Skip to main content

Comparison of CPU performance between minicomputer and x86 server

Yesterday I posted an article comparing the performance of different CPU DMIPS. It happened that a friend in the group who was an Oracle asked about the performance difference between minicomputers and x86 servers in terms of database performance, which also attracted the attention and discussion of many people, so by the way today Write another article about the CPU performance comparison of minicomputers and x86 servers. By the way, I also do a survey to see what everyone thinks about this issue:

For many comprehensive benchmark performance tests, especially for the TPC-C database type Benchmark test, the test is actually the performance of the whole machine or the overall system, not just a single aspect of software and hardware. Performance, the test results are not suitable for direct use as a basis for hardware configuration selection, but if it is only used as a horizontal performance comparison, it is still very suitable. Borrowing a film from Oracle on OOW Shanghai in 2013 to compare the performance of servers with different CPUs in an OLTP database environment:
The first one is Oracle's SPARC SuperCluster. Although the performance value of tpmC is as high as 30M, it consists of 27 4-way SPRAC T3-4 minicomputers. If the overall system's 30M performance is averaged to a single machine, it is actually Not much.
The second place is the IBM Power 780 system, with three nodes, each node is configured with 8 Power 7 CPUs. The entire system has a total of 24 CPUs, 192 cores and 768 threads. The test performance of the topic exceeds 10M tpmC. If the average is counted as a single machine If it is only 3 o'clock M.
Although the previous two models occupy the top two positions in the tpmC performance ranking, they are all cluster systems and rely on more to win. Although Oracle's newly launched minicomputer T5-8 in 2013 only achieved a time score of 8.55M, it still claims that it has set a world record for "single system" OLTP performance.
The highest-ranking of the x86 server on this list is Oracle's 8-way server 4800. This machine is equipped with 8 Intel Xeon E7-8870, each with a 10-core 2.4G frequency. It is also used in the Oracle Exadata database. Onboard, the better-known name is X2-8. Its stand-alone tpmC score reached 5.05M. As the leader of the x86 server, the performance of the 8-way E7 server is second only to the flagship models of the two minicomputer manufacturers.
On the low-end 2-way server models, the tpmC performance of the x86 server significantly exceeds that of the minicomputer. The x86 server using 2-way E5 2690 CPU can reach 1.6M, while the Power780 using 2-way Power7 can only reach 1.2M , The performance difference between the two is obvious, and if you count the price between them, the difference in cost performance between the two is even greater.
In fact, the data on the Oracle film comes from public information on tpc.org. More TPCC performance test results can also be found at http://www.tpc.org/tpcc/results/tpcc_results.asp?print=false&orderby =tpm&sortby=desc, but in the following years, Sparc T5-8 has not been updated, HP Itanium has disappeared, and now only IBM Power is still struggling to update products, but the latest The Power8 system has never been tested for tpmC performance, but only disclosed the performance values ​​of some models of SPEC CPU2006:
The most high-end E880 in the Power8 system, SPEC CPU2006 performance with 8-way and 16-way CPU configurations:
E870 has only 8 channels in a configuration, and released the SPEC CPU2006 performance value:
On spec.org, we can also query the SPEC CPU2006 performance values ​​of 8-way and 16-way servers using Intel’s latest Xeon E7-8890 v4 CPU as follows:
A simple comparison of the two performance values ​​is not difficult to draw the following conclusions:
Power is still slightly better than x86 on the highest-end models, but the performance of the two is very, very close. On 8-way (and below) models, the performance of x86 servers has surpassed PowerPower’s main frequency and multi-threading capabilities. Stronger, but the number of cores has fallen behind x86 a lot. As Intel Xeon keeps the Tick-Tock rhythm constantly updated, I believe that x86's performance over Power on the highest-end models is just around the corner, and this surpassing time point may happen. In terms of the next-generation products that Intel and IBM will release this year, of course, it is still too early to say these. Let's wait and see the performance of E7 v5 and Power 9.
The IBM Power system now uses the so-called reference performance value (rPerf) to indicate the performance of different models, which is expressed by the multiple of the performance of this model relative to the comprehensive performance of IBM eServer? pSeries? 640 (rPer value of prizes 640=1.0), The performance between different models can also be converted and compared by rPerf performance value. For example, the maximum performance of the following E850C system is 853 times that of the P640 system, while the maximum performance of the E870 system is 1712 times that of the P640, and the maximum performance of the E880 is that of the P640. 3905 times.

Comments

Popular posts from this blog

AMD's GPU technology enters the mobile phone chip market for the first time

In addition to the release of the Exynos2100 processor, Samsung also confirmed a major event at this Exynos event, that is, the custom GPU that they have worked with AMD for many years will soon appear and will be used on the next flagship machine. The current Exynos2100 processor uses ARM’s Mali-G78GPU core with a total of 14 cores, so the GPU architecture developed by Samsung will be the next Exynos processor, and the GPU will be the focus. This is probably the meaning of Exynos2100’s GPU stacking. The key reason. Dr. InyupKang, president of Samsung’s LSI business, confirmed that the next-generation mobile GPU in cooperation with AMD will be used in the next flagship product, but he did not specify which product. Samsung is not talking about the next-generation flagship but the next one, so it is very likely that a new Exynos processor will be available this year, either for the GalaxyNote21 series or the new generation of folding screen GalaxyZFold3. In 2019, AMD and Samsung reached

Apple and Intel want to join the game, what happened to the GPU market?

Intel recently announced that it will launch Xe-LP GPU at the end of this year, officially entering the independent GPU market, and will hand over to TSMC for foundry. At the 2020 WWDC held not long ago, Apple also revealed that it is possible to abandon AMD's GPU and use a self-developed solution based on the ARM architecture. It will launch a self-developed GPU next year. What happened to the GPU market? Why are the giants entering the game?    Massive data calls for high-performance GPU    Why has the demand for GPUs increased so rapidly in recent years? Because we are entering an era where everything needs to be visualized. Dai Shuyu, a partner of Aiwa (Beijing) Technology Co., Ltd., told a reporter from China Electronics News that visualization requires a large amount of graphics and image computing capabilities, and a large amount of high-performance image processing capabilities are required for both the cloud and the edge.    Aiwa (Beijing) Technology Co., Ltd. is an enterp

SSD vs. HDD: that is nice for You?

Deciding on the proper garage isn’t just about comparing ability and value. The type of garage your pc uses topics for overall performance, inclusive of electricity utilization and reliability. Stable country drives (SSDs) and difficult disk drives (HDDs) are the 2 major storage options to consider and it’s critical to realize the satisfactory use for each and how they compare facet by using aspect. What is an HDD? An HDD is a statistics garage tool that lives within the pc. It has spinning disks inside where statistics is saved magnetically. The HDD has an arm with several "heads" (transducers) that study and write statistics at the disk. It is much like how a turntable file participant works, with an LP record (tough disk) and a needle on an arm (transducers). The arm moves the heads throughout the surface of the disk to get admission to distinct information. HDDs are considered a legacy generation, that means they’ve been round longer than SSDs. In popular, they may be dec