|
|
楼主 |
发表于 2019-11-6 15:18:20| 字数 8,377| - 中国–北京–北京 联通
|
显示全部楼层
感谢您的回复 我在贴吧曾看到过一种说法，调节滑块位置可以改变处理器和显卡的TDP和功耗墙，从而改善使用thinkpad玩常用游戏过程中严重掉帧和CPU锁频。
看来这是intel最近几年的一种新技术，我搜索了一下intel Speed Shift，结果喜人。看来这就是本疑问的正确答案——改变静默状态下的默认工作频率与电压，提高cpu时钟频率与计算机运行性能切换速度。
有空试一下使用电池时的区别。
找到了一篇英文的详细介绍文章，说的非常全面。我搬运过来。
========================================================================
Modern computer processors are constantlychanging their operating frequency (and voltage) depending on workload. ForIntel processors, this is often handled by the operating system which willrequest a particular level of performance, known as the Performance State orP-State, from the processor. The processor then adjusts its frequencies andvoltage levels to accomodate, in a DVFS (dynamic voltage and frequency scaling)sort of way, but only at the P-states fixed at the time of production. Whilethe best for performance would be to run the system at the maximum all thetime, due to the high voltage, this is the least efficient way to run aprocessor and wasteful in terms of energy used, which for mobile devices meansa shorter battery life or thermal throttling. With the P-state model, toincrease efficiency, the operating system can request lower P-states in orderto save power, but if a task requires more performance, and the power/thermalbudgets are sufficient, the P-State can be changed to accomodate. This'technology' on Intel processors has historically been called 'Speed Step'.
With Skylake, Intel's newest 6th generationCore processors, this changes. The processor has been designed in a way thatwith the right commands, the OS can hand control of the frequency and voltageback to the processor. Intel is calling this technology 'Speed Shift'. We’vediscussed Speed Shift before in Ian’s Skylake architecture analysis, butdespite the in-depth talk from Intel, Speed Shift was noticably absent at thetime of the launch of the processors. This is due to one of the requirementsfor Speed Shift - it requires operating system support to be able to hand overcontrol of the processor performance to the CPU, and Intel had to work withMicrosoft in order to get this functionality enabled in Windows 10. As of rightnow, anyone with a Skylake processor is actually not getting the benefit of thetechnology, at least right now. A patch will be rolled out in November forWindows 10 which will enable this functionality, but it is worth noting that itwill take a while for it to roll out to new Windows 10 purchases.
Compared to Speed Step / P-statetransitions, Intel's new Speed Shift terminology, changes the game by havingthe operating system relinquish some or all control of the P-States, andhanding that control off to the processor. This has a couple of noticablebenefits. First, it is much faster for the processor to control the ramp up anddown in frequency, compared to OS control. Second, the processor has much finercontrol over its states, allowing it to choose the most optimum performance levelfor a given task, and therefore using less energy as a result. Specific jumpsin frequency are reduced to around 1ms with Speed Shift's CPU control from20-30 ms on OS control, and going from an efficient power state to maximumperformance can be done in around 35 ms, compared to around 100 ms with thelegacy implementation. As seen in the images below, neither technology can jumpfrom low to high instantly, because to maintain data coherency throughfrequency/voltage changes there is an element of gradient as data is realigned.
The ability to quickly ramp up performanceis done to increase overall responsiveness of the system, rather than linger atlower frequencies waiting for OS to pass commands through a translation layer.Speed Shift cannot increase absolute maximum performance, but on shortworkloads that require a brief burst of performance, it can make a bigdifference in how quickly that task gets done. Ultimately, much of what we dofalls more into this category, such as web browsing or office work. As anexample, web browsing is all about getting the page loaded quickly, and thengetting the processor back down to idle.
For this short piece, Intel was able toprovide us with the Windows 10 patch for Speed Shift ahead of time, so that wecould test and see what kind of gains it can achieve. This gives us a somewhatunique situation, since we can isolate this one variable on a new processor andmeasure its impact on various workloads.
To test Speed Shift, I’ve chosen severaltasks which have workloads that could show some gain from Speed Shift. Testswhich run the processor at its maximum frequency for long periods of time arenot going to show any significant gain, since you are not limited by theresponsiveness of the processor in those cases. The first test is PCMark 8,which is a benchmark which attempts to represent real-life tasks, and theworkload is not constant. In addition, I’ve run the system through severalJavascript tests, which are the best case scenario for something like SpeedShift, since the processor has to quickly complete a task in order to allow youto enjoy a website.
