How CPU Features Affect CPU Performance, Part 5
|
3D rendering
Without Turbo Boost and Hyper-Threading
|
1 core |
2 cores |
3 cores |
4 cores |
3ds max ↑ |
4.71 |
9.17 |
95% |
12.44 |
36% |
15.6 |
25% |
Lightwave ↓ |
332.32 |
182.23 |
82% |
124.98 |
46% |
94.48 |
32% |
Maya ↑ |
10:35 |
05:22 |
97% |
03:36 |
49% |
02:44 |
32% |
Group Score ↑ |
40 |
77 |
93% |
110 |
43% |
143 |
30% |
Rendering is the quietest group of our tests. It's hard to imagine that this operation may cause problems with performance scalability as the number of cores grows. Especially as rendering is not sensitive to insignificant fluctuations of cache size (see the explanations above). However, we can see no super-linear performance growth here, which was demonstrated by Phenom II X4 in 3ds max and Lightwave.
With Turbo Boost and Hyper-Threading
|
2 cores |
4 cores |
6 cores |
8 cores |
3ds max ↑ |
6.02 |
11.61 |
93% |
15.08 |
30% |
18.69 |
24% |
Lightwave ↓ |
275.89 |
136.88 |
102% |
94.06 |
46% |
71.5 |
32% |
Maya ↑ |
08:25 |
04:07 |
104% |
02:50 |
45% |
02:09 |
32% |
Group Score ↑ |
50 |
100 |
100% |
140 |
40% |
181 |
29% |
Super-linear performance growth appears as soon as we switch to the normal mode. But the set of applications is a little different here: there is Lightwave, as in case of Phenom II X4, but it's accompanied with Maya instead of 3ds max. As the number of core is doubled with Hyper-Threading, 3ds max rendering engine (V-Ray) starts having difficulties with scalability -- it's noticeably worse than in Maya and Lightwave.
Compared to AMD Phenom II X4
However, even in this group, despite Intel HT technology, which performs well in rendering, Core i7 fails to outperform Phenom II X4 in relative scalability in any mode. We already discussed a possible reason in the analysis of results in the previous group of tests.
Write a comment below. No registration needed!
|
|
|
|
|