Intel Pentium G2020: Ivy Bridge with very low power consumption
When not stated otherwise, the values in the charts represent average fps. In almost all cases, the details are maxed out, with the exception of anti-aliasing, which is disabled.
High resolution
Should someone criticize me in the forum for not having the "true" full HD, I would like to point out to him that the difference (in megapixels) between 1680 × 1050 pixels and 1920 × 1080 pixels is a small one, and even more so in the context of a CPU test. You will see that even when the details are set to maximum (with the exception of anti-aliasing, which is provided solely by GPU), the resolution, even when compared to 800 × 600 pixels, for instance, has a rather negligible effect on the results – when using GeForce GTX 680, the selected tests depend heavily on a given processor.
Call of Duty 4: Modern Warfare
This installment of CoD is still the most popular one (maybe together with the second one). I test it via timedemo included in the TpuBench tool (the development of this automated software was discontinued). The details are maxed out, anti-aliasing is disabled.
Enemy Territory: Quake Wars
In the OpenGL ET:QW, I use a time-proven timedemo; the quality is set to "high". This setting will max out almost all of the values, with the exception of lighting (set to "normal") and anti-aliasing (disabled). Soft particles are also disabled; anisotropic filtering is set to 8×. I remove the FPS limit via the console. The results in the 640 × 480 pixels resolution can be worse than in 1680 × 1050 pixels due to different aspect ratio.
Left 4 Dead
A representative of the Source engine is tested with its own timedemo, so I have to be careful about preserving my version of the game (Steam has a strong tendency to always make everything up-to-date, even your settings). I max out the details and set texture filtering to "trilinear" (anti-aliasing is disabled). Importantly, multicore rendering is turned on while vertical synchronization is switched off.
The Elder Scrolls V: Skyrim
I chose two saves for this test – Whiterun (you can see a large part of the city and even certain things beyond the walls) and Riverwood (trees, water, the village itself). The ZIP file contains another save, Ragged Flagon (an underground inn beneath Riften), which I didn't use. After loading a save, I wait for the FPS rate to "calm down" (loading from the disk; it usually takes two or three seconds), and then I measure for ten seconds (using FRAPS). I load the other save and do the procedure once again. The values in the chart represent an average of both ten-second FPS measurements. I used the 1.7.7.0.6 version of the game with no added texture packages or other expansions.
Unreal Tournament 3
Even though UT3 is an ancient game, its engine is still used in an enormous number of games. I test it via the VCTF-Suspense flight through. Details are maxed out (number 5 in the game, which means 16× AF), software physics. I count the better one out of two results.
World in Conflict
Unlike in the previous methodology, I have a GPU with enough to power to set graphic details to the highest setting. This includes DX10 rendering and 4× MSAA. Only the "water reflects clouds" option is disabled. Like ET:QW, World in Conflict also supports FOV and the results obtained with lower resolution and an aspect ratio of 4:3 might be higher than those obtained with the 1680 × 1050 pixels and a ratio of 16:10.
World of Tanks
I run a replay of one of the most demanding and processor-dependent maps – Ensk – in the 8.0 version of a client. I go through a large part of the map (problematic bushes, most of a city) in the rather agile SU-100 tank-destroyer. The sniper mode is employed frequently, mainly at the beginning of the game, to zoom out – in those moments, there are lots of objects on the screen. Rendering Improved is enabled, Graphics Quality is set to "Maximum", and virtually everything is maxed out (with the exception of Enhanced Shadows).
X-Plane 10
In this demanding flight simulator, I use a benchmark script that runs the game with the --fps_test=1 (2, 3) parameter; an average of those three tests is represented in the chart. In Setting, Rendering Options, virtually all details in the upper three sections are maxed out; the number of objects and pathways is set to "tons". The airport details are set to default, the water reflections details and "HDR anti-aliasing" are disabled. Clouds are also set to 100%. I use the 10.05r1 version of the game.
Low resolution
The second set of game tests was done in the lowest resolution possible, but details were almost always maxed out (in case CPU participated on one of them).
Skyrim setting:
Values in the chart represent an average of performances in Whiterun and Riverwood:
Summary
All game tests are included in the final summary game chart; I included values from both high-resolution and low-resolution measurements. Of course, minimum FPS measured in World in Conflict and World of Tanks are not included.
Related articles
| 13. 9. 2013 Comments (0) | 31. 7. 2013 Comments (0) |
| 31. 7. 2013 Comments (0) | 31. 7. 2013 Comments (0) |





















Comments
Ivy Bridge G2020
This review is even better than Tom's Hardware! Excellent! G2020 beats FX6100 and equals FX6300, amazing.
This is an awesome review.
This is an awesome review. Been scouring for a review like this because I'm going to have my first gaming PC. Hahahaha. I'm so poor.
cpu
btw a fx 6300 destroy any pentium i3 offer from intel in gaming.....
with my 7950 bf4 at ultra 50fps + with direct x and with mantle 60-120 fps
Add new comment