Blog Post

TPC-E Benchmark Results for Intel Xeon E5-2697 v2

,

Two new TPC-E benchmark results have shown up for two-socket database servers using the brand new, 22nm Intel Xeon E5-2697 v2 (Ivy Bridge-EP) processor. The first result is for an IBM System x3650 M4 with 512GB of RAM, that has a score of 2590.93. The second result is for a Fujitsu Primergy RX300 S8 with 512GB of RAM, that has a score of 2472.58.

These two systems are shown in Table 1, along with the best result for an earlier, two-socket, HP ProLiant DL380p Gen 8 system with 256GB of RAM, with a score of 1881.76, using the earlier eight-core, 32nm Intel Xeon E5-2690 (Sandy Bridge-EP) processor.

VendorProcessorTPC-E ScoreScore/Core
IBME5-2697 v22590.93107.96
FujitsuE5-2697 v22472.58103.03
HPE5-26901881.76117.61

Table 1: Comparing Three Different TPC-E Results

These results show pretty linear scalability as we move from an eight-core processor, to the newest twelve-core processor (that is actually running at a slightly lower clock speed). This gives me more confidence about running even bigger workloads on a new two-socket system, and getting better performance and much lower SQL Server 2012 license costs than you can get from any current, four-socket database server.

The Xeon E5-2697 v2 processor has twelve physical cores, plus hyper-threading, for a total of 24 logical cores per processor. It runs at a base clock speed of 2.7GHz, with the ability to Turbo Boost up to 3.5GHz.

This processor is the new “top-of-the-line” Xeon E5-2600 v2 series processor. If you want to minimize your SQL Server 2012 licensing costs, you can still pick a Xeon E5-2600 v2 series processor with a lower physical core count, such as a six-core Xeon E5-2643 v2.

Keep in mind that you can only use 16 physical cores with an instance of SQL Server 2012 Standard Edition, so a new, two socket server with these 12-core processors will go over that limit.

Filed under: SQL Server 2012, TPC-E Tagged: Ivy Bridge-EP

Rate

You rated this post out of 5. Change rating

Share

Share

Rate

You rated this post out of 5. Change rating