Summary of Key Findings
This section summarizes key z/VM 6.3 performance items and contains links that take the reader to more detailed information about each one.
Further, the Performance Improvements article gives information about other performance enhancements in z/VM 6.3.
To compare the performance of z/VM 6.3 to previous releases, IBM ran a variety of workloads on the two systems. For the base case, IBM used z/VM 6.2 plus all Control Program (CP) PTFs available as of September 8, 2011. For the comparison case, IBM used z/VM 6.3 at the "code freeze" level of June 14, 2013.
Regression measurements comparing these two z/VM levels showed nearly identical results for most workloads. Variation was generally less than 5%.
Key Performance Improvements
z/VM 6.3 contains the following enhancements that offer performance improvements compared to previous z/VM releases:
Storage Management Scaling Improvements: z/VM 6.3 can exploit a partition having up to 1024 GB (1 TB) of real storage. This is an improvement of a factor of four over the previous limit. The storage management chapter discusses the performance characteristics of the new storage management code. Workloads that on earlier releases were suffering from the constraints present in the z/VM Control Program's real storage manager should now experience correct performance in real storage sizes up to the new limit. Workloads that were not experiencing problems on earlier releases will not experience improvements. See the chapter for more information.
z/VM HiperDispatch: The z/VM 6.3 HiperDispatch function improves performance for amenable workloads. At the partition level, z/VM exploits PR/SM's vertical mode partitions to help increase the logical CPUs' proximity to one another and to help reduce the motion of the partition's logical CPUs within the physical hardware. At the guest level, changes in the z/VM dispatcher help to increase the likelihood that a guest virtual machine will experience reduced motion among the logical CPUs of the partition. These changes provided up to 49% performance improvement in the workloads measured. See the chapter for more information.
System Dump Improvements: Increasing the system's maximum real storage size required a corresponding improvement in the system's ability to collect system dumps. The system dump chapter provides a brief discussion of the changes made in the area of dumps, especially as those changes contribute to better performance. In the experiments conducted, data rates for dumps improved by 50% to 90% for dumps to ECKD and by 190% to 1500% for dumps to EDEV. See the chapter for more information.