Intel Core i3-3110M Ivy Bridge versus i3-2370M Sandy Bridge - BeHardware
>> Processors

Written by Guillaume Louel

Published on July 26, 2012


Page 1

Sandy Bridge versus Ivy Bridge, Core i3 versus i5

After our first look at mobile processors (see our article here), we now also wanted to cover the Intel Core i3s. Very commonly used in laptops, these dual core processors can seem relatively similar to certain Core i5 models.

We have looked at two popular and relatively recent Core i3 models. Firstly in the Sandy Bridge range, we opted for the Core i3-2370M. Launched in the first quarter of this year, this dual core processor, which supports HyperThreading, comes with a 3 MB cache and is clocked at 2.4 GHz. We chose its exact equivalent from the Ivy Bridge range, namely the Core i3 3110M, a 2C/4T processor also clocked at 2.4 GHz and with a 3 MB cache. Like the processors we've tested in the past, these chips both have a TDP of 35W and, given their similar spec, it will be particularly interesting to observe the difference in energy consumption of the two architectures. As you know, Sandy Bridge processors use a 32nm manufacturing process while the Ivy Bridges are engraved at 22nm. Beyond this, thereís also another difference with respect to memory support. As already mentioned for the models further up the range, Sandy Bridge is limited to DDR3 1333 MHz while Ivy Bridge supports DDR3 1600 MHz memory.

On the left, the Core i3 2370M, on the right, the Core i3 3110M

Now letís move on to the multiple differences between Core i5 and Core i3. The most significant of these concerns the absence of turbo on the CPU side on Core i3s. Given the importance of Turbo in laptop configurations Ė remember it can clock an active core on the Core i5 3210M up from 2.5 to 3.1 GHz Ė the performance impact will be noticeable. Note that the graphics Turbo is still present on the Core i3s equipped with HD 3000s (Core i3 2370M) and HD 4000s (Core i3 3110M).

There are however other features on which sometimes surprising segmentation does exist. For example the Ivy Bridge Core i3s donít support AES-NI instructions that enable the acceleration of encryption/decryption operations (as long as youíre using applications that support this instruction set of course). Note also that while the Intel Core i3 desktop processors are limited to PCI Express 2.0, this isnít the case for the mobile range: sixteen PCI Express 3.0 lanes are available on the Ivy Bridge Core i3s, just like on the models further up the range.

The last notable difference is that the maximum authorised temperature (TJunction in Intel speak) is slightly lower: 90į on the Core i3 3110M and up to 105į on the Core i5 3210M. The lack of Turbo on the Core i3s of course means that the role of the TJunction is much less important here. To recap, the Core i3 2370M has a TJunction at 85į.

Page 2
Configuration recap

Configuration recap

In our tests we used three mobile platforms for each of the sockets represented. In order to put all the configurations on an equal footing with respect to energy consumption, we chose machines without additional graphics cards.

For FS1, the socket used by the AMD Llanos, we used a Lenovo Thinkpad Edge E525 laptop. For the AMD Trinity solutions (socket FS2), we used a test machine supplied by AMD. For the Intel platform, used for Sandy Bridge and Ivy Bridge (PGA988), we used a Clevo W270EU equipped with the HM76 chipset.

Our three platforms were equipped with three different but relatively similar chipsets:

On the AMD side the only difference between the A60M, which equips the Llano platform, and the A70M, which equips the Trinity platform, is the addition of an NEC block that supports USB 3.0. The chip TDP hasnít changed.

The Intel HM76 also supports USB 3.0, like its desktop equivalents. Still in line with the Intel desktop chipsets, only two SATA 6 Gb/s ports are supported here, which isnít too much of an issue on the mobile side. RAID is however absent and is only used on very few laptops (Intel has some higher end versions of its chipset with RAID). The Intel chipset supports up to 8 PCI Express 2.0 lanes to interconnect additional chips. The AMD chipsets settle for four lanes. Once again, this is plenty for this type of configuration.

We tested nine different processors which have been grouped in this table.

As you can see, comparing the Core i3s to the Core i5s, whether Sandy Bridges or Ivy Bridges, will be particularly interesting and this is what we will be focusing on most when we look at the benchmarks. Note also that the Core i3 2370Mís graphics turbo clock is 50 MHz lower than that on the 2410M. This i3/i5 difference doesnít exist in the Ivy Bridge range.

Page 3
Memory bandwidth, energy consumption

First we measured the memory performance of our respective platforms.

Monothreaded bandwidth

We started with the memory bandwidth accessible via a single thread. The readings were taken with the Aida64 memory test.

The lack of Turbo in monothreaded mode has an impact on memory performance in dual channel mode. It can be seen on the Core i3 2370M and is marked on the Core i3 3110M! While moving from one to two channels on the Core i5 3210M gives an increase in bandwidth of 43% in reads, the gain is just 26% on the Core i3 3110M!

