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
- Determinism Control=Manual
- Performance Determinism=Power Deterministic
- Memory Patrol Scrubbing=Disabled
- Numa Memory Domains Per Socket(NPS)=Four Memory Domains Per Socket
- Last-Level Cache(LLC) As NUMA Node=Enabled
- L1 Stream HW Prefetcher=Disabled
- L2 Stream HW Prefetcher=Disabled
- Minimum Processor Idle Power Core C-State=No C-states
- Package Power Limit=400
|
Notes |
None
|
|
JVM Instances |
jvm_Ctr_1(1), jvm_Backend_1(12), jvm_TxInjector_1(12)
|
OS Image Description |
os_1
|
Tuning |
- tuned-adm profile throughput-performance
- ulimit -n 1024000
- UserTasksMax=970000
- DefaultTasksMax=970000
- write_schedstat rt_runtime_us 960000
- write_schedstat latency_ns 40000000
- write_schedstat migration_cost_ns 40000
- write_schedstat min_granularity_ns 900000000
- write_schedstat wakeup_granularity_ns 200000000
- write_schedstat nr_migrate 9000
- 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 0 > /proc/sys/vm/numa_stat
- echo always > /sys/kernel/mm/transparent_hugepage/enabled
- echo always > /sys/kernel/mm/transparent_hugepage/defrag
|
Notes |
None
|
Parts of Benchmark |
Controller
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms3g -Xmx3g -Xmn2g -XX:+UseParallelGC -XX:ParallelGCThreads=2 -XX:CICompilerCount=2
|
Tuning |
Used numactl to interleave memory on all CPUs
|
Notes |
None
|
Parts of Benchmark |
Backend
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms30g -Xmx30g -Xmn29g -XX:AllocatePrefetchInstr=2 -XX:+UseParallelGC -XX:ParallelGCThreads=16 -XX:LargePageSizeInBytes=2m -XX:-UseAdaptiveSizePolicy -XX:+AlwaysPreTouch -XX:+UseLargePages -XX:SurvivorRatio=20 -XX:TargetSurvivorRatio=95 -XX:MaxTenuringThreshold=15 -XX:InlineSmallCode=11k -XX:MaxGCPauseMillis=100 -XX:LoopUnrollLimit=200 -XX:+UseTransparentHugePages -XX:ThreadStackSize=140 -XX:CompileThresholdScaling=120 -XX:CICompilerCount=4
|
Tuning |
Used numactl to affinitize each Backend JVM to physcical cores in a NUMA node. - numactl --physcpubind=0-7,96-103 --localalloc
- numactl --physcpubind=8-15,104-111 --localalloc
- numactl --physcpubind=16-23,112-119 --localalloc
- numactl --physcpubind=24-31,120-127 --localalloc
- numactl --physcpubind=32-39,128-135 --localalloc
- numactl --physcpubind=40-47,136-143 --localalloc
- numactl --physcpubind=48-55,144-151 --localalloc
- numactl --physcpubind=56-63,152-159 --localalloc
- numactl --physcpubind=64-71,160-167 --localalloc
- numactl --physcpubind=72-79,168-175 --localalloc
- numactl --physcpubind=80-87,176-183 --localalloc
- numactl --physcpubind=88-95,184-191 --localalloc
|
Notes |
None
|
Parts of Benchmark |
TxInjector
|
JVM Instance Description |
jvm_1
|
Command Line |
-Xms3g -Xmx3g -Xmn2g -XX:+UseParallelGC -XX:ParallelGCThreads=2 -XX:CICompilerCount=2
|
Tuning |
Used numactl to affinitize each Transaction Injector JVM to physical cores in a NUMA node. - numactl --physcpubind=0-7,96-103 --localalloc
- numactl --physcpubind=8-15,104-111 --localalloc
- numactl --physcpubind=16-23,112-119 --localalloc
- numactl --physcpubind=24-31,120-127 --localalloc
- numactl --physcpubind=32-39,128-135 --localalloc
- numactl --physcpubind=40-47,136-143 --localalloc
- numactl --physcpubind=48-55,144-151 --localalloc
- numactl --physcpubind=56-63,152-159 --localalloc
- numactl --physcpubind=64-71,160-167 --localalloc
- numactl --physcpubind=72-79,168-175 --localalloc
- numactl --physcpubind=80-87,176-183 --localalloc
- numactl --physcpubind=88-95,184-191 --localalloc
|
Notes |
None
|
|