---

Results

Results are broken down into subcategories. 2D Performance covers ‘ye olde’ x86 instructions. System Performance runs a number of newer whole-system tests. Workstation focuses on using the system for creating digital content while 3D Synthetic is to measure up to other devices in relation to gaming. Media Encoding is very important for many systems but especially for one an ITX board that could very well be ending up in an HTPC. Lastly Bootracer, just because boot speed is tied to much more than just storage devices.

2D performance

SuperPi 1.5 Mod

To measure 2D performance we include both SuperPi 1M and 32M tests. The 1M test pushes the CPU and cache to it fullest potential as a single threaded app. 32M brings in memory bandwidth as a major factor. Stock and OCed results are included.

wPrime 1.55

wPrime is a multithreaded 2D app that really pushes a CPU to its limits. Memory bandwidth has less of an effect on these tests and too tight can actual degrade performance. 1.55 is used as it is accepted at hwbot for comparison with other systems.

2D Results

Limited to CPU mostly with some benefits from increase memory speed, lower latency and increased cache speed, these tests perform slightly better than the OC percentage. Bouncing between 3 and 4% better performance than stock.

3D Synthetic tests

3dMark Vantage

3DMark Vantage is the industry standard performance benchmark for DirectX 10 gaming PCs. It includes two graphics tests, two CPU tests and six feature tests. Trusted by gamers worldwide to give accurate and unbiased results, 3DMark Vantage is the best way to consistently and reliably test DirectX 10 under game-like loads.

Heavily GPU dependent a high CPU score will not effect the over all total very much. Focuses on DX10 gaming performance and CPU physics. CPU score has a minor role in determining total score. Testing Consisted of Performance results only on the IGP.

3d Mark 06

3DMark06 is a PC benchmark suite designed to test the DirectX9 performance of your graphics card.

A 3DMark score is an overall measure of your system’s 3D gaming capabilities, based on comprehensive real-time 3D graphics and processor tests. By comparing your score with those submitted by millions of other gamers you can see how your gaming rig performs, making it easier to choose the most effective upgrades or finding other ways to optimize your system.

An Oldie but a goodie. 3dMark 06 focuses on dx9 performance and is heavily influenced by CPU speed as GPU capabilities increase. CPU score also has much more of an effect on the total score as well.

3D Results

With expectations of ~30% improvement for GPU tests when OCed 3d06 was no shocker there.  3d Vantage showed a 40% increase in performance and this was quite a surprise. Vantage CPU clocks were nearly identical OCed or not.

System Tests

PCMark Vantage

New enough to still be pertinent and provide accurate storage test results. Taking over an hour the PCMark Suite focuses on all aspects of a systems performance from CPU to GPU to storage and memory.

Because of the wide range of tests Vantage performs Overclock results are all over the place. From reduced performance in the Music category, almost no change in HDD. Productivity, TV and Movies hovered around system overclock of 3%.  Memories increased almost 5%. Gaming and communications saw the biggest increase with 7+% higher performance.

Performance Test 7

A newcomer to our lineup PT7 performs a variety of 2D, 3D, memory and storage tests. Passmark includes a break down of each test and includes graphs to compare results for each of the subtests. Check them out at passmark.com

This test was a bit more consistent, performing around the 3-5% mark in all tests. 3D Gaming was right where it should be approaching 30%

AIDA64 Engineer Edition

This test has been around for many years under a few names. AIDA32, then Everest Ultimate and now AIDA64. This test focuses mainly on CPU and Memory performance and includes “standard” results for other CPUs in the results so the end user can simply compare them to other types of stock systems.

Stressing only CPU or Memory (rarely both) AIDA64 show linear improvement of 3% across the board from FPU, to CPU and Memory tests. Latency actually improved slightly more at 5%.

Workstation Tests

Editing of digital photos, home movies and music is covered under system performance. A rise in the digital arts leads to many people creating their own 3D content. These tests usually focus on CPU performance but may also include 3D performance, usually in the form of OpenGL/OpenCL testing.

Cinebench 11.5 x64

Created by Maxthon as a means of evaluating a system capabilities for running their own Cinema 4D software. There are no direct comparison charts at maxon.net however cbscores.com does host a searchable chart for both render and OpenGL results. Single and multithread CPU tests are displayed as well as the OpenGL benchmark for video.

SpecViewPerf11

Developed by SPEC.org SVP11 focuses on a number of different 3D/4D applications to give an idea of overall system performance. It requires at least 512MB of Vram and 1080P resolution monitor to run. a Full breakdown of the tests can be found at their website.. Catia-03 will not run Intel HD graphics at default settings and thus failed to complete the test.

Workstation Results

All of the testing improved with the OC. CPU scores in cinebench as well as the PROE-05 and TCVIS-02 tests were completely CPU bound. As such their performance was at or less than the overclock of 3% on the CPU. Most of the other testing scaled better although not to the point we would have hoped considering the 32% Overclock on the GPU. OpenGL fared best with a nearly 30% performance increase while the rest of the 4D tests hovered around 5%. SW-02 had a nice boost of just over 12%.

Media Encoding Tests

Media Espresso 6.5

Media Espresso is a simple to use video and music encoding software that is capable of taking advantage of Quick sync features for faster HW decoding and encoding. Focused at converting HD content quickly for use uploading a to the web, or downloading to a portable device, as such stereo audio is the only option at this point. However transcoding can be done at up to 1080P using a custom code set.

Handbrake .94

A free open source software used for converting files to another format. Includes a number of presets like paid software but is highly customizable for serious encoders. Supports multi channel audio as well as AC3 outputs. Great free software that supports up to 8 threads but does not support Intel Quick Sync. profiles used were ipod classic (320×180), Ipodtouch (640×480), AppleTV2 (1280×720) and High profile (1920×1080).

Big Buck Bunny Results

Using the Big Buck Bunny 1080P surround file downloaded from bigbuckbunny.org. We have included filesizes in the chart below to show how close were were keeping our presets. As well as any flaws appearing one form of encoding or another. This test will probably see a few more changes over the next few months as well.

Times stay grouped together rather well for Media Espresso.  Enabling Quicksync improves times and slightly reduced filesize on most tests. Quicksync at 720P despite running almost 35 seconds faster than CPU only ME encoding, results in the same file size as the 1080P file. Or double the size of what Media Espresso without Quick Sync being enabled.

Handbrake just rocks the low res encoding with slightly better filesizes and drastically improved times. Over 3 times faster than QuickSync at 320×180 resolution. However the opposite happens at 1080P settings where handbrake slows to over 14 minutes compared to Quick Syncs 4’ 12”.

TechARP x264 4.0

A new test used here at TechREACTION. Designed to benchmark media encoding performance it performs two dual pass encoding runs and results can be compared to others performance at techarp.com Download the benchmark here. It uses the OpenSource x264 codec version 1913 and supports parallel encodes (multi-threaded). Results are not comparable to other media encoding tests and so are charted separately.

BootRacer

Bootracer monitors the time it takes from OS initialization to the login screen and follows it up with time to usable desktop. Bear in mind that if you use a password for login it does NOT affect the results in any way that was noticeable in these tests.

This software was deemed more accurate than using a stopwatch to measure loading time and why power off to desktop times will be longer than those used for this application. Software installs and add-on hardware devices will add to the time shown in these tests.

Time to Logon remained unchanged among the Best bootracer time. Total time to desktop OC improved by 10% knocking 3 full seconds off the clock.

 

Next: This is the End

Pages: 1 2 3 4 5 6 7