|Introduction and Specifications|
|About this time last year, Intel offered us a complete processor revamp and architecture update for both the desktop and mobile markets. Intel called it their evolutionary "tick" step in their manufacturing process migration from 45 to 32nm. The "tock," as it were, follows along in cadence offering refinement and feature enhancement that completes the product evolution. So here we are, about 12 months or so later, and the "tock" cometh. Intel's tock architecture, known as Sandy Bridge, is still based on 32nm manufacturing process technology but offers critical features and performance enhancement, as well as higher levels of architectural integration. Intel is set launch a refresh of both desktop and mobile architectures and on the following pages here, we're aiming to give you a detailed view of what Sandy Bridge is going to bring for 2011 in the notebook arena. If you haven't been by our desktop line-up coverage, do yourself a favor and bone-up on the details we offer there as well, for what you're about to see here is all about how Intel intends to enhance your mobile computing experience.
Intel's Core i7-2820QM processor is the vehicle that we'll be using as a means of evaluating Intel's new architecture. It's not the highest-end SKU in the line-up but it has all the bells and whistles enabled and about 90% of top-end clock speed that Intel will offer in their "Extreme" version mobile chip. In a 45 Watt power envelope, this is the Sandy Bridge chip you'll likely see in some of the more capable multimedia targeted notebooks coming to market in 2011. And we'd dare say, at this early juncture, it packs a healthy serving of beef-cake computing muscle for just about anything you could throw at it. But enough of the genuflecting, let's get a look at the specifics and then put Sandy Bridge mobile to the test.
Cranking up the speeds and feeds, on the fly:
You'll have to excuse our cut-and-paste ways with the above eye-charts but in the interest of brevity we felt they laid things out nicely. As you'll note, there are dual-core and quad-core variants of Sandy Bridge mobile processors, all of which have Intel's new Turbo Boost 2.0 dynamic overclocking at a single core granularity if required. The Core i7-2820QM chip that we're testing today has a base clock of 2.3GHz with Turbo Boost speeds of 3.1GHz, 3.2GHz and 3.3GHz for quad-core, dual-core and single-core modes respectively. Also note that integrated graphics cores, which are now resident on-die versus the previous gen Arrandale discrete die, multi-chip module design, also have dynamic clock gating that scales from 650MHz to 1.3GHz. Finally, these new chips support Intel HyperThreading technology for up to 8 threads of processing resources in quad-core designs and 4-thread throughput in dual-core chips.
New engines under the hood:
All chips also support Intel's new AVX extension for increased SSE-based floating point performance as well as Intel AES (Advanced Encryption Standard) instructions for hardware-based encryption processing. Another very significant feature add is Intel's new Quick Sync hardware video encode engine that offloads video file conversion in hardware (we'll cover more on this in detail shortly). Finally, as is detailed in the bottom chart, Intel plans to release an entire family of low voltage and ultra low-voltage versions of Sandy Bridge that will drop into as little as a 17 Watt TDP power envelope. These chips will provide serious direct competition for AMD's low power Zacate Fusion processor, at least with respect to AMD's 18 Watt variant of the chip.
Let's look at chip-level and system-level architecture details next...
|Intel Sandy Bridge Mobile Architecture and Features|
|If you've been around long enough to remember Intel's 486 processor, you just have to marvel at the level of integration on the modern day CPU, whether you consider Intel or AMD's latest. 8MB of cache, four full execution cores and an integrated memory controller; these functional blocks alone offer orders of magnitude more on-chip resources than the basic function devices of yesteryear. And the very fact that they're resident, on-die, versus requiring system-level communication to access them, makes these processing elements even faster.
Intel Sandy Bridge Die Map
The most significant integration effort for Intel with Sandy Bridge of course, was to integrate their graphics core on die which meant the architecture would have to undergo a process migration from the previous generation Arrandale 45nm graphics core to the new 32nm technology found on the main processor die itself. As you can see, along with its logic and memory structures, the Intel HD Graphics 3000 core found on the Core i7-2820QM consumes a large chunk of die real estate. Another note-worthy design change here is that the chip's shared L3 cache is also now available to the graphics core for lightning-fast local memory access.
