The Test

As always we're only presenting a subset of our entire CPU test suite here. For all of the numbers as well as other comparison options check out Bench.

CPU: AMD A8-3850, AMD A6-3650, AMD Athlon II X4 635, AMD Athlon II X3 455, AMD Athlon II X2 265
Intel Core i3-2100, Intel Core i3 540, Intel Pentium G850, Intel Pentium G840, Intel Pentium G620, Intel Pentium G620T
Motherboard: Intel DH67BL (Intel H67)
ASRock A75 Extreme6 (AMD A75)
Chipset Drivers: Intel 9.2.0.1025
AMD Catalyst 8.862 RC1
Hard Disk: Intel X25-M (80GB), Corsair P256 SSD (256GB)
Memory: G.Skill DDR3-1866 2 x 4GB
Video Drivers: AMD Catalyst 8.862 RC1
Intel 2372
Desktop Resolution: 1920 x 1200
OS: Windows 7 x64

Adobe Photoshop CS4 Performance

To measure performance under Photoshop CS4 we turn to the Retouch Artists’ Speed Test. The test does basic photo editing; there are a couple of color space conversions, many layer creations, color curve adjustment, image and canvas size adjustment, unsharp mask, and finally a gaussian blur performed on the entire image.

The whole process is timed and thanks to the use of Intel's X25-M SSD as our test bed hard drive, performance is far more predictable than back when we used to test on mechanical disks.

Time is reported in seconds and the lower numbers mean better performance. The test is multithreaded and can hit all four cores in a quad-core machine.

Adobe Photoshop CS4 - Retouch Artists Speed Test

The Pentium G620 does a lot better than its Athlon II X2 counterpart, there's no contest here. The race is a lot closer between the G840/850 and the Athlon II X3 455, but in the end the Pentiums are faster. Both perform a bit better than the more expensive AMD A8-3850.

x264 HD Video Encoding Performance

Graysky's x264 HD test uses the publicly available x264 encoder to transcode a 4Mbps 720p MPEG-2 source. The focus here is on quality rather than speed, thus the benchmark uses a 2-pass encode and reports the average frame rate in each pass.

x264 HD Encode Test - 1st pass - x264 0.59.819

x264 HD Encode Test - 2nd pass - x264 0.59.819

Once again the G620 is faster than the Athlon II X2 265 despite the latter's clock speed advantage. The G840 and G850 are faster than the 620 but not fast enough to overcome the extra core advantage of the X3 455. If you're spending $80 on a CPU for video encoding go for the Athlon II X3 455.

3dsmax 9 - SPECapc 3dsmax CPU Rendering Test

Today's desktop processors are more than fast enough to do professional level 3D rendering at home. To look at performance under 3dsmax we ran the SPECapc 3dsmax 8 benchmark (only the CPU rendering tests) under 3dsmax 9 SP1. The results reported are the rendering composite scores:

3dsmax r9 - SPECapc 3dsmax 8 CPU Test

Not all heavily threaded tasks are going to favor more cores. In this case the IPC advantages of Sandy Bridge give the G850 equal performance to the X3 455.

Cinebench R10

Created by the Cinema 4D folks we have Cinebench, a popular 3D rendering benchmark that gives us both single and multi-threaded 3D rendering results.

Cinebench R10 - Single Threaded Benchmark

Single threaded performance is going to be a big advantage of the SNB Pentiums. Even the G620 is a good 14% faster than AMD's Athlon II X2 265 and the rest scale up with clock speed. What this translates to is great general use performance as well as solid performance in those apps that are still bound by the performance of a single thread.

Cinebench R10 - Multi-Threaded Benchmark

Multithreaded apps however start to benefit from more cores. In this case the Athlon II X3 455 isn't much faster than a G850 and only 4.5% faster than a G840, despite its core count advantage.

PAR2 Multithreaded Archive Recovery Performance

Par2 is an application used for reconstructing downloaded archives. It can generate parity data from a given archive and later use it to recover the archive

Chuchusoft took the source code of par2cmdline 0.4 and parallelized it using Intel’s Threading Building Blocks 2.1. The result is a version of par2cmdline that can spawn multiple threads to repair par2 archives. For this test we took a 708MB archive, corrupted nearly 60MB of it, and used the multithreaded par2cmdline to recover it. The scores reported are the repair and recover time in seconds.

Par2 - Multi-Threaded par2cmdline 0.4

Our Par2 extraction test is another multithreaded scenario where two Sandy Bridge cores manage to slightly outperform three Athlon II cores. Whereas before it was a pretty simple "more cores, better multithreaded performance" argument, the SNB Pentiums do complicate things a bit.

WinRAR - Archive Creation

Our WinRAR test simply takes 300MB of files and compresses them into a single RAR archive using the application's default settings. We're not doing anything exotic here, just looking at the impact of CPU performance on creating an archive:

WinRAR 3.8 Compression - 300MB Archive

Once again we've got a multithreaded test with a murky outcome. The G850 is hot on the heels of the Athlon II X3 455, as is the G840.

