In this post I will compare maximum workloads between lpars using POWER7 and POWER8 processors.

Detailed processor information for the two test lpars

Processor Type: PowerPC_POWER7
 Processor Implementation Mode: POWER 7
 Processor Version: PV_7_Compat
 Number Of Processors: 10
 Processor Clock Speed: 4228 MHz
 System configuration: type=Shared mode=Uncapped smt=4 lcpu=40 mem=92160MB psize=16 ent=1.00

Processor Type: PowerPC_POWER8
 Processor Implementation Mode: POWER 8
 Processor Version: PV_8_Compat
 Number Of Processors: 8
 Processor Clock Speed: 4157 MHz
 System configuration: type=Shared mode=Uncapped smt=8 lcpu=64 mem=102400MB psize=16 ent=0.80

My expectation for processor performance is based on specint_cpu_explorer from Karl Arao

I picked the closest processor from the list based on clock speed and type

ResultPerCore: 54.63
 Operating System: IBM AIX V7.1
 Processor MHz: 4228
 Processor: POWER7+
 System: IBM Power 730 Express (4.2 GHz, 16 core)

ResultPerCore: 79.00
 Operating System: IBM AIX V7.1
 Processor MHz: 4192
 Processor: POWER8
 System: IBM Power E870 (4.19 GHz, 80 core)

Expected workload increase between POWER7 and POWER8 processors based on specint_rates


Test results  generated by my hog load script for both test lpars

lpar using POWER7 with 10 cores: 1821 tx/s, 182 tx/s/core
lpar using POWER8 with  8 cores: 2521 tx/s, 315 tx/s/core

Workload ratios between POWER7 and POWER8 when lpar-s are maxed out

lpar 2521/1821 = 1.38
core 315/182 = 1.73


Lpar test
Lpar test POWER8