Comparison of processor speeds for dedicated servers
(on the graph on the x-axis logarithmic scale)
During the discussion of the announcement of rented servers based on atoms at 1500 rubles, the question arose how much slower is the atom than other platforms.
The question turned out to be interesting. Here are the test results. We used sysbench as a test (it is available in many Linux distributions, including Ubuntu and Squeeze).
The tests distinguish between two cases - single-threaded and multi-threaded load. A typical load on a visited web server is multi-threaded. A typical load for a single single-threaded application (for example, gzip on big data) is single-threaded.
')
It is clearly seen that atoms substantially lose their Core / Xeon processors, which in a single thread turn out to be one and a half to two times faster than an atom with two cores and hypertreaming.
Another interesting observation is that the 32-bit and 64-bit modes on the atom show themselves the same, on all other processors the 64-bit architecture significantly outperforms the 32-bit one.
As a benchmark, we used nice -20 sysbench - test = cpu --cpu-max-prime = 40000 --num-threads = X run (X is the number of threads, from 1 to 16).
On the graph, this data. Note that the semilog scale (in the linear difference between the atoms and the rest was so great that comparing the other processors with each other did not make sense). It is also clear that CoreDuo is the fastest single-threaded version (which is understandable, because it is a desktop processor, sharpened for resource-intensive single-threaded applications, like games). Xeon'y show their speed with maximum parallelization of tasks performed (typical server load).
Of particular interest is the increase, which gives HT (with HT two models - Atom D150 and Xeon 5620):
This comparison applies only to processors and does not fully characterize the platform. Atoms come with 2 GB of memory, crusts - 6 GB, zeons - 12 and 24 GB, respectively; also different equipment discs.
PS Since the logarithmic scale caused many questions, here’s another view of the same data: processor performance is relatively better (that is, 100%). The scale is linear.
During the discussion of the announcement of rented servers based on atoms at 1500 rubles, the question arose how much slower is the atom than other platforms.
The question turned out to be interesting. Here are the test results. We used sysbench as a test (it is available in many Linux distributions, including Ubuntu and Squeeze).
The tests distinguish between two cases - single-threaded and multi-threaded load. A typical load on a visited web server is multi-threaded. A typical load for a single single-threaded application (for example, gzip on big data) is single-threaded.
')
It is clearly seen that atoms substantially lose their Core / Xeon processors, which in a single thread turn out to be one and a half to two times faster than an atom with two cores and hypertreaming.
Another interesting observation is that the 32-bit and 64-bit modes on the atom show themselves the same, on all other processors the 64-bit architecture significantly outperforms the 32-bit one.
As a benchmark, we used nice -20 sysbench - test = cpu --cpu-max-prime = 40000 --num-threads = X run (X is the number of threads, from 1 to 16).
| Model | Atom D510 1.66 GHz | Core2Duo E8400 3.0GHz | Core2Quad Q8300 2.5GHz | 2 x Xeon 5504 2.0 GHz | 2 x Xeon 5620 |
| Core | 2 + HT | 2 | four | eight | 8 + HT |
| Single threaded computation (32/64) | 412.83 / 412.82 | 61.89 / 54.98 | 73.47 / 66.03 | 90.31 / 83.07 | 75.28 / 76.43 |
| Multithreading (32/64) | 124.55 / 124.57 | 30.91 / 27.49 | 18.43 / 16.56 | 11.29 / 10.45 | 8.91 / 7,21 |
On the graph, this data. Note that the semilog scale (in the linear difference between the atoms and the rest was so great that comparing the other processors with each other did not make sense). It is also clear that CoreDuo is the fastest single-threaded version (which is understandable, because it is a desktop processor, sharpened for resource-intensive single-threaded applications, like games). Xeon'y show their speed with maximum parallelization of tasks performed (typical server load).
Of particular interest is the increase, which gives HT (with HT two models - Atom D150 and Xeon 5620):
| CPU | Atom D510 1.66 GHz | 2 x Intel Xeon 5620 |
| without HT (multithreaded) | 206.47 | 11.35 |
| with HT (multithreaded) | 124.57 | 7.21 |
| Speed increase (%) | 65.7% | 57.4% |
This comparison applies only to processors and does not fully characterize the platform. Atoms come with 2 GB of memory, crusts - 6 GB, zeons - 12 and 24 GB, respectively; also different equipment discs.
PS Since the logarithmic scale caused many questions, here’s another view of the same data: processor performance is relatively better (that is, 100%). The scale is linear.
Source: https://habr.com/ru/post/107378/
All Articles