Napa system does run 44% longer than the Sonoma system, which is well within the glowingly positive estimates provided throughout last year by Intel.

For our second test, our engineers connected a USB 2.0 external combo drive (IBM DN-224E), and ran Mobilemark 2005 - Reader Benchmark on both systems again. The external drive was powered by its own external power adapter, which means it should avoid draining any power from the system to which it's attached. To keep the USB port awake, we continuously read data from each device using a standard electricians' tool. CPU load on both notebooks, according to numbers registered by Windows Task Manager, was about 2 - 3%. Current drain on the USB port of both systems was about 170 mA, which translates into a negligible power consumption of about 1 watt.

The second test yielded the more astonishing result: While the Sonoma notebook only lost 17 minutes of battery running time (2:51 total) with the drive connected, the Core Duo notebook's running time decreased by a stunning 76 minutes (3:08 total). In short, the power drain from the device attached to the Napa system reduced running time by 347% (76 minutes / 17 minutes) of the power drain for the same device attached to the Sonoma system. With a USB device connected, the Napa notebook with the USB device connected ran only 8% longer than the Sonoma rig.

In a second battery of tests, we were able to replicate these same results for both platforms. However, our engineers aborted further tests using these two systems, after the charge regulator in our Asus W5F Napa notebook failed to work properly. We have since been unable to secure a replacement system that would meet our requirements of matching the specifications of the Sonoma system, except with the newer CPU and chipset.

Our USB 2.0 device: IBM DN-224E

Although the Napa platform notebook system used in our tests was, admittedly, a sample, it's important to note it is not what we typically understand to be a "prototype." Its contents very likely represent the technology that will soon be shipping in volume through retail channels, and which you'll soon be able to purchase for yourself. The Napa system in our tests was received by our labs two weeks after Intel's announcement of the Core Duo processor and related platform; with the official production date so close on the horizon, it is simply improbable that this system is very different from a mass production unit. Clearly, our concern was that Intel could find itself shipping a major product that contains a major problem.

The scheduler that won't stop

Earlier this week, Greg Myers, Intel's product marketing manager for Core Duo, told us "We believe we have properly implemented the USB 2.0 power management across Centrino, including support for the chipset family 945 and ICH-7." He argued that the guidelines for USB 2.0 technology apply to multiple devices simultaneously. "Each single piece has to power-manage USB 2.0 correctly," he said. "If one of the pieces is not correctly implemented, the whole system will be impacted."

The specific piece which Intel and Microsoft now say is responsible for the power drain is Windows' Advanced Configuration and Power Management Interface (ACPI) driver, which is a software component provided through the operating system. This is the component responsible for instructing a device to power down when it is not in use. Under ACPI, there is a series of sleep states that the operating system's ACPI driver recognizes and supports. Hardware manufacturers such as IBM must build their external devices to support the sleep states that the ACPI driver recognizes; manufacturers cannot expect to implement their own proprietary drivers, while still expecting the computer to be able to fully support all USB devices that could be connected to it. Microsoft provides the ACPI driver for Windows which manufacturers rely upon as the reference driver - the software component with which their hardware must deal, if their setup is to be expected to work.

Since Microsoft's drivers are now believed to be directly involved, then all of Apple's upcoming MacBook Pro systems - which use the Core Duo processor and 945 chipset - should be unaffected by this issue. We have yet to attain access to a MacBook Pro to verify this.

To make certain the power drain problem was unrelated to the external drive, Tom's engineers ran the same battery of tests as with the Asus notebooks, on a pre-production sample of MSI's Megabook S270. This is an ultra-portable notebook system with an AMD Sempron 3000+ CPU, and which uses ATI's Radeon RS480/SB400 chipset. Its battery capacity is 26 Wh (design capacity 30 Wh); and CPU load, based on a read of Windows Task Manager, was 3 - 4%. Current drain on the USB port was 170 mA - the same figure as for both Asus systems.

In the first run, using Mobilemark 2005 and with no external USB device connected, battery life was clocked at 1 hour, 26 minutes. For the second run, we connected the same IBM external drive we used before. In that instance, battery life only diminished to 1:12 - a 16% drop, compared to the 44% drop measured when the same drive is connected to the Core Duo/945 GM unit.

Late this afternoon, representatives from Intel contacted Tom's Hardware Guide's Munich Labs, stating they believed the significantly reduced battery could have occurred because hardware in the Napa platform may be exhibiting problems reaching its deeper sleep states - which are regulated by Microsoft's ACPI driver. Then, at about a quarter-to-five Eastern time, Microsoft apparently lifted a confidentiality agreement on its Knowledge Base article and Intel was able to provide its contents to TG Daily.

The issue, according to Microsoft, concerns the asynchronous scheduler component - a part of the USB 2.0 driver that determines when devices can access local memory. With the revision to that driver implemented in Windows XP Service Pack 2, the scheduler can inadvertently be left running. As a result, Windows' internal task scheduler (a separate item) treats the asynchronous scheduler as a running process involving the attached device, and thus stops itself from ever giving the processor the signal to power down, or power lower - to slip into one of its ACPI sleep states. Because the scheduler is running, Windows thinks the system is continually busy. As a result, the computer can use more battery power.