PCMark 8
The processor in question is an Intel Corei7-6600U, with a base frequency of 2.6 GHz, and turbo frequency of 3.4 GHz.Despite the base frequency being rated on the box at 2.6 GHz, the processor cango all the way down to 400 Mhz when idle, so being able to ramp up quicklycould make a big impact even on the U-series Skylake processors. My guess isthat it will be even more beneficial to the Y series Core m3/m5/m7 parts sincethey have a larger dynamic range, and typically more thermal constraints.
Both the Home and Work tests show a verysmall gain with Speed Shift enabled. The length of these benchmarks, which arebetween 30 and 50 minutes, would likely mask any gains on short workloads. Ithink this illustrates that Speed Shift is just one more tool, and not a holy grailfor performance. The gain on Home is just under 3%, and the difference on theWork test is negligible.
JavaScript Tests
JavaScript is one of the use cases whereshort burst workloads are the name of the game, and here Speed Shift has a muchbigger impact. All tests were done with the Microsoft Edge browser.
The time to complete the Kraken 1.1 test isthe least affected, with just a 2.6% performance gain, but Octane's scoresshows over a 4% increase. The big win here though is WebXPRT. WebXPRT includessubtests, and in particular the Photo Enhancement subtest can see up to a 50%improvement in performance. This bumps the scores up significantly, withWebXPRT 2015 showing an almost 20% score increase, and WebXPRT 2013 has a 26%gain. These leaps in performance are certainly the kind that would benoticeable to the end user manipulating photographs in something like Picasa orwatching web-page based graph adjustments such as live stock feeds.
Power Consumption
The other side of the coin is powerconsumption. Having a processor that can quickly ramp up to its maximum frequencycould mean that it will consume more power due to the greater penalty ofincreasing the voltage, but if it can complete the task quickly and get back toidle again, there is a chance to be more efficient when work is done in 10s ofmilliseconds rather than 100s of milliseconds, as the frequency ramps up anddown again before the old P-state method has decided to do anything. Theprinciple of 'work fast, finish now' was the backbone of Intel's 'Race ToSleep' strategy during the ultrabook era and focused on the impulse ofresponse-related performance, however the drive for battery life means thatefficiency has tended to matter more, especially as devices and batteries getsmaller.
Due to the way modern processors work, wedon’t have the tools to directly measure the SoC power. Intel has told us thatSpeed Shift does not impact battery life very much, one way or the other, so toverify this, I've run our light battery life test with the option disabled andenabled.
This task is likely one of the best case scenariosfor Speed Shift. It consists of launching four web pages per minute, withplenty of idle time in between. Although Speed Shift seems to have a slightedge, it is very small and would fall within the margin of error on this test.Some tasks may see a slight improvement in efficiency, and others may see aslight regression, but Speed Shift is less of a power savings tool than otherpieces of Skylake. Looking at it another way, if, for example, the XPS 13 withSkylake was to get 15 hours of battery life, Speed Shift would only change theresult by about 7 minutes. Responsiveness increases, but net power use remainsabout the same.
Final Words
With Skylake, while there was not the largeleap in clock for clock performance gain that we have become accustomed to withnew Intel microarchitectures, but when you look at the overall package, therewas a decent net gain in performance combined with new technologies. Forexample, being able to maintain higher Turbo frequencies on multiple cores hasincreased the stock to stock performance more than the smaller IPC gains.
Speed Shift is just one small part of theoverall performance gain, and one that we have not been able to look at untilnow. It does lead to some pretty big gains in task completion, if the workloadsare bursty and short enough for it to make a difference. It can’t increase theabsolute performance of the processor, but it can get it to maximum performancein a much shorter amount of time, as well as get it back down to idle quicker.Intel is billing it as improved responsiveness, and it’s pretty clear that theyhave achieved that.
The one missing link is operating systemsupport. We’ve been told that the patch to enable this is coming to Windows 10in November. While this short piece looks at what Speed Shift can bring to thetable in terms of performance, if you'd like to read more about how it isimplemented, please check out the Skylake architecture analysis which goes intomore detail.
|
|
20 NB资产值
狗仔卡
离线
提升卡
置顶卡
变色卡