At the system level, Intel's new notebook platform architecture based on Sandy Bridge and the H67 Express chipset, is also a two chip solution like Arrandale was configured. The processor has 16 PCI Express 2.0 lanes available as well as a dual channel DDR3 memory controller capable of speeds up to 1333MHz. Intel's H67 Express chip offers the rest of the IO functionality for the system, including 6 SATA ports with option RAID capability, a Gigabit Ethernet MAC, display interface interfaces for HDMI, DVI and Displayport, as well as eight channel HD audio and eight additional PCI Express 2.0 lanes. Unfortunately the platform has been delivered sans USB 3.0 or 6Gps SATA connectivity at this juncture, which seems like an oversight but when you consider how long these chips have probably been in design, it's almost hard to fault Intel for the omission; we said almost. This is Intel we're talking about after all, right?
Finally, here's a quick look at Intel's Sandy Bridge Mobile roadmap moving forward. Dual-core and quad-core variants will replace Arrandale and Clarksfield versions starting in Q1 2011.
|Vital Signs and Turbo Boost 2.0 Overclocking|
|Intel's Turbo Boost dynamic clock gating and overclocking technology has been around since the early days of Intel's original Nehalem architecture. With the Sandy Bridge mobile architecture, Intel is taking things up and down a notch or two, depending on your workload at the time.
Sandy Bridge mobile chips have the ability to scale core clock speeds at a single, dual and four-core level, depending on thread level and workloads. The range of speed bins can vary widely as well, with our 2.3GHz Core i7-2820QM test chip offering top clock speeds of 1GHz+ faster in pure single-thread processing mode. At a basic level, the power management intelligence that Intel built into these chips, allows the architecture to scale clock speeds for optimal throughput depending on thread-level workload conditions, while still remaining within their thermal and power envelope requirements for platform design goals.
Intel Core i7-2820QM Sandy Bridge - Idle - Click for high res.
Here we see the Core i7-2820QM Sandy Bridge chip at idle on the desktop. As you can see, the core clock speed has ramped all the way down to a sleepy 800MHz while core-level thermals sit around a cool 40ºC temperature.
Intel Core i7-2820QM Sandy Bridge - Full Load - All Threads 3.1GHz - Click for high res.
Here you're looking at Core i7-2820QM under 100% full load, with all eight threads pegged. The Sandy Bridge mobile architecture takes the entire processor core complement up to 3.1GHz across all cores.
Intel Core i7-2820QM Sandy Bridge - Two Thread Load - 3.3GHz - Click for high res.
And finally, here's an example of Sandy Bridge mobile under a lightly threaded condition, specifically our Lame MT audio conversion test, which utilizes only two threads in its processing workload. The Core i7-2820QM takes its two active cores up to 3.3GHz to chew through this test with the fastest score we've seen yet from any mobile chip. More details on the performance results in this scenario, coming up.
|HD Media Playback and Encoding|
|Possibly one of our favorite additions to the Sandy Bridge architecture would be Intel's Quick Sync technology. Quick Sync essentially offers new functional blocks and hardware engines to offload video encoding on the CPU.
Rather than relying on brute force execution in software, Intel's new Sandy Bridge graphics core takes on HD video encode workloads within its graphics processing pipeline. That said, what's more interesting is how efficient and fast these new functional blocks are and how impressive our first experience with a Core i7-2820QM mobile processor turned out to be.
Cyberlink's Media Show Espresso is a video conversion tool that imports various video media files types and converts them to other standard video types for publication, distribution and streaming. In this test, we take a 224MB high definition 720p AVCHD video clip and convert it to a 720p H.264-encoded video file. Times are measured in minutes:seconds with lower times representing faster throughput in the video conversion process.
The top two gold bars in this graph are scores taken in Cyberlink Media Show Espresso with Sandy Bridge mobile crunching the file with its hardware engines enabled and also disabled. Intel's Quick Sync video encoding technology offers an amazing performance advantage in this application under our test conditions. The Quick Sync-enabled test run was performed in just a few seconds, versus over a minute required for the Core i7-2820QM to process the file without Quick Sync. What's more impressive however is how much faster Sandy Bridge was versus the NVIDIA GeForce GTX 280M graphics chip along with Intel's previous generation Core i7-920XM. Even NVIDIA's CUDA video processing engines can't keep up with Quick Sync and also cut it close versus Sandy Bridge without Quick Sync enabled.
Finally, with a powerful chip like the Core i7-2820QM, we didn't expect HD video playback to be an issue but regardless, we looked at CPU utilization playing back a 1080p H.264 QuickTime clip. We fired up the .MOV file with both Apple's QuickTime player as well as Windows Media Player, taking note of thread activity in Windows Task Manager Performance Monitor.
