Pkafkas
asked on
How to judge which processor is faster?
Hello:
At my work we are researching a new Server to purchase and we are looking a a Dell R730 server. We have a quote from last year and this year. The processors for both quotes are listed below.
2016: Intel Xeon E5-2660 v3 2.6 Ghz, 25M Cache, 9.60GT/s QPI, Turbo, HT, 10C/20T (105W) Max Mem 2133Mhz (338-BFFG).
Upgrade to 2 Intel Xeon E5-2660 V3 2.6Ghz… (374-BBGN)
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
My question is, how can we tell which of the 2 processors provides better performance? Are there other factors to include as well such as system board?
A brief summary is as follows:
2016 - 10 cores per socket
2017 = 14 cores per socket
2016 - Version 3 with 2.6 Ghz
2017 - Version 4 with 2.0 Ghz
In the past we have been burned by purchasing new ESXi hosts that performed much slower with our applications than older models. For example a Dell T710 with 6 cores per socket and 2.8 Ghz out performs an HP ProLiant Gen9 with 8 cores per socket and 2.6 Ghz. I do not wish to purchase a server that will provide lower performance than the Dell T710 (6 yers old server).
How can we judge how the processors or how the server will perform compared to another server?
At my work we are researching a new Server to purchase and we are looking a a Dell R730 server. We have a quote from last year and this year. The processors for both quotes are listed below.
2016: Intel Xeon E5-2660 v3 2.6 Ghz, 25M Cache, 9.60GT/s QPI, Turbo, HT, 10C/20T (105W) Max Mem 2133Mhz (338-BFFG).
Upgrade to 2 Intel Xeon E5-2660 V3 2.6Ghz… (374-BBGN)
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
My question is, how can we tell which of the 2 processors provides better performance? Are there other factors to include as well such as system board?
A brief summary is as follows:
2016 - 10 cores per socket
2017 = 14 cores per socket
2016 - Version 3 with 2.6 Ghz
2017 - Version 4 with 2.0 Ghz
In the past we have been burned by purchasing new ESXi hosts that performed much slower with our applications than older models. For example a Dell T710 with 6 cores per socket and 2.8 Ghz out performs an HP ProLiant Gen9 with 8 cores per socket and 2.6 Ghz. I do not wish to purchase a server that will provide lower performance than the Dell T710 (6 yers old server).
How can we judge how the processors or how the server will perform compared to another server?
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
If the 2 CPU's that I provided above are not that different, then they both may not do. I just left a meeting with Dell and they told me that some SQL Server Database applications prefer a higher CPU speed and not so much cores per processor. Can you tell me the difference in performance from:
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
A Proof of concept is a good idea. The easiest ting to do is perhaps do as how the Vendor suggests.
Can anyone tell me the difference in performance from:
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
Can anyone tell me the difference in performance from:
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
ASKER
I could not figure out how to use the above mentioned web sites, so I have been waiting for a response to my last question:
Can anyone tell me the difference in performance from:
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
Can anyone tell me the difference in performance from:
2017: Intel Xeon E5-2660 v4 2.0GHz,35M Cache,9.60GT/s QPI,Turbo,HT,14C/28T (105W) Max Mem 2400MHz
vs
Recently suggested: Intel Xeon E5-2667 3.2 Ghz with 25M Cach, 8 cores/processor ?
It seems i can't find benchmarks for E5-2667 (v4) as single but here are the benchmarks for dual
Dual CPU Intel Xeon E5-2667 v4
Dual Intel Xeon E5-2660 v4
The performance difference is very small while the price gap is considerable
Dual CPU Intel Xeon E5-2667 v4
Dual Intel Xeon E5-2660 v4
The performance difference is very small while the price gap is considerable
ASKER
On the Dell opening up the surgery pictures for a test patient will take 3 seconds; but on the HP Servers if one is on a client program accessing the same test patient's surgery pictures it take 9-10 seconds. The performance has been tested and documented and it is consistent. It performs better from the Dell Servers and not when on the HP Servers.
In face I moved a Virtual Machine that was on the HP to use resources on the Dell then the performance went up immidiately. Even other normal operations such as logging in was much faster (13 seconds vs 21 seconds). It was determined that since the CPU is used a lot when these modules are triggered that that must be the bottle neck. We even moved teh test patient's pictures from the Dbase server to another server on DELL (for shared file storage to test IOPS) and no change in performance.
It performed better with the Dell so CPU it is.