Multithreaded bandwidth

Let's now move on to the multithreaded test included in Rightmark.

When all threads (4) are used on a Core i3, the difference in clock only has a minimal impact and both memory channels can be correctly used. It will be interesting to see if this difference has an impact in our applications tests.

Energy consumption

We measured the energy consumption of our platforms in different scenarios at the socket:
  • Machine at idle
  • Playback of an H.264 720p video file in DXVA mode via MPC-HC
  • Processor load (Cinebench)
  • GPU load (Furmark)
  • Processor + GPU load (Cinebench + Furmark)
  • Load in games (F1 2011)

Letís look at the results!

As we said in our previous article, Intel allows its Turbo equipped processors to go over the TDP for 28 seconds. Itís this maximum energy consumption that we have noted in this graph and this is why the energy consumption of the dual core processors is particularly high in ĎCinebench+Furmarkí. This is the only instance where these processors exceed their TDP.

Variability between one mobile processor and another is relatively significant. On the Sandy Bridge side, while our Core i3 does a bit better at idle than the Core i5, in load theyíre almost on a par in spite of the fact that the Core i3 doesnít have Turbo. For the Ivy Bridges itís the other way round. While the Core i3 is pretty much on a par with the Core i5 3210M at idle, its energy consumption in load is significantly down. In F1 2011 it even looks very low, but it should be remembered that the energy consumption in games is directly linked to the number of frames per second calculated. Weíll see what this means in practice.

When we compare the Core i3 2370M Sandy Bridge to the Core i3 3110M Ivy Bridge, we can see that as with models further up the range, the Sandy Bridge models have a small energy consumption advantage at idle or in video playback. Here itís in the order of 4 or 5 Watts. In load, the trend is reversed and the 3110M consumes 4 to 5 Watts less in Cinebench and Furmark. When both are considered together, the 3110M is particularly economical, consuming 11 Watts less than its predecessor.

Page 4
CPU performance: Cinebench, x264, Visual Studio

We began our performance tests by looking at processor-side performance.

Cinebench R11.5

We used Cinebench version R11.5 to measure 3D rendering performance. To recap, the software uses the Cinema 4D rendering engine.

[ Monothreaded ]  [ Multithreaded ]

At equal clocks, our Core i3 Ivy Bridge does better than the Sandy Bridge model, by 5% in monothreaded mode and 3.4% in multithreaded mode.
Staxrip - x264 b2197

Moving frontend, we used Staxrip to transcode a scene from Avatar via x264 in build 2197. We carried out a medium type 2 pass encoding on a 720p source, re-encoded at a bitrate of 6 Mbits/s. To recap, the second pass is the one that benefits most from multithreading.

[ 1st pass ]  [ 2nd pass ]

Compared to its Sandy Bridge clone, the Ivy Bridge model has an 8.8% advantage on the second pass. This makes the Core i3 Ivy Bridge faster than the quad-core A10-4600M Trinity in this test.

Visual Studio 2011 beta

We went for the 2011 beta version of Visual Studio. We compiled the latest version (1.7.4) of the source code of the 3D Obre engine (examples included). Parallel compilation was activated for each project in VS.

The Core i3 Ivy Bridge has around a 4% gain over its predecessor. The Turbo mode of the Core i5s still represents a significant advantage here.

Page 5
CPU performance: 7-Zip, Bibble, DxO Optics Pro, WinZIP

7-Zip 9.20

We used version 9.20 of 7-Zip to compress a large volume of files using the LZMA2 algorithm.

The Core i3 Ivy Bridge has around a 5% gain over its predecessor. Using a Core i5 Ivy Bridge reduces the compression time by around 14%.

Bibble 5.2.3

Letís bring our purely processor tests to an end with the photo processing software, Bibble. We processed a lot of 48 RAW photos, exported as JPEGs.

The gap between the Core i3 Ivy Bridge and Sandy Bridge is up to 6% here. The Core i5 3210M does better by around 17.2%.

We also measured performance in our two OpenCL benches, in CPU mode only for the Intel processors because of the lack of compatibility of the driver.

DxO Optics Pro 7.2.3

We used version 7.2.3 of this photo processing application to carry out RAW to JPEG exports on a series of 48 files. Note that version 7.5 was recently released.

[ CPU ]  [ OpenCL ]

DxO Optics Pro only allows activation of OpenCL by default on a platform where the CPU is slower than the OpenCL acceleration. A benchmark was run at launch of the application to measure the respective performance levels. This benchmark canít be run on the quad-core Ivy Bridge solutions. We were however able to run it on the dual core Core i5 3210M, but activation of OpenCL systematically results in a blockage during conversion of the first photo. The Intel OpenCL driver is particularly capricious as weíll see further on!

If we focus on pure processor performance, there's a marked advantage for the Core i5 3210M over the Core i3 3110M: almost 24.3%! The Ivy Bridge/Sandy Bridge gap is within the norm.

