Multi-generation Intel processor performance comparison

Almost every year a new generation of Intel Xeon E5 CPUs comes onto the market. In each generation alternately change the socket and the process. The nuclei are getting bigger and bigger, and the heat release is gradually decreasing. But a natural question arises: “What does the new architecture give to the end user?”

For this, I decided to test the performance of similar processors of different generations. I decided to compare the models of the mass segment: 8-core processors 2660, 2670, 2640V2, 2650V2, 2630V3 and 2620V4. Testing with a similar spread of generations is not entirely fair, because between V2 and V3 there is a different chipset, the memory of the new generation with a greater frequency, and most importantly, there are no direct peers in frequency among the models of all 4 generations. But, in any case, this study will help to understand the extent to which the performance of new processors has increased in real-world applications and synthetic tests.

The selected processor line has many similar parameters : the same number of cores and threads, 20 MB SmartCache, 8 GT / s QPI (except 2640V2) and the number of PCI-E lines equal to 40.

To assess the feasibility of testing all processors, I turned to the results of the PassMark tests .
')
Below is a summary schedule of results:



Since the frequency is significantly different, comparing the results is not entirely correct. But despite this, the findings suggest the move:

1. 2660 is equivalent in performance to 2620V4
2. 2670 exceeds the performance of 2620V4 (obviously, due to the frequency)
3. 2640V2 sags, and 2650V2 hits everyone (also because of the frequency)

I divided the result into frequency and got a certain performance value at 1 GHz:



Here, the results have already turned out to be more interesting and visual:

1. 2660 and 2670 - an unexpected run for me within the framework of one generation, the 2670 justifies only that its overall performance is very high
2. 2640V2 and 2650V2 - a very strange low result, which is worse than the 2660
3. 2630V3 and 2620V4 - the only logical growth (apparently just due to the new architecture ...)

After analyzing the result, I decided to weed out some of the uninteresting models that have no value for further testing:

1. 2640V2 and 2650V2 - intermediate generation, and not very successful, in my opinion - I remove from candidates
2. 2630V3 is an excellent result, but it costs unreasonably more expensive than 2620V4, considering the similar performance and, moreover, this is the outgoing generation of processors
3. 2620V4 is an adequate price (comparing with 2630V3), high performance and, most importantly, it is the only model of the latest-generation 8-core processor with Hyper-threading in our list, so we definitely leave it for further tests
4. 2660 and 2670 - an excellent result in comparison with the 2620V4. In my opinion, the comparison of the first and last (at the moment) generation in the Intel Xeon E5 line is of particular interest. In addition, we have enough stocks of first-generation processors in stock, so for us this comparison is very important.

The cost of servers based on 2660 and 2620V4 processors may differ by up to 2 times not in favor of the latter, therefore, by comparing their performance and choosing a server on V1 processors, you can significantly reduce the budget for purchasing a new server. But I will tell about this offer after the test results.

For testing, 3 stands were assembled:

1. 2 x Xeon E5-2660, 8 x 8Gb DDR3 ECC REG 1333, SSD Intel Enterprise 150Gb
2. 2 x Xeon E5-2670, 8 x 8Gb DDR3 ECC REG 1333, SSD Intel Enterprise 150Gb
3. 2 x Xeon E5-2620V4, 8 x 8Gb DDR4 ECC REG 2133, SSD Intel Enterprise 150Gb

PassMark PerformanceTest 9.0

When selecting processors for tests, I already used the results of synthetic tests, but now it is interesting to compare these models in more detail. Comparison made by groups: 1st generation vs. 4th.



A more detailed test report allows us to draw some conclusions:

1. Mathematics, incl. and floating point, mainly dependent on frequency. The difference of 100 MHz allowed the 2660 to outperform the 2620V4 in computational operations, in encryption and compression (and this despite the significant difference in the memory frequency)
2. Physics and calculations using advanced instructions on the new architecture are performed better, despite the low frequency
3. Well, of course, the test using memory was in favor of the V4 processors, since in this case different generations of memory competed - DDR4 and DDR3.

It was synthetic. Let's see what specialized benchmarks and real applications will show.

