Looking at the VMmarks :-
HP ProLiant DL385 G6 VMware ESX v4.0 | VMmark v1.1 15.54@11 tiles |
HP ProLiant DL380 G6 VMware ESX v4.0 | VMmark v1.1 24.15@17tiles |
How do you decide which is is more suitable to running Vsphere? I'm slightly confused on understanding the figures. From the doc's it looks like the first figure is a performance figure and the second the number of tiles the server was able to run without undue performance hits. In which case, is it that the 385 runs faster VM's but can't run as many VM's as the 380? Is it that simple?
It is a fair comparision between two socket servers from Intel and AMD
Intel | AMD | ||
Socket | 2 | 2 | |
Cres per CPU | 4 | 4 | |
Hypertreading | yes | no | |
Logical CPU | 8 | 4 | |
´VMware License | 2 | 2 | |
Vmmark Value | 17 tiles | 11 tiles | |
1 Tile | 6 different workloads | 6 different worloads | |
Tiles x Workload = standard VM's | 6 x 17 = 102 VM'S | 6 x 11 = 66 VM's | |
Result | 36 more VM'S than AMD | ||
Read although about
Anatech report
Intel VMdq support
I would recommend not looking too deeply into the benchmark results for
CPU. CPU most likely won't be your bottleneck. Usually memory and
disk IO will be performance factors before CPU. To maximize CPU, you
must hit very high amounts of memory which ruins your ROI (compared to
less memory and more servers). Instead, consider sticking with what
you already have within your datacenter or compare price/performance
between models.
not fair. DL385G6 have 64GB DDR2 667MHz Registered ECC but DL380G6 have 96GB DDR3 1066MHz Registered ECC.