1080p H.264 Movie Trailer - Apple QuickTime Player
1080p H.264 Encoded Movie Trailer - Windows Media Player
In the QuickTime player the Core i8-2820QM shows about 5% utilization playing back this 1080p content. In Windows Media Player, resource utilization isn't even registering on the CPU usage meter.
|Test Systems and SiSoft SANDRA|
The Packard Bell machine that Intel sent us for testing with Sandy Bridge, had easy access to primary system components via a couple of panels on the underside of the notebook. However, since we were under the gun with this piece, we weren't able to do a full tear-down for you. You can thank Intel for pulling in NDA times and making us work over the holiday break . Our significant others certainly did. Break? What break?
Intel Core i7-2820QM Mobile Processor and H67 Express IO Hub Chip
The Core i7-2820QM comes packaged in a 1224 pin rPGA (Pin Grid Array) with alternative BGA (Ball Grid Array) versions available for direct solder to the PCB. Regardless, like previous designs, a half twist with a flat head screw driver will release the processor from its socket. On the top right you can see Intel's new H67 Express IO chip for the rest of the system's multitude of connectivity options.
Our Sandy Bridge Mobile Test System...
We began our testing with SiSoftware's SANDRA 2011B, the
System ANalyzer, Diagnostic
and Reporting Assistant. We ran three of the
built-in subsystem tests that partially comprise the SANDRA 2011 suite with
Intel's new Core i7-2820QM Sandy Bridge processor (CPU
Arithmetic, Multimedia, and Memory Bandwidth). All of the scores reported below were
taken with the processor running at its default clock speed of 2.3GHz with Turbo Boost scaling to 3.4GHz depending on thread-level workloads.
SiSoft SANDRA CPU Test
SiSoft SANDRA Multimedia Test
SiSoft SANDRA Memory Bandwidth Test
In terms of raw CPU throughput, the new Core i7-2820QM mobile processor shows it's faster than Intel's previous generation Core i7 975 desktop chip under SANDRA's Processor Arithmetic and Multi-Media tests. These early simple test results are obviously encouraging. In terms of memory bandwidth, Intel's new mobile processor architecture is also able to keep pace with the triple channel DDR3 memory and processor cache performance of an X58 chipset-based system in some cases, offering over 18GB/s of aggregate memory bandwidth. Remember that the Core i7-2820QM is doing this with a dual channel DDR3 configuration as well. Again, Intel's Sandy Bridge mobile architecture is stacking up very well in these tests.
|Futuremark PCMark Vantage Benchmarks|
Next up, we ran a number of different test systems through Futuremark’s latest system performance metric, PCMark Vantage. PCMark Vantage runs through a host of different usage scenarios to simulate different types of workloads including HD TV and movie playback and video manipulation, gaming, image editing and manipulation, music compression, communications, and productivity. Most of the tests are multi-threaded as well, so the tests can exploit the additional resources offered by multi-core CPUs.
The short story here is, obviously longer bars indicate better performance. In this case, the new Core i7-2820QM handily out-paces Intel's legacy architectures in almost every test, with the exception of the NVIDIA GPU-equipped Core i7-920XM system in the Gaming test suite and even then it's a relatively close race. In addition, if you look at the huge advantage the Sandy Bridge system has over the others in the Communications tests, it's obvious to see that Intel's new dedicated AES instruction set with dedicated hardware processing is kicking into gear nicely, nearly doubling performance of the previous generation Intel Arrandale dual-core architecture.
|LAME MT Audio Encoding|
In our custom LAME MT MP3 encoding test, we convert a large WAV file to the MP3 format, which is a popular scenario that many end users work with on a day-to-day basis to provide portability and storage of their digital audio content. LAME is an open-source mid to high bit-rate and VBR (variable bit rate) MP3 audio encoder that is used widely around the world in a multitude of third party applications.
In multi-threaded performance (Lame MT actually makes use of two threads), the new Core i7-2820QM puts up times some 42% faster than the next fastest processor in our test, the Core i5 540M. In single-threaded performance Sandy Bridge is about 14% faster than the previous generation Core i7-920XM.
|Cinebench R11.5 and POV-Ray Rendering|
|Cinebench R10 is an OpenGL 3D rendering performance test based on Cinema 4D from Maxon. Cinema 4D is a 3D rendering and animation tool suite used by 3D animation houses and producers like Sony Animation and many others. It's very demanding of system processor resources and is an excellent gauge of pure computational throughput.
This is a multi-threaded, multi-processor aware benchmark that renders a single 3D scene and tracks the length of the entire process. The rate at which each test system was able to render the entire scene is represented in the graph below.
