Server Products
Data Center Products including boards, integrated systems, Intel® Xeon® Processors, RAID Storage, and Intel® Xeon® Processors
4761 Discussions

Now a day, how to estimate the speed of a CPU

idata
Employee
1,374 Views

Since by now, Mhz seem not a very reference-able label for a CPU, then how we to estimate / compare the preformance of a CPU ?

0 Kudos
2 Replies
William_L_Intel1
Employee
510 Views

Hi Ming,

Thanks for your question. I think a good place to start would be to look at our http://www.intel.com/performance/ performance website.

Please let us know if this has been helpful to you.

0 Kudos
Christophe_P_Intel
510 Views

Ming, A lot of factors go into evaluating the "speed" of a CPU since the reality is the performance of the overall server the processor goes in is most important to business owners and IT professionals.

In general, some factors that boost performance of the server beyond the raw speed or frequency of the processor.

ð Interconnect Type and Speed: This is the speed at which data is able to move in/out of the processor to memory, the network or between CPUs. The new Intel Quickpath interconnect with Xeon 5500 is measured in GT/s (giga transfers per second). Fastest speed is 6.4GT/s and slowest speeds are 4.8GT/s. Older Xeon processors use the Front Side Bus and speeds were measured in MHz (1333, 1066, 800)

ð Memory Type and Speed: This is the speed at which the server can access onboard memory storage. The new Xeon 5500 servers have an integrated memory controller and speed are at 1333, 1066 or 800 MHz per DIMM.

ð I/O speed: This is the speed at which your processor connects to the network and other input/output devices (storage arrays, etc). the new Xeon 5500 servers use PCI Express Generation 2. PCIe2 runs at 5 GB/s compared to previous PCIe which ran at 2.5GB/s

ð CPU cache: cache is memory storage that is on-die (processor). Cache serves an important role in that is stores instructions and data that the processor needs to process or compute information. In general, both how large and how well architected the cache is will boost performance since the processor does not have to wait for the system to retrieve data from main memory before processing.

ð Hyper-Threading: This feature has been re-introduced with the Xeon 5500 and enables two software threads to run per core on the processor. This allows IT to run more code in parallel so for software that is well threaded (and most is for enterprise server software), applications will run faster, getting more work done per every clock cycle

ð Turbo-Boost: This new feature was added with the Xeon 5500 family. This feature will increase the CPU clock speed when the OS sees a peak demand period, meaning that your processor can run faster when your business demands are the highest.

You can look at the Xeon 5500 family of processors here http://ark.intel.com/ProductCollection.aspx?codeName=33163 http://ark.intel.com/ProductCollection.aspx?codeName=33163

In the end, I believe that delivered performance for a given application environment is the best way to access performance for your environment – however, when testing every application is not practical, there are many industry standard benchmarks to help assess the value and performance for different environments. For links to that information, use Hank's link above.

Chris

0 Kudos
Reply