One of our stack holders did some testing and it would appear that we can get a performance decrease with the upgrade. Do you have any input on Rogers findings ?
"Further to our conversation this morning and my subsequent email, I believe that I have been able to carry out a test that definitively shows the difference between the old CPU and new Core licensing model we have accepted from SAP. Based on my testing below, any hope that we had of v17 providing better utilization of the server processing power and therefore a possible performance increase is questionable, and it would appear that many clients will go backwards in server processor utilisation or face considerable costs of buying more core licenses.
We need a urgent conversation about this.
Test Conclusion
- The assumption that 1 CPU is equivalent to 1 Core is completely wrong
- SAP’s assertion that 4 cores is equivalent to 1 CPU may have some merit but it will still penalise many of our clients that have more than 8 cores
- Tests on other configurations show that for example a machine with 2 Core with 4 logical processors, a 2 Core license will access all 4 logical processors
- I have no metric to definitely know how 2,4,6 or 8 cores impact on performance
- Grand Prix Mazda continually complain of poor performance, they have 14 cores which are all accessible using Sybase 11, our version of SAP 17 is limited to 8 cores, what impact will a 40% reduction in processing power have on performance?
Test Conditions
Server: Windows 2012 R2
Type: Amazon Hosted Environment
CPU: 1
Cores: 8
Databases: 20GB Production Sybase 11
20GB Production SAP v17
Software: Sybase 11, licensed as 1 CPU
SAP v17, licensed as 4 Cores"
Test Results
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.