Tuesday, September 29, 2009

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) ReviewIntroduction and Specifications


For quite some time now, Intel has been the undisputed leader when it comes to laptop CPU performance. This advantage comes from the fact that Intel has successfully scaled its desktop processor technology for use in the mobile platform. That trend continues with the latest mobile processor platform being unleashed today from Intel: the "Clarksfield"-based Core i7 Mobile processor family and the new PM55 Express Chipset.

This marks the first time that the "Nehalem" Core i7 microarchitecture has been ported over to the mobile side. The fact that the mobile version of Nehalem makes its debut only two weeks following the launch of Intel's Core i5 ("Lynnfield") mainstream desktop CPU and P55 Express chipset is no coincidence. As it turns out, the Core i7 Mobile processor die is identical to the desktop version, but uses mobile packaging--as far as the microarchitecture is concerned, Clarksfield and Lynnfield are essentially the same thing. The lower power demands of Lynnfield (as opposed to the original Nehalem architecture) are in large part what enable it to also be used in notebooks.

What this means is that Core i7 Mobile-based notebooks will see a number of the same benefits that Core i7 desktops already have, such as integrating the memory controller into the processor die, using a three-level cache hierarchy, utilizing Hyper-Threading technology, and taking advantage of Intel Turbo Boost Technology. What Lynnfield/Clarksfield adds to the equation is on-die PCI Express connectivity, getting rid of the Northbridge chip, and improving the Intel Turbo Boost Technology (from the original Nehalem architecture). Additionally, whereas only the previous Intel Core 2 mobile ("Penryn") microarchitecture had primarily dual-core offerings, all of the Core i7 Mobile solutions are quad cores. The three Penryn-based quad-core mobile processors (the 2.53GHz Core 2 Extreme QX9300, 2.26GHz Core 2 Quad Mobile Q9100, and 2.0GHz Core 2 Quad Mobile Q9000) are made from two dual-core chips merged together in a single CPU package, while Clarksfield uses a single-chip (monolithic) design.



Core i7 Mobile processors are available in three versions. The flagship processor is the Core i7-920XM, which has a base speed of 2.0GHz and goes up to 3.2GHz using Turbo Boost. The middle CPU is the Core i7-820QM, which has a base speed of 1.73GHz and can go up to 3.06GHz with Turbo Boost. Last, but not least, is the Core i7-820QM, with a base speed of 1.6GHz and maximum Turbo Boost speed of 2.8GHz. The Core i7-920XM and Core i7-820QM both have 8MB of shared L3 Smart Cache, while the Core i7-720QM has 6MB of L3 Smart Cache. The other notable difference between these three processors is that the Core i7-920XM has a maximum TDP of 55W, while the Core i7-820QM and Core 720QM both have a maximum 45W TDP. (For comparison, the three quad-core Penryn processors all have a maximum TDP of 45W).

...................................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Core i7 Mobile Processor Details


The Core i7 has a die size of approximately 296mm2 and contains roughly 774-million transistors. The die includes four physical execution cores, three levels of cache, an integrated memory controller, and PCI Express interconnects.
Lynnfield/Clarksfield Die Map

Since the Core i7 Mobile processor integrates the memory controller and PCI Express interconnects onto the processor, there is no need for a traditional Northbridge chip. This means that the PM55 chipset acts primarily as a Southbridge chip, handling most of the device I/O. With no Northbridge chip to communicate with, the Core i7 Mobile processor doesn't need the Intel QuickPath Interconnect (QPI) that the Core i7-900 series ("Bloomfield") processors use to communicate with the chipset. Instead, the Core i7 Mobile uses the slower Direct Media Interface (DMI) interconnect to communicate with the PM55 chipset. Another major difference between the Core i7 Mobile and the Core i7 900-series, is that unlike the Core i7 900-series support for triple-channel DDR3 memory, the Core i7 Mobile instead supports just dual-channel DDR3. The Core i7 Mobile processor supports configurations of 16 PCI Express 2.0 lanes per GPU or two sets of 8 PCI Express 2.0 lanes.
With the memory controller and PCI Express interconnects on the processor die,
the Core i7 Mobile doesn't need a Northbridge chip.


The Core i7-920XM and Core i7-820QM processors have the same three-level cache architecture as all other existing Core i7 and Core i5 processors. The cache configuration is made up of a 4 x 32K instruction L1 cache, 4 x 32K data L1 cache, 4 x 256K L2 cache, and an 8MB L3 cache that is shared between all four cores. The Core i7-720QM has the same L1 and L2 cache amounts, but instead has 6MB L3 shared cache.

All Core i7 Mobile CPUs have three levels of cache; the
Core i7-920XM and Core i7-820QM have 8MB of shared
L3 cache (the Core i7-720QM has 6B of shared L3 cache).