The open question: Why Core Duo?

While Microsoft's Knowledge Base article is conclusive and states that the current ACPI driver is a "known problem," we are still scratching our heads why only our Core Duo/945-based platform provoked the error and not the Sonoma or the AMD/ATI Radeon platforms. Also, according to the Windows Performance Monitor utility, the Core Duo notebook showed a similar deeper sleep performance as the single-core Sonoma platform: The dual-core notebook remained in the "C3" sleep state for 99.82% of the time during the USB 2.0 device was connected, the single-core notebook 99.84% - which may indicate that there was no difference of when certain sleep staes were reached. However, we were not able to confirm whether the dual-core notebook in fact reached one of its two remaining "deeper sleep" states (C4 and C5).

For now, we are told that the solution to the matter could be simple: a simple addition of a single key to the Windows System Registry. At the time of this writing, neither Microsoft, Intel, nor Tom's Hardware Guide has had the opportunity to attempt this suggested fix to our Napa system. The engineers of Tom's Hardware plan on testing a Napa notebook with modified Registry settings and we will publish our findings as soon as they are available.

Intel's Greg Myers did not have an explanation, why we have seen two different effects of USB 2.0 on our single- and dual-core platforms. "This is an issue that is not specific to Napa. We expect this problem to occur on any system - whether it is a Sonoma or any other system. We are still not sure why you have been seeing a different result in the Sonoma notebook," he said.

However, the reason that one notebook is impacted by the bug and the other is not, may be be that there are several more components that can cause excessive power consumption through USB 2.0. "It is just one piece of the puzzle," Myers said. "All the vendors have to design their products according to the power management specifications. If one componet is not working properly, the whole system may be impacted."

What about the consumer?

Buyers of Core Duo-based systems may still find themselves encountering this problem, however, as systems with Windows XP SP2 installed are already making their way to retail channels. When they do, they may find themselves implementing a fix, which is not necessarily easy for an amateur user, unless Microsoft makes it available as a "patch" that installs automatically.

According to Myers, Intel has asked Microsoft "to fix the registry to make the problem go away." In the meantime, he said, Intel is working to fix the issue by itself: "We are doing additional work on our end to see if there is anything we can do to overcome the same challenge Microsoft is having," Myers said.

If the patch can be as simple as a modification to the registry, the question arises why Microsoft hasn't issued a registry patch. We were not able to reach Microsoft representatives to answer this question and Myers said that he was unsure why the problem has not been fixed so far. "Battery life of Napa is a big deal for us. Microsoft said it is investigating and looking into this issue."

Still, Myers could not guarantee that there will be a fix for the current bug: "I am not certain at this point. But battery life impact is pretty dramatic and significant. It is something we have asked our engineers to put a high priority on. At this time, we may be able to solve the problem through drivers, firmware and software. If there is no solution from a software persepctive, we will look into hardware fixes for future platforms to prevent this issue."

In any case, one may wonder why Intel announced and apparently is shipping its platform, while this USB 2.0 issue still exists. According to Myers, Intel believes it has "done a great job job in reducing overall power consumption." The existing problem is "more of a USB problem" than a processor or chipset issue, he said: "It's an issue that we all [vendors] have to try to solve."

It's difficult for anyone to say, at this time, whether Intel or Microsoft is more at fault than one another; and in the end, the answer to that question may be insignificant. Not every element of hardware in an Intel chipset is made by Intel, and the efficiency of all that hardware is largely dependent on drivers in Microsoft Windows. But when an Intel chipset works well - as it so often does - the company rarely shares credit with its outside suppliers, the names of which are buried deep within Intel's engineering literature, generally within the small print, sometimes headed by a phrase such as, "If you have a problem, please call..." In the end, Intel's message remains "Intel Inside," as it should be. We're heartened to learn that the companies that so readily seize claim for any and all factors that lead to their success, now appear ready to be responsible for what could have become a monumental problem for Intel, had the ultimate cause not been discovered: We still have to verify this claim, but we remain optimistic that the solution Microsoft provides at this time will solve the power drain issue.

Our focus as technical journalists who would ourselves purport to be responsible, should not be upon the interests of corporations, or whether one corporation appears to champion over the other, or whether they look upon us favorably, or whether any of them are capable of saving face in the end. The principal issue of our concern is making sure this power drain does not affect you. When a customer such as yourself spends as much as $2,000 on a top-of-the-line system, you should be able to expect fair and honest treatment from the companies that profit from your generosity. We'll stay on top of this issue, continuing to test our systems and cooperate with manufacturers and developers, so that the responsible parties have a clear channel to you - their customers, and our readers - so that someone's inadvertent oversight doesn't impact your work and your life.