WinZIP 16.5

As highlighted by AMD at the launch of its Radeon HD 7000s, version 16.5 of WinZIP includes OpenCL support. We compressed the same set of files as that used in our 7-Zip test.

[ CPU ]  [ OpenCL ]

Remember that OpenCL canít be activated on the Ivy Bridges as the option doesnít appear in the application. Perhaps the developer hasnít yet validated the Intel driver.

Processor performance here is within the norm, namely 5% between the Core i3 Sandy Bridges and Ivy Bridges, and around 15% between the Core i5 3210M and Core i3 3110M.

Lets now move on to the games tests!

Page 6
Gaming performance: F1 2011, Civilization V, Battlefield 3

We measured performance in five modern games on our mobile platforms. To recap, we took our readings at 1366 x 768. We have taken Diablo III out of our comparative as the arrival of a new patch has changed the rendering and performance on the Intel processors.

F1 2011

We started with the Codemasters game, F1 2011. We tested the game both in DirectX 9 and DirectX 11 modes. We used the Medium/Intermediary setting.

[ DirectX 9 ]  [ DirectX 11 ]

Only having a single memory channel on a machine isnít a very good idea. It is interesting to note however that in DirectX 9 mode, although the Core i5 3210M and the Core i3 3110M are almost on an equal footing on a single channel, the gap between the two processors is significantly higher when a second is available. This isnít the case on Sandy Bridge where the memory bandwidth is more of a limit to performance than the processor clock.

Civilization V

We measured the graphics performance on a scene loaded at the end of a game. All details were set to minimum.

Performance in dual channel mode is relatively limited on the Core i3 3110M Ivy Bridge when compared to the Core i5 3210M, which seems to benefit from the Turbo mode and its additional memory bandwidth. In practice, the Core i3 3110M and Core i3 2370M give very similar results on two channels.

Battlefield 3

We used the lowest setting in this title.

In Battlefield 3, the Core i3 3110M does better than its predecessor. This demanding title is not however playable on any of these platforms, even at our modest resolution!

Page 7
Gaming performance: Batman Arkham City, Crysis 2

Batman Arkham City

We used the built-in benchmark with the options set at minimum.

Batman is relatively playable on mobile platforms and the HD 4000 does particularly well, including on our Core i3 3110M, which gives a 27.5% gain over the previous generation. The Core i5 3210M is still 10% faster however.

Crysis 2

We measured performance in a scene in "High" modeÖ which is the lowest graphics mode available on this title! Marketing speak againÖ

None of the platforms except Trinity allow you to play this title though the Core i3 3110M marks its best performance here with a 38% lead over its predecessor.

Page 8

Letís be clear, removing Turbo mode on the x86 cores of the Core i3s means a lower running clock in practice. It's therefore no surprise to see an impact on performance; in our tests thereís around a 15% difference between the Core i5 3210M and the Core i3 3110M. This is a significant gap but it should be noted that the base performance is still relatively solid. In practice this Core i3 3110M remains in front of the AMD quad cores in all our processor tests, in spite of its low clock and the fact that it has two less cores.

Compared to the equivalent Sandy Bridge model, the 2370M, there's not a very big gap in performance. Around 5% separates the two chips, corresponding to an architectural gain and the gain in bandwidth given by the DDR3-1600.

When playing games however, the gap between Sandy Bridge and Ivy Bridge is extended. This is no surprise as apart from the points highlighted above and which remain relevant here, HD 4000 often gives Ivy Bridge the advantage. Often, but not all the time. The lower clock has an impact on graphics performance to the point where HD 4000 graphics isnít exploited to its full potential. In practice however, HD 4000 performance, whether on a Core i3 or a Core i5, only really gives access to games that arenít graphically demanding. This terrain is one where AMD stands out with its Trinity processors, which offer significantly higher performance.

Of course, when we start talking pricing, the mobility question becomes much more complicated! Intel does of course provide prices on its site, but in practice it should perhaps not bother. The Core i3 2370M, 3110M and Core i5 2410M and 3210M are according to Intel's site, priced at the same level!

In practice, the only reliable measure is the price of laptop PCs, though even then itís impossible to find identical machines which use the Core i3 2370M, 3110M and Core i5 3210M. If we consider that the Core i3 3110M is priced at the same level as the 2370M and therefore should be at the same price in new configurations on the market, the price difference between Core i3 and Core i5 machines is very variable. Manufacturers often take advantage of including, say, a slightly bigger hard drive to increase prices by as much as Ä100!

Obviously you have to weigh up this sort of increase against needs, while keeping in mind the fact that the Core i3 already offers a comfortable level of performance for most uses and that while an i5 does bring a real gain in performance (in the order of 15% between the Core i5 3210M and the Core i3 3110M), this gain isn't anything like that generally found between i3 and i5 on desktop.

Copyright © 1997-2014 BeHardware. All rights reserved.