Also, as with all other Core i7 processors, the Core i7 Mobile includes Hyper-Threading support. Hyper-Threading allows each core to process two simultaneous threads. As a result, the Core i7 Mobile can run up to eight simultaneous threads.
With Hyper-Threading, the Core i7-920MX has four physical execution cores and
four virtual cores, for a total of eight cores available for multi-threaded workloads.


As you can see in the Windows 7 Resource Monitor screenshot above, the system sees a total of eight cores: four physical execution cores plus four virtual cores via Hyper-Threading.

..............................................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Intel Turbo Boost Technology



Intel Turbo Boost Technology has been part of the Core i7 (Nehalem) architecture from day one. However, Lynnfield and Clarksfield bring a significant update to the Turbo Boost technology, making it an even more flexible and powerful performance booster.

A graphical example of how Core i7 Mobile Turbo Boost works.

The Core i7 Mobile processor is actually capable of three different speeds above its base speed, depending on the nature of the active single or multi-threaded workloads (and as long as the processor is operating within specified thermal and power limits). When a single-threaded workload is detected, the Core i7-920XM's clock speed on just a single execution core increases up to 3.2GHz, while the remaining three cores remain inactive and operate and close to zero power. When a lightly-threaded workload is detected, only two cores are active and they operate at up to 3.06GHz. With a high-threaded workload, all four cores are active and operate at up to 2.26GHz.









A single-threaded workload (left) vs. a heavily-threaded workload (right).

The image on the left was recorded while Cinebench R10 was running its single-threaded workload. As you can see in Windows 7's Resource Monitor, CPU2 is getting pegged heavily. There is also some activity on CPU1, CPU5, and CPU7, which can be attributed to Windows 7 background tasks and other apps running in the background. The Intel Turbo Boost Technology Monitor gadget shows the processor running at 3.2GHz. While this screenshot shows that the Core i7-920XM is capable of scaling up to 3.2GHz, the chip actually ran at 3.07GHz during the vast majority of the test run.

The screenshot on the right was captured while Cinebench R10 was running its multi-threaded workload. Resource Monitor shows all four physical execution cores and all four Hyper-Threading virtual cores completely saturated, with the Intel Turbo Boost Technology Monitor showing the processor running at 2.0GHz. Curiously, the processor did not run at the 2.26GHz speed that Turbo Boost supports for highly-threaded workloads, and instead ran at the processor's 2.0GHz base speed.


....................................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Intel PM55 Express Chipset


If you read our recent coverage of the Lynnfield processor (Intel Core i5 and the Core i7 800-series), then the diagram below of the Clarksfield platform should look very familiar (in fact, it is the same diagram). As we previously mentioned, Clarksfield and Lynnfield have the identical architecture. The same can be said for the P55 and PM55 chipset feature sets.

Prior to the release of Lynnfield and Clarksfield, the Core i7's PCI Express interconnects were still located on the chipset; therefore the Intel X58 chipset for Bloomfield processors still required a Northbridge chip and a speedy QPI interconnect between the processor and the Northbridge chip. Now with the PCI Express interconnects and the memory controller located on the processor die, the Northbridge chip is gone from the equation and the CPU communicates directly with the Southbridge chip via a DMI interconnect.

The PM55 chipset supports up to 14 USB 2.0 ports and six SATA 3Gb/sec ports; it also supports Intel Matrix Storage Technology RAID support and Intel High Definition Audio. The PM55 chipset also includes support for up to eight PCI Express x1 ports. PCI Express ports 1 through 4 can be configured as either four x1, two x2, or one x4 lane groups; the same is true for PCI Express ports 5 through 8. This means that the PM55's PCI Express lanes could be configured as two x4. Unfortunately, it is not possible to gang all eight PCI Express lanes into single x8 skit.


Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Test System and SiSoft SANDRA

In order to test the performance of the new Intel Core i7 Mobile platform, we got our hands on a Core i7-920XM-based whitebook from Intel. Strictly speaking, the unit we tested is not an actual shipping configuration. But since the whitebook came in a Style-Note chassis (model W870CU) from Taiwan-based notebook manufacturer Clevo, it is likely that we will see shipping Core i7-920XM-based notebooks that closely resemble the configuration and design of the whitebook. Clevo manufacturers notebooks for a number of OEMs, such as AVADirect and Eurocom. In fact, Eurocom has already announced that it will be offering a Core i7 Mobile-based system using the same Clevo W870CU notebook design come October.






































