Call of Juarez DX10
This relatively old game with Direct3D 10 support cannot use multiple cores at all, as we found out in our previous review (we actually used a D3D9 version then). Performance may be limited by the speed of computing units in a CPU because of the relatively big number of draw calls. But Call of Juarez is mostly limited by a graphics card. It carries the heaviest load, while a CPU works on simple physics and AI.
As we can see, nothing has changed since our previous tests in this game -- no significant differences between configurations with one, two, three, and four cores. A measly difference can be explained with a reduced load on a processor by system and background processes. In other respects, the game proves that its rendering performance almost fully depends on graphics card's capacity.
But let's analyze the load on each core separately, to make sure we are right:
|
Core 1 |
Core 2 |
Core 3 |
Core 4 |
Average |
60.6 |
2.9 |
2.7 |
5.0 |
Maximum |
87.9 |
16.4 |
19.5 |
35.0 |
The numbers speak for themselves. Only one core of the processor was loaded by our tests (the average usage of 3-5% for the other cores is almost nothing). Besides, that only core was working at 2/3 of its full capacity. And the load never jumped up to 100%. So performance in this game is not limited by CPU speed, not even by a single core. This is essentially a single-threaded application, so it runs fine on a single-core processor. The total CPU usage below 75% is another proof of this fact.
Company of Heroes: Opposing Fronts
Perhaps, games of other genres (RTS) will benefit from multiprocessor configurations, and we'll see higher performance gains. Let's take a look at the benchmark in Company of Heroes, which also got Direct3D 10 support in its Opposing Fronts addon. Unfortunately, the built-in benchmark does not show gaming performance, as it runs a script scene, which does not resemble the game itself. But it will be still interesting to see performance differences between different configurations.
A single 2.4GHz core in our Core 2 processor is apparently not enough here. The second core increases the speed by 15-30%. So we can draw a conclusion that performance of script scenes in Company of Heroes: Opposing Fronts grows, if a dual-core processor is installed. Perhaps, performance gains partially originate from driver optimizations and more efficient distribution of background and system processes, even if we speak of truly multithreaded applications. Let's analyze the CPU usage:
|
Core 1 |
Core 2 |
Core 3 |
Core 4 |
Average |
18.3 |
21.0 |
57.1 |
35.4 |
Maximum |
42.1 |
50.6 |
100 |
72.5 |
These values prove that it's a multi-threaded application -- all four cores are loaded. Even if the core loads differ -- two cores do more work than the others. On the whole, the game runs well on two cores in our test processor, the total CPU usage recalculated for a single core (Core 2 architecture) was a tad above 130%. That is, the application really benefits from the second core (a tad higher frame rate). However, further increase of the number of cores makes no sense for this game.
Need for Speed: ProStreet
Need for Speed: ProStreet is one of the games from a famous series, based on a multiplatform engine. It's notable for good parallelizing, which is important for game consoles. Unfortunately, this game also lacks benchmarking tools and options to record and repeat replays, so we have to use the FRAPS method with a bigger measurement error. On the other hand, the game is very interesting for our article -- it's a multiplatform game. So it should benefit from multi-processor systems. We couldn't miss it.
There is a noticeable difference between single-core and other configurations even in the heaviest mode. Even three cores are faster than two. However, the difference is not so big there, and we should treat it with caution because of measurement errors.
The triple- and quad-core processors have almost a twofold advantage over the single-core processor in a lower resolution. In other cases there is also a noticeable difference between them -- 20-40%, which apparently shows that the game uses several threads distributed by the system between CPU cores. Let's see how effective this distribution was:
|
Core 1 |
Core 2 |
Core 3 |
Core 4 |
Average |
89.4 |
66.3 |
57.2 |
68.9 |
Maximum |
100 |
84.8 |
81.6 |
90.8 |
Well, that's another game that apparently needs more than two cores in a 2.4GHz Core 2 processor. It holds true even with high graphics quality settings. All four cores were loaded, one of them was used up to 100%, the other one -- up to 90%. The other cores are less loaded, but the average usage was still above 50%.
It makes it more than 280% for a single core. That is the game fully utilizes three out of four cores in our Core 2 Quad processor. So Need for Speed: ProStreet needs a triple- or a quad-core processor. The difference appears with complex gaming and graphics settings.
Write a comment below. No registration needed!