The new Core i7-2820QM was easily the fastest machine in Cinebench's CPU test, punching up over 67% more performance than the next fastest mobile chip, the Core i7-920XM. What's perhaps more impressive is how much faster the OpenGL test ran with Intel's latest, fully integrated graphics core in Sandy Bridge. According to Cinebench, in OpenGL rendering,z the Core i7-2820QM's 32nm-built HD Graphics 3000 core is orders of magnitude faster than Arrandale's 45nm graphics core. As an aside, our Sandy Bridge mobile scores here are about 2X the performance of what we've seen from NVIDIA's 16 CUDA core-powered Ion 2 GPU.
POV-Ray continues to highlight the new Core i7-2820QM's dominance, showing over 45% more performance versus their previous generation quad-core mobile chip in the multi-threaded workload test and over 24% more throughput in the single-threaded test.
|Unoptimized H.264 Video Encoding|
x264 HD benchmark measures how fast a system can encode a short, DVD quality MPEG-2
video clip into a high-quality H.264 HD video clip.
This test reports the compression results in frames per second for each pass of the video encoding process. The test is also multi-threaded so it takes advantage of the additional resources afforded by multi-core processors.
Similar to what we saw in our earlier runs with Media Show Espresso, the Core i7-2820QM offers a massive performance kick in this test, which incidentally isn't remotely optimized for Intel's new Quick Sync video engine. Though it's a crude sort of measurement, it affords us a look at what legacy code might run like on Intel's new architecture, versus previous generation processors. Here the new Sandy Bridge Core i7 mobile CPU offers up well in excess of 50% more performance versus the Core i7 mobile chip of old.
|Gaming with Intel HD Graphics 3000 - Far Cry 2|
|Before we dive into our Core i7-2820QM gaming tests, let's take a quick scan through a feature comparison of Intel's new HD Graphics 2000/3000 engine versus the graphics core found on Intel's 2010 Core processors.
Though there are the same number of execution units available in the HD Graphics 3000 engine found in the Core i7-2820QM, there are additional processing resources like a dedicated math box and of course the Quick Sync media processing engine for HD video that we showed you earlier. The graphics core itself also supports DX10.1 as well as the OpenGL 3.0 spec and Shader Model 4.1. Also, you'll note that clock speed, which is again dynamically scaled based on workload, has also been increased 450MHz higher than the previous generation Arrandale architecture. Enough with the paperwork though, let's see what the frame rates tell us.
Here we took the liberty of folding in one other machine we happen to have lying around the lab for testing. Lenovo's ThinkPad T410s with NVIDIA Optimus technology comes equipped with an NVS3100M, which is essentially the Quadro version of NVIDIA's GT 310M GPU, which is a 16 CUDA core GPU clocked at 600MHz. It's essentially about as close of an apples-to-apples comparison of NVIDIA's nearest competitor that we could find, only in discrete form obviously.
Interestingly enough, the long and short of it is, at least with FarCry 2, Intel's new fully integrated HD Graphics 3000 core has the chops to keep pace with a solid low-end discrete mobile graphics solution, and even offers a bit more performance in the DX9 version of this game engine, likely due to driver optimization with NVIDIA's solution being more focused on the DX10 path. Playable frame rates are to be had with FarCry 2 at 720p resolutions with high image quality settings and no AA turned on. AA (Anti-Aliasing) turned out to be a frame rate killer for Intel here due to the fact that the chip needs to go out to system memory for larger capacity frame buffer access. Regardless, if we dialed back eye candy a bit further to more moderate levels, frame rates in excess of 30fps are very realistic. And look at the performance gain over Intel's previous generation IGP; over three times faster at 720p resolution is simply impressive no matter how you slice it.
Remember as well, this is a very early look at what Sandy Bridge mobile can do. Hopefully, with time, we could even see more performance optimization made at the driver level by Intel. And hopefully they're listening intently (or perhaps reading here rather). Intel's graphics driver team has never been known for earth-shattering execution in driver drops that provide new features and optimizations. Here's to a solid New Year's resolution, perhaps?
|Gaming with Intel HD Graphics 3000 - Metro 2033|
|Next we turn up the heat a little bit with Metro 2033, which is known to be a GPU-crusher if we ever saw one.
