Amazon EC2 m8a.metal-48xl 536478 SPECjbb2015-Composite max-jOPS
442189 SPECjbb2015-Composite critical-jOPS
Tested by: Infobell IT Solutions Pvt. Ltd. Test Sponsor: Infobell IT Solutions Pvt. Ltd. Test location: Bengaluru,IND Test date: May 22, 2026
SPEC license #: 6723 Hardware Availability: Oct-2025 Software Availability: Feb-2026 Publication: Thu Jun 25 13:02:27 EDT 2026
Benchmark Results Summary
 
Overall Throughput RT curve
Overall SUT (System Under Test) Description
VendorAmazon Web Services (AWS)
Vendor URLhttps://aws.amazon.com/ec2/instance-types/m8a/
System SourceSingle Supplier
System DesignationServer Rack
Total Systems1
All SUT Systems IdenticalYES
Total Nodes1
All Nodes IdenticalYES
Nodes Per System1
Total Chips2
Total Cores192
Total Threads192
Total Memory Amount (GB)768
Total OS Images1
SW EnvironmentNon-virtual
 
Hardware hw_1
NameAmazon EC2 m8a.metal-48xl
VendorAmazon Web Services (AWS)
Vendor URLhttps://aws.amazon.com/ec2/instance-types/m8a/
AvailableOct-2025
Modelm8a.metal-48xl
Form FactorNone
CPU NameAMD EPYC 9R45 96-Core Processor
CPU Characteristics96 core, 2.6 GHz, 384MB L3 Cache (Max frequency 4.5 GHz)
Number of Systems1
Nodes Per System1
Chips Per System2
Cores Per System192
Cores Per Chip96
Threads Per System192
Threads Per Core1
Version1.0
CPU Frequency (MHz)2600
Primary Cache32KB(I) + 48KB(D) per core
Secondary Cache1MB (I+D) per core
Tertiary Cache384MB (I+D) on chip per chip
Other CacheNone
Disk1x 500GB EBS
File Systemext4
Memory Amount (GB)768
# and size of DIMM(s)24 x 32GB
Memory DetailsDDR5 6400 MT/s
# and type of Network Interface Cards (NICs)75 Gbps
Power Supply Quantity and Rating (W)None
Other HardwareNone
Cabinet/Housing/EnclosureNone
Shared DescriptionNone
Shared CommentNone
Notes
  • NA: The test sponsor attests, as of date of publication, that CVE-2017-5754 (Meltdown) is mitigated in the system as tested and documented.
  • Yes: The test sponsor attests, as of date of publication, that CVE-2017-5753 (Spectre variant 1) is mitigated in the system as tested and documented.
  • Yes: The test sponsor attests, as of date of publication, that CVE-2017-5715 (Spectre variant 2) is mitigated in the system as tested and documented.
  • The submitter confirms that the benchmark configuration and execution comply with the applicable Terms of Service and usage policies of the hardware, software, and service providers on which the system under test was deployed
Other Hardware network_1
NameNone
VendorNone
Vendor URLNone
VersionNone
AvailableNone
BitnessNone
NotesNone
Operating System os_1
NameUbuntu 24.04.4 LTS
VendorUbuntu
Vendor URLhttps://ubuntu.com/
Version6.17.0-1012-aws
AvailableFeb-2026
Bitness64
NotesNone
Java Virtual Machine jvm_1
NameOracle Java SE 25.0.2
VendorOracle
Vendor URLhttp://oracle.com/
VersionJava HotSpot 64-bit Server VM, version 25.0.2
AvailableJan-2026
Bitness64
NotesNone
Other Software other_1
NameNone
VendorNone
Vendor URLNone
VersionNone
AvailableNone
BitnessNone
NotesNone
Hardware
OS Images os_Image_1(1)
Hardware Description hw_1
Number of Systems 1
SW Environment non-virtual
Tuning None
Notes None
OS Image os_Image_1
JVM Instances jvm_Composite_1(1)
OS Image Description os_1
Tuning

  • tuned-adm profile throughput-performance
  • ulimit -n 1024000
  • echo 950000 > /proc/sys/kernel/sched_rt_runtime_us
  • echo 250000 > /proc/sys/kernel/sched_base_slice_ns
  • echo 200000 > /proc/sys/kernel/sched_migration_cost_ns
  • echo 32 > /proc/sys/kernel/sched_nr_migrate
  • echo NO_WAKEUP_PREEMPTION > /sys/kernel/debug/sched/features
  • 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 1 > /proc/sys/kernel/numa_balancing
  • echo always > /sys/kernel/mm/transparent_hugepage/enabled
  • echo always > /sys/kernel/mm/transparent_hugepage/defrag