Gaming Performance

Our first set of gaming performance results come using a discrete GPU. In the case of the Crysis Warhead results it's a GeForce GTX 280, while everything else uses a Radeon HD 5870. Across the board the Pentium manages to do better than its Athlon II competitors, of course with a discrete GPU that's not unexpected. Next we'll see how these stack up with processor graphics.

Crysis Warhead - 1680 x 1050 - Mainstream Quality (Physics on Enthusiast) - assault bench

Dragon Age Origins is another very well received game. The 3rd person RPG gives our CPUs a different sort of workload to enjoy:

Dragon Age Origins - 1680 x 1050 - Max Settings (no AA/Vsync)

World of Warcraft needs no introduction. An absurd number of people play it, so we're here to benchmark it. Our test favors repeatability over real world frame rates, so our results here will be higher than in the real world with lots of server load. But what our results will tell you is what the best CPU is to get for playing WoW:

World of Warcraft

Starcraft 2

Power Consumption

The new Pentiums draw very little power, not only compared to their AMD counterparts but also the rest of the Sandy Bridge lineup.Without Hyper Threading, overall execution resource utilization is lower and thus power consumption is lower. It doesn't look like any of these chips come close to hitting their max TDPs as the Pentium G620T only draws 4W less than the G620 despite being rated with a 30W lower TDP.

Load Power Consumption x264 HD 3.03 1st pass (Win7/Radeon HD 5870)

The Matchup Processor Graphics Performance
Comments Locked

110 Comments

View All Comments

  • yo2020 - Tuesday, August 23, 2011 - link

    for both idle and load. :D
  • Havor - Tuesday, August 23, 2011 - link

    For anyone that wants to build a low power system, dont look for a 620T or 2100T, they will only cost you actually more money and save's you almost no energy or even cost you extra if you do heavy loads like encoding.

    After looking up a lot of sites and comparing idle loads i made the following list, and radical changed the dissension on what CPU to use for my server.

    CPU (total system use) Idle/Load
    E-350 26W /38W
    Pentium G620 Undervolt @ 0.9GHz 31W/43W
    Pentium G620 31W /60W
    Pentium G840 31W /60W
    Pentium G850 32W /65W
    Core i3 2100 33W /69W
    Athlon X2 240e 34W/84W
    i3 2120 33W /84W
    i5 2400 34W /99W
    i5 2500K 34W /110W
    i7 2600K 35W /125W
    i7-2600K (4.85GHz) 39W/190W

    The idle difference between a G620T and a i7 2600K is only 4 watts, and yes under load the 2600K uses 2x the amount of watts, but at the same time it also dose 4x the amount of work, so with the "instruction per watt" (IPW) you are better of with a 2500K/2600K

    So now I turn my main system of when i dont need it, and let my server that's 24/7 on anyway, do all the work.

    So if a system is needed that needs to do heavy CPU work like encoding, but you still want a energy lean machine, a fast CPU could not be suds a bad idea after all.

    Just something to think about ;-)
  • Marlin1975 - Tuesday, August 23, 2011 - link

    These are well suited for HTPC but you did not compare them in that regard?

    Very few care about synthetic benchmarks, test these for real world uses like HTPC where many are looking to use low-end cpus like these.
    Also you can still overclock the intel cpus just not much. Why not overclock the A6/A8 and the intel chips to compare as well?
  • dubyadubya - Tuesday, August 23, 2011 - link

    Seems everyone has forgotten about 64 bit programs. There is zero mention of 64 bit programs in most reviews on the net. Testing using mainly 32 bit programs and not stating which programs are 32/64 bit give the impression that Intel's SB is way faster than AMD's offerings. AMD offerings perform much better running 64 bit code than Intel's SB so lets see some multi-threaded 64 bit testing.
  • Traciatim - Tuesday, August 23, 2011 - link

    Benchmark links?
  • A5 - Tuesday, August 23, 2011 - link

    You're going to have to back that up.
  • silverblue - Tuesday, August 23, 2011 - link

    Either you're saying that AMD's CPUs perform better in 64-bit mode than 32-bit mode, or that AMD's CPUs are better at x86-84 than Sandy Bridge. Regardless, as stated, you'll need to provide some evidence to back this up.
  • AngryMonkeyMob - Tuesday, August 23, 2011 - link

    While it is true that AMD can keep competitive on price/performance, when it comes to power consumption AMD gets it the hot bullets of shotgun when compared to Intels baked goods. That matters to some people ( it does to me ). I'd use this in a tiny ITX machine with something like this http://www.logicsupply.com/products/ga620ibk_a1

    ...But I'd imagine that makes me a minority around here :)
  • Death666Angel - Friday, August 26, 2011 - link

    If you do something that specific, surely you know how to undervolt a CPU? :-) AMDs Llano can be undervolted quite heavily as far as I know and the 65W TDP chips (A8-3800 and A6-3600 afaik) will already perform much more energy conscious out of the box. :-)
  • CeriseCogburn - Monday, June 25, 2012 - link

    Since you quibble on 5 or 20 bucks when it comes to a crap apu, and spend minimum $124.00+ on a tiny computer case, yes, I'd say, God Willing, you are in the minority around here, but I fear, that you definitely are not.

    Oh the irony

Log in

Don't have an account? Sign up now