We tested Metro 2033 in DX9 mode, just to kick frame rates up as far as we could. Still, with medium quality and AA disabled, we couldn't get playable frame rates in this game engine on any of our mobile test systems here for that matter. Regardless, the Core i7-2820QM offers nearly 3X the performance of Intel's previous gen integrated graphics core in this test as well. We were impressed with these results to be sure regardless.
|Power Consumption and Battery Life Projections|
Before bringing this article to a close, we'd like to cover a few final data points--namely power consumption. Throughout all of our benchmarking and testing, we monitored how much power our Intel Sandy Bridge notebook test system was consuming using a power meter. Our goal was to give you all an idea as to how much power each configuration used while idling and while under a heavy workload.
A mere 16 Watts at idle and that's with the LCD panel on? Your eyes do not deceive you. 16 Watts is on the order of what an Atom-based netbook consumes under light workloads and this is a quad-core, eight thread capable processor that will scale in excess of 3GHz when called on. Recall our Vital Signs and Overclocking with Turbo Boost section from a few pages back. Sandy Bridge's 800MHz clock gating capability will afford notebook battery life advantages not seen on any Intel-based architecture thus far. Though since we really weren't looking at a retail ready notebook in this test, we're omitted our traditional battery life tests for this piece. We'll definitely explore this area more deeply with retail products from various major OEMs in the future.
At full load, even scaled to over 3.1GHz with all eight threads pegged at 100%, the Core i7-2820QM system didn't even break the 90 Watt level. When we fired up a more graphics intensive app like Cinebench's OpenGL test, we found the Sandy Bridge consumed about 56 Watts under heavy graphics loads that call less upon the chip's CPU engines. It may sound a bit like marketing speak but the Core i7-2820QM's performance-per-watt profile is literally unrivaled thus far in our humble estimation.
|Performance Summary and Conclusion|
Performance Summary: Intel's new Core i7-2820QM processor put up the fastest benchmark numbers we've seen from any notebook processor to date, with respect to general purpose computing workloads. In these test conditions, Intel's new Sandy Bridge mobile architecture offered performance gains of anywhere from 25 to over 50%, depending on thread-level workload. Where the Core i7-2820QM really excelled was within specialties like video encoding, gaming and encryption processing. Intel's new on-die IGP offered competitive performance to existing entry-level discrete and integrated mobile graphics solutions and out-paced previous generation Intel IGPs by as much as three-fold in our tests with FarCry 2 and Metro 2033. We saw a similar result within PCMark Vantage's Communications test, where Sandy Bridge's new AES instruction set offered over two times the performance of Intel's previous generation quad-core mobile architecture. Finally, our video encoding tests showed Sandy Bridge Quick Sync technology offers video encoding performance that rivals and surpasses dedicated GPU solutions in the market today, at least in the application we tested that were optimized to support the technology.
We're not sure what more there is to say that hasn't already been said in our evaluation and benchmark testing. If you're the type that likes to cut to the chase and just surf performance graphs, then, frankly, you've seen all there is to know about Intel's new mobile processor architecture for mainstream and high-end notebooks. Sandy Bridge is fast, crazy fast in fact. It consumes less power under load and has more aggressive clock gating when you're not using your machine but want to keep it at the ready with the lid up. Better still, Intel's latest, fully integrated HD Graphics core in the processor we tested, offers legitimate entry-level graphics capabilities for multimedia, video processing and light duty gaming. If you were previously looking for notebooks with discrete GPU solutions, just so you had some level of basic, dedicated graphics horsepower with current capabilities, we'd offer that you might not need to weigh that into your buying decision any longer. True mobile gamers are going to still going to want discrete GPUs but yes, Intel's integrated graphics solution, especially if they decide to resource driver development a bit more, could very well be, "good enough." And good enough doesn't mean you have to sacrifice much, in our book anyway.
Our only regret was that the test system we had to work with unfortunately won't see the light of day, as it was not intended to be a production model. However, though you can't buy the same system we tested, we're told Intel Sandy Bridge mobile processor-based notebooks will be launching in abundance over the next few weeks as we witness a multitude of unveilings at this year's Consumer Electronics show in Las Vegas. Typically we reserve our ratings and awards for products that you can actually buy but there are rare few exceptions where we offer them to a technology that will enable products to come in the market. This is definitely the case with Intel's new Core i7-2820QM processor and Intel's Sandy Bridge mobile notebook processor architecture. If we were in the market
for a notebook, whether it be thin-and-light or midrange, there's little
question a machine with one of Intel's new Sandy Bridge mobile chips
would be on our short list, more than once actually. If this is Intel's "tock" processor technology refresh, we can't wait to see the next "tick."