*Intel Core i7-920XM 2.0GHz (2.26/3.06/3.2 Max Turbo Frequencies (4C/2C/1C))

* 4GB DDR3-1333 (2x2GB) SDRAM

* NVIDIA GeForce GTX 280M graphics with 1GB VRAM

* 17-inch 1600x900 display

* Intel 80GB X25-M SATA SSD

* DVD-Multi Recorder

* Intel Ultimate N WiFi Link 5300 WLAN

*
GbE Ethernet

* 42 WHr Battery

* Windows 7 Ultimate

* ExpressCard/54 slot

* eSATA port

* DVI out

* HDMI out

* 4 USB ports

* 1394 port
The Clevo W870CU whitebook unit measures 2.25x16.25x11.0-inches (HWD) and weighs about 8 pounds. While we keep calling it a whitebook--as the images show--it is not acutally white ("whitebook" is just an expression to indicate that it is a custom design that is not available for sale.). The W870CU's chassis is an almost all-black affair with red trim. The left side of the unit includes the DVD-RW drive, USB port, FireWire port, MMC/SD/MS card slot, TV antenna input, and RJ-11 modem jack. The right side of the unit includes four audio ports, USB port, ExpressCard/54 slot, eSATA slot, DVI out, and lock slot. The back of the unit has an HDMI out port, power connector, two USB ports, and an RJ-45 Ethernet port. The bottom of the unit features no less than three separate exhaust ports for the two internal fans (one for the CPU and one for the GPU).

The estimated street price of the Core i7-290XM-based Clevo W870CU is around $3,299--depending on how the OEM configures it. Before the Intel techs sent us the unit for testing, they swapped out the notebook's 250GB SATA hard drive for an 80GB Intel X-25-M SATA SSD. Such an upgrade raises the price by about another $275 or so.

We began our testing with SiSoftware's SANDRA 2009 SP4, the System ANalyzer, Diagnostic and Reporting Assistant. We ran four of the built-in subsystem tests that partially comprise the SANDRA 2009 SP4 suite: CPU Arithmetic, Multimedia, Memory Bandwidth, and Cache and Memory.























On all of the SiSoftware's SANDRA tests, the Core i7-920XM puts in a very strong showing. While our whitebook wasn't the top performer in the Processor Arithmetic test, a closer look shows that the CPU that SANDRA automatically picked to compare the whitebook's performance too was actually another Core i7-920XM--but one running with Turbo Boost bumping up the processor speed to 3.2GHz. The whitebook is also not the winner on the Processor Multi-Media test; but the automatically-generated comparison CPU in this case is a 3.33GHz Intel Core i7 975 Extreme Edition desktop processor--which is not exactly a fair comparison.

..............................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Futuremark PCMark Vantage



Next up, we ran a number of different test systems through Futuremark's PCMark Vantage system performance metric. PCMark Vantage runs through a host of different usage scenarios to simulate different types of workloads including High Definition TV and movie playback and manipulation, gaming, image editing and manipulation, music compression, communications, and productivity. Most of the tests are multi-threaded (up to three simultaneous threads), so the tests can exploit the additional resources offered by a multi-core CPU.

We pitted the Core i7-920XM against a number of other different processors and platforms. It is important to note that unlike when we test desktop processors, we can't just swap out the processor from the notebook's motherboard and pop in a new one for comparison, while keeping all other components the same. Therefore, all of the test results presented here and on the following pages include systems with different configurations, including different chipsets, memory types and speeds, hard disk drives, and GPUs. These comparisons are still valid, but the numbers represent specific configurations and models, and are not necessarily representative of all systems that use these processors.

Our primary comparison system in these pages a Dell XPS M1730, which is powered by a 2.8GHz Core 2 Duo X9000, 4GB 667MHz DDR2 SDRAM, 80GB Intel SSD hard drive, and dual Nvidia GeForce 9800M GTX GPUs using SLI. But we also included scores for a number of desktop processors as well, such as a 2.66GHz Core i5-750, 2.66GHz Core i7-920, and a 3.4GHz AMD Phenom II X4 965.

On the PCMark Vantage test, the Core i7-920XM's score of 12,517 PCMarks easily bests all of the comparison systems by a significant margin. The Core i7-920XM falters a bit on the Communications and TV & Movies workloads, but more than makes up for it on the other tests.

.........................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
LAME MT and X264 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 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 this test, we created our own 223MB WAV file (a hallucinogenically-induced Grateful Dead jam) and converted it to the MP3 format using the multi-thread capable LAME MT application in single and multi-thread modes. Processing times are recorded below, listed in seconds. Shorter times equate to better performance.