Notes None
JVM Instance jvm_Composite_1
Parts of Benchmark Composite
JVM Instance Description jvm_1
Command Line

-Xms720g -Xmx720g -Xmn700g -server -XX:+UseParallelGC -XX:+AlwaysPreTouch -XX:+UseLargePages -XX:LargePageSizeInBytes=2m -XX:MaxTenuringThreshold=15 -XX:-UseAdaptiveSizePolicy -XX:ParallelGCThreads=96 -XX:SurvivorRatio=80 -XX:TargetSurvivorRatio=95 -XX:+UseNUMA -XX:+UseTransparentHugePages -XX:InlineSmallCode=11k -XX:TLABAllocationWeight=2 -XX:ThreadStackSize=200 -XX:AutoBoxCacheMax=32 -XX:OnStackReplacePercentage=35

Tuning

Used numactl to affinitize JVM across NUMA nodes

  • numactl --cpunodebind=0-1 --interleave=0-1
Notes None
max-jOPS = jOPS passed before the First Failure
Pass/Fail Pass Pass Pass Fail Fail
jOPS 522360 529419 536478 543537 550596
critical-jOPS = Geomean ( jOPS @ 10000; 25000; 50000; 75000; 100000; SLAs )
Response time percentile is 99-th
SLA (us) 10000 25000 50000 75000 100000 Geomean
jOPS 285180 442359 504712 511771 518830 442189
  Percentile
  10-th 50-th 90-th 95-th 99-th 100-th
500us 289416 / 35295 - / 7059 - / 7059 - / 7059 - / 7059 - / 7059
1000us 437653 / 444712 190591 / 35295 14118 / 21177 14118 / 21177 7059 / 14118 - / 7059
5000us 536478 / - 494124 / 501183 409417 / 416476 374123 / 367064 211768 / 218827 42354 / 14118
10000us 536478 / - 529419 / 536478 480007 / 487065 437653 / 444712 296475 / 268239 49412 / 14118
25000us 536478 / - 536478 / - 515301 / 522360 501183 / 508242 451771 / 437653 105884 / 28236
50000us 536478 / - 536478 / - 529419 / 536478 522360 / 529419 501183 / 508242 105884 / 28236
75000us 536478 / - 536478 / - 529419 / 536478 522360 / 529419 508242 / 515301 240003 / 28236
100000us 536478 / - 536478 / - 529419 / 536478 529419 / 536478 515301 / 522360 275298 / 134119
200000us 536478 / - 536478 / - 536478 / - 529419 / 536478 529419 / 536478 331769 / 303534
500000us 536478 / - 536478 / - 536478 / - 536478 / - 536478 / - 465889 / 430594
1000000us 536478 / - 536478 / - 536478 / - 536478 / - 536478 / - 522360 / 529419
Probes jOPS / Total jOPS
Request Mix Accuracy
Note
(Actual % in the Mix - Expected % in the Mix) must be within:
'Main Tx' limit of +/-5.0% for the requests whose expected % in the mix is >= 10.0%
'Minor Tx' limit of +/-1.0% for the requests whose expected % in the mix is < 10.0%
There were no non-critical failures in Response Time curve building
Delay between status pings
IR/PR Accuracy
This section lists properties only set by user
Property Name Default Controller
specjbb.comm.connect.selector.runner.count 0 12
specjbb.controller.maxir.maxFailedPoints 3 1
specjbb.controller.rtcurve.warmup.step 0.1 0.2
specjbb.customerDriver.threads 64 {probe=90, saturate=140}
specjbb.forkjoin.workers 192 {Tier1=48, Tier2=1, Tier3=14}
specjbb.group.count 1 4
specjbb.heartbeat.period 10000 50000
specjbb.heartbeat.threshold 100000 500000
specjbb.mapreducer.pool.size 192 4
View table in csv format
 
Level: COMPLIANCE
Check Agent Result
Check properties on compliance All PASSED
 
Level: CORRECTNESS
Check Agent Result
Compare SM and HQ Inventory All PASSED
High-bound (max attempted) is 705892 IR
High-bound (settled) is 588310 IR