OS Images |
os_Image_1(1)
|
Hardware Description |
hw_1
|
Number of Systems |
1
|
SW Environment |
Non-virtual
|
Tuning |
- Workload Profile=High Performance Compute(HPC)
- Thermal Configuration=Maximum Cooling
- Performance Determinism=Power Deterministic
- Memory Patrol Scrubbing=Disabled
- NUMA memory domains per socket=Four memory domains per socket
- Last-Level Cache(LLC) As NUMA Node= Enabled
- Maximum memory Bus Frequency=2933
|
Notes |
None
|
|
JVM Instances |
jvm_Ctr_1(1), jvm_Backend_1(8), jvm_TxInjector_1(8)
|
OS Image Description |
os_1
|
Tuning |
- tuned-adm profile=throughput-performance
- ulimit -n 655350
- echo 15000000 > /proc/sys/kernel/sched_wakeup_granularity_ns
- echo 80000000 > /proc/sys/kernel/sched_min_granularity_ns
- echo 1000 > /proc/sys/kernel/sched_migration_cost_ns
- echo 16 > /proc/sys/kernel/sched_nr_migrate
- echo 950000 > /proc/sys/kernel/sched_rt_runtime_us
- echo 10000000 > /proc/sys/kernel/sched_latency_ns
- echo 10000 > /proc/sys/vm/dirty_expire_centisecs
- echo 1500 > /proc/sys/vm/dirty_writeback_centisecs
- echo 40 > /proc/sys/vm/dirty_ratio
- echo 10 > /proc/sys/vm/dirty_background_ratio
- echo 10 > /proc/sys/vm/swappiness
- echo 0 > /proc/sys/kernel/numa_balancing
- echo always > /sys/kernel/mm/transparent_hugepage/defrag
- echo always > /sys/kernel/mm/transparent_hugepage/enabled
|
Notes |
None
|
Parts of Benchmark |
Controller
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms3g -Xmx3g -Xmn2g -XX:+UseParallelOldGC -XX:ParallelGCThreads=1 -XX:CICompilerCount=2
|
Tuning |
None
|
Notes |
numactl --interleave=all
|
Parts of Benchmark |
Backend
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms30g -Xmx30g -Xmn27g -XX:AllocatePrefetchInstr=2 -XX:+UseParallelOldGC -XX:LargePageSizeInBytes=2m -XX:-UseAdaptiveSizePolicy -XX:+AlwaysPreTouch -XX:-UseBiasedLocking -XX:+UseLargePages -XX:SurvivorRatio=30 -XX:TargetSurvivorRatio=99 -XX:MaxTenuringThreshold=15 -XX:ParallelGCThreads=32 -XX:+UseTransparentHugePages -XX:MaxInlineSize=900 -XX:FreqInlineSize=900 -XX:LoopUnrollLimit=30 -XX:InlineSmallCode=3k -XX:AutoBoxCacheMax=5000
|
Tuning |
None
|
Notes |
Used numactl to affinitize each Backend JVM to two NUMA nodes. - numactl --cpunodebind=0,1 --localalloc
- numactl --cpunodebind=2,3 --localalloc
- numactl --cpunodebind=4,5 --localalloc
- numactl --cpunodebind=6,7 --localalloc
- numactl --cpunodebind=8,9 --localalloc
- numactl --cpunodebind=10,11 --localalloc
- numactl --cpunodebind=12,13 --localalloc
- numactl --cpunodebind=14,15 --localalloc
|
Parts of Benchmark |
TxInjector
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms4g -Xmx4g -Xmn3g -XX:+UseParallelOldGC -XX:ParallelGCThreads=3 -XX:CICompilerCount=2
|
Tuning |
None
|
Notes |
Used numactl to affinitize each Transaction Injector JVM to a NUMA node. - numactl --cpunodebind=1 --localalloc
- numactl --cpunodebind=3 --localalloc
- numactl --cpunodebind=5 --localalloc
- numactl --cpunodebind=7 --localalloc
- numactl --cpunodebind=9 --localalloc
- numactl --cpunodebind=11 --localalloc
- numactl --cpunodebind=13 --localalloc
- numactl --cpunodebind=15 --localalloc
|
|