The Core i7-920XM puts in a very strong showing on our LAME MT test, but its multi-threaded performance is still only on par with that of the 2.8GHz Core 2 Duo X9000-based Dell notebook. In fact, with the exception of the Core 2 Quad 9400, all of our comparisons systems were at least as fast as the Core i7-920XM or faster when it comes to multi-threaded performance on this encoding test. If you look at the scores, however, you will see that there is a very tight grouping here with very similar multi-threaded performance among them. There is a wider range of performance with the single-threaded encoding performance, and here the Core i7-920XM comes out on top--in part because the test's single-threaded workload enabled the CPU to scale up to 3.2GHz using Turbo Boost. Even though the Core i5-750 also scales up to 3.2GHz with Turbo Boost, the Core i5 processor lacks Hyper-Threading--which is what gives the Core i7-920XM the edge over the Core i5-750 here.
The x264 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. The application reports the compression results in frames per second for each pass of the video encoding process, and it is threaded so it can take advantage of the additional resources afforded by multi-core processors.

On this HD video encoding test, the Core i7-920XM was significantly faster that the 2.8GHz Core 2 Duo X9000-based Dell notebook. This is an indication that the new Core i7 Mobile platform should be much more efficient at encoding video than the Core 2 Duo processor--an increasingly important task for all platforms. On the other hand, all but one of the desktop-based CPUs handily beat the Core i7-920XM with their encoding speed. While the mobile processor may have come a long way in its performance capabilities, a modern desktop CPU is still the better choice when it comes to encoding video quickly.

............................................................................................

Monday, September 28, 2009

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Cinebench R10 and POV-Ray


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 is 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.
Three of the four comparison desktop CPUs outperformed the Core i7-920XM on the multi-threaded iteration of Cinebench R10; but the Core i7-920XM was faster than every single one of the comparison systems on the single-threaded version of the test. Once again this is direct result of the Core i7-920XM's ability to scale its processor speed up to 3.2GHz using Turbo Boost Technology.

POV-Ray, or the Persistence of Vision Ray-Tracer, is a top-notch open source tool for creating realistically lit 3D graphics artwork. We tested with POV-Ray's standard 'all-CPU' benchmarking tool on the test machines, and recorded the scores reported for each. Results are measured in pixels-per-second throughput; higher scores equate to better performance.
Our results with the POV-Ray benchmark nearly mirror those of Cinebench R10 above. The only notable difference is that on this benchmark, the 3.4GHz AMD Phenom II X4 965 edges out the Core i7-920XM for top honors on the single-threaded workload. As this test is very sensitive to raw CPU speed, it's not a complete surprise that the processor running at the highest speed was the one to put in the fastest single-threaded performance.

................................................................................................

Intel Core i7 Mobile CPU (Clarksfield) Review

Intel Core i7 Mobile CPU (Clarksfield) Review
Futuremark 3DMark Vantage


3DMark Vantage
Futuremark's synthetic 3D gaming benchmark, 3DMark Vantage, uses some advanced visual technologies that are only available with DirectX 10. 3DMark Vantage isn't simply a port of 3DMark06 to DirectX 10 though. With this latest version of the benchmark, Futuremark has incorporated two new graphics tests, two new CPU tests, several new feature tests, in addition to support for the latest PC hardware. We tested the systems here with 3DMark Vantage's Extreme preset option, which uses a resolution of 1,920x1,200 (as this exceeded the native screen resolution of the Core i7-920XM whitebook, we conducted this test with the notebook sending its video out to an external display).


Even though the Core i7-920XM has what is arguably the fastest mobile CPU and mobile GPU (the Nvidia GeForce GTX 280M) available today, we still wouldn't expect this amped-up mobile platform to beat high-end desktop gaming rigs when it comes to 3D graphics. We ran this test primarily to see how close the platform could come to desktop-like 3D graphics performance. We'll be the first to admit that the Core i7-920XM whitebook's 3D graphics performance wouldn't make it a candidate as our first-choice system for a LAN party, but its gaming performance is about some of the most-powerful we've seen for a non-SLI notebook.

3DMark Vantage
3DMark Vantage's CPU Test 2 is a multi-threaded test designed for comparing relative game physics processing performance between systems. This test consists of a single scene that features an air race of sorts, with a complex configuration of gates. There are aircraft in the test that trail smoke and collide with various cloth and soft-body obstacles, each other, and the ground. The smoke spreads, and reacts to the planes as they pass through it as well and all of this is calculated on the host CPU. We disabled the benchmark's Nvidia PhysX in order to better isolate the performance of the CPU.

The Core i7-920XM whitebook handily beat the 2.8GHz Core 2 Duo X9800-based Dell notebook, as well as holding its own against the 2.66GHz Core i5-750 and 2.66GHz Core 2 Quad Q9400 systems. Not surprisingly, the most powerful system on this test uses the highest-end desktop processor of our comparison systems, the Core i7-920 desktop CPU.

..............................................................................................