7ZIP Archiver

Here the results echo the previous test - direct reference to the frequency of the processor. It does not matter that a slower memory is installed - the V1 processors confidently take precedence over the frequency.

CINEBENCH R15

CINEBENCH is a benchmark for assessing the performance of a computer for working with a professional animation program MAXON Cinema 4D.



Xeon E5-2670 pulled in frequency and beat 2620V4. But the E5-2660, which has a not so visible advantage in frequency, lost to the 4th generation processor. Hence the conclusion - this software uses useful additions to the new architecture (although everything is possible in memory ...), but not so much that it was the decisive factor.

3DS MAX + V-Ray

To evaluate the performance of processors when rendering in a real application, I took a bunch: 3ds Max 2016 + V-ray 3.4 + real scene with multiple light sources, specular and transparent materials, and an environment map.



The results were similar to CINEBENCH: The Xeon E5-2670 showed the lowest rendering time, and the 2660 could not bypass the 2620V4.

1C: SQL / File

At the end of testing I attach the results of gilev tests for 1C.



When testing a database with file access, the E5-2620V4 processor is confidently leading. The table shows the average values ​​of 20 runs of the same test. The difference between the results of each stand in the case of the file base was no more than 2%.

The single-threaded SQL database test showed very strange results. The difference turned out to be insignificant, considering the different frequencies of the 2660 and 2670, and the different frequencies of the DDR3 and DDR4. There was an attempt to optimize the SQL settings, but the results were worse than they were, so I decided to test all the stands on the basic settings.



The results of the multi-threaded SQL test turned out to be even stranger and more controversial. The maximum speed of 1 stream in MB / s was equivalent to the performance index in the previous single-threaded test.

The next parameter was the maximum speed (of all flows) - the result was almost identical for all stands. Since the results of different runs strongly fluctuated (+ -5%) - sometimes they were at different stands with a significant margin in one way or the other. The same average results of a multi-threaded SQL test lead me to 3 thoughts:

1. This situation is caused by a non-optimized SQL configuration.
2. SSD became the bottleneck of the system and did not allow processors to overclock
3. There are almost no differences between the frequency of memory and processors for these tasks (which is highly unlikely)

If you have reliable explanations for such results, please share them in the comments.

Also, the result on the parameter “Recommended number of users” turned out to be inexplicable. The average result for the 2660 turned out to be the highest - and this is with the low results of all tests.
On this issue I will also be glad to see your comments.

findings

The results of several versatile computational tests showed that the processor frequency in most cases turned out to be the most important generation, architecture, and even the memory frequency. Of course there is a modern software that uses all the improvements of the new architecture. For example, video transcoding is sometimes done including. using AVX2.0 instructions, but this is specialized software — and most server applications are still tied to the number and frequency of cores.

Of course, I do not declare that there is absolutely no difference between the processors, I just want to note that for certain applications there is no point in a “planned” transition to a new generation.

If you do not agree with me or you have suggestions for testing, the stands have not yet been disassembled, and I will be happy to test your tasks.

Economic benefit

As I wrote at the beginning of the article - we offer a line of servers based on Xeon E5 processors of the first generation, which at a cost significantly lower than the servers on the E5-2620V4.
These are the same new servers (not to be confused with used ones) with a 3 year warranty.

Below is a rough calculation:

The STSS Flagman RX227.4-008LH server in a 2 x Intel Xeon E5-2620V4 + 8 x 8Gb DDR4 ECC REG configuration at retail costs 265065 rubles today.

A similar configuration of STSS Flagman EX227.3-008LH based on 2 x Intel Xeon E5-2660 + 8 x 8Gb DDR3 ECC REG per share is available for 175275r.

Habr's readers can get an additional 5% discount when ordering. To do this, select the desired form factor of the case from the list of models on our website . The EX217.3-004LH is built in a 1U package, the EX227.3-008LH is 2U, and the EX240.3-008LH is built on the base of a Tower / 4U case.
In the model configurator, you can select the necessary parameters for memory, disk subsystem and additional devices. When sending an application for a calculation, you must specify the promotional code HABRAHABR .

Thanks for attention! I will wait for your comments and suggestions on the tests.

Article writing and testing: Usikoff
Testing 1C: sarge74