SPECpower_ssj2008 Copyright (C) 2007-2024 Standard Performance Evaluation Corporation Dell Inc. PowerEdge R7725 SPECpower_ssj2008 = 35,920 overall ssj_ops/watt Test Sponsor: Dell Inc. SPEC License #: 6573 Tested By: Dell Inc. Test Location: Round Rock, TX, USA Test Date: Sep 25, 2024 Publication: Oct 28, 2024 Hardware Availability: Nov-2024 Software Availability: Sep-2024 Test Method: Single Node System Source: Single Supplier System Designation: Server Power Provisioning: Line-powered Benchmark Results Summary Performance | Power | Performance | | to Power | | Ratio Target Load | Actual Load | ssj_ops | Average | | | | Active Power | | | | (W) | ------------- | ------------- | ------------- | ------------- | ------------- 100% | 99.3% | 40,017,802 | 861 | 46,456 90% | 90.0% | 36,252,933 | 812 | 44,647 80% | 80.0% | 32,246,773 | 771 | 41,847 70% | 70.0% | 28,199,831 | 710 | 39,722 60% | 60.0% | 24,183,842 | 661 | 36,608 50% | 50.0% | 20,153,061 | 593 | 34,009 40% | 40.0% | 16,123,742 | 515 | 31,326 30% | 30.0% | 12,092,032 | 438 | 27,584 20% | 20.0% | 8,050,595 | 367 | 21,944 10% | 10.0% | 4,024,453 | 297 | 13,531 Active Idle | 0 | 138 | 0 sum of ssj_ops / sum of power = | 35,920 System Under Test Set: 'sut' Set Identifier: sut Set Description: System Under Test # of Identical Nodes: 1 Comment: SUT HARDWARE Hardware Vendor: Dell Inc. Model: PowerEdge R7725 Form Factor: 2U CPU Name: AMD EPYC 9965 2.25 GHz CPU Characteristics: 192 core, 2.25 GHz, 384 MB L3 Cache, Max Boost Clock up to 3.70 GHz CPU Frequency (MHz): 2250 CPU(s) Enabled: 384 cores, 2 chips, 192 cores/chip Hardware Threads: 768 (2 / core) CPU(s) Orderable: 2 chips Primary Cache: 32 KB I + 48 KB D on chip per core Secondary Cache: 1 MB I+D on chip per core Tertiary Cache: 384 MB I+D on chip per chip, 32 MB shared / 16 cores Other Cache: None Memory Amount (GB): 768 # and size of DIMM: 24 x 32 GB Memory Details: 24x32GB 2Rx4 PC5-6400-BR running at 5200 MT/s Slots A1-12 B1-12 populated Power Supply Quantity and Rating 2 x 1500 (W): Power Supply Details: 1500W Titanium Efficiency (DPN 1PKMH) Disk Drive: 1 x 480GB NVMe Disk Controller: BOSS N DCMHS card # and type of Network Interface 1 x 4P Broadcom 25G Cards (NICs) Installed: NICs Enabled in Firmware / OS / 4/1/1 Connected: Network Speed (Mbit): 25000 Keyboard: None Mouse: None Monitor: None Optical Drives: No Other Hardware: None SOFTWARE Power Management: Enabled(see SUT notes) Operating System (OS): Microsoft Windows Server 2022 Datacenter OS Version: 10.0.20348 Build 20348 Filesystem: NTFS JVM Vendor: Oracle JVM Version: Java HotSpot(TM) 64-Bit Server VM 17.1(build 11.0.28, mixed mode) JVM Command-line Options: -server -Xms1500m -Xmx1500m -Xmn1300m -XX:+UseParallelGC -XX:ParallelGCThreads=2 -XX:+UseLargePages -XX:LargePageSizeInBytes=2m -XX:+UseBiasedLocking -XX:InlineSmallCode=1500 -XX:UseAVX=1 -XX:AutoBoxCacheMax=20000 -XX:+OptimizeFill -XX:+AggressiveHeap -XX:AllocatePrefetchDistance=256 -XX:AllocatePrefetchLines=16 JVM Affinity: start /NODE[0-23] /AFFINITY [0xF to 0xF0000000] JVM Instances: 192 JVM Initial Heap (MB): 1500 JVM Maximum Heap (MB): 1500 JVM Address Bits: 64 Boot Firmware Version: 0.2.6 Management Firmware Version: 10.1.15.00 A00 Workload Version: SSJ 1.2.10 Director Location: Controller Other Software: None Boot Firmware Settings * Logical Processor enabled * L1 Stream HW Prefetcher disabled * L2 Stream HW Prefetcher disabled * MADT Core Enumeration set to Linear * L3 Cache as NUMA Domain set to Enabled * NUMA nodes per socket set to 4 * System Profile set to Custom * CPU Power Management set Max * Memory Frequency set to 5200 MT/s * Memory Patrol Scrub disabled * Memory Refresh Rate set to 1x * Determinism Slider set to Performance Determinism * Efficiency Optimized Mode Enabled * ApbDis enabled * ApbDis Fixed Socket P-state set to P2 * xGMI x20 Management Firmware Settings Minimum Power (Performance per Watt Optimized) System Under Test Notes * Each JVM instance was bound to 4 logical processors * Using the local security settings console, "lock pages in memory" was enabled for the user running the benchmark * Run initiated via Windows Remote Desktop Connection * Hard Disk set to turn off after 1 minute * Display set to turn off after 1 minute * SPECpower_ssj.props input.load_level.number_warehouses set to 768 since processor has 192 cores * Power Plan Minimum processor state set to 0% * Power Plan Maximum processor state set to 100% * Using Windows PowerShell, set the following powercfg properties: * Powercfg /setacvalueindex scheme_current sub_processor PROCTHROTTLEMIN 0 * Powercfg /setacvalueindex scheme_current sub_processor PROCTHROTTLEMAX 100 * Powercfg /setacvalueindex scheme_current sub_processor PERFEPP 0 * Powercfg /setacvalueindex scheme_current sub_processor perfautonomous 1 * Powercfg /setactive scheme_current * The system ROM used for this result contains AMD AGESA version 2110100B00000000 for the AMD EPYC 9965 processor. Controller System HARDWARE Hardware Vendor: Dell Inc Model: PowerEdge T150 CPU Description: Intel Xeon E5620, 2.40 GHz Memory amount (GB): 32.0 SOFTWARE Operating System (OS): Microsoft Windows Server 2022 Standard 10.0.20348 N/A build 20348 JVM Vendor: Oracle Corporation JVM Version: Hotspot(TM) Client VM, version 1.8.0_121 CCS Version: 1.2.6 Measurement Devices POWER ANALYZER PWR1 Hardware Vendor: Yokogawa Meters and Instruments Corporation Model: WT210 Serial Number: 91KA21978 Connectivity: RS-232 Input Connection: Default Metrology Institute: NIST Accredited by: TesCom Calibration Label: T76421 Date of Calibration: 13-Oct-2023 PTDaemon Host System: same as CCS PTDaemon Host OS: same as CCS PTDaemon Version: 1.11.0-a4047d62-20231205 Setup Description: Connected to power supply TEMPERATURE SENSOR TEMP1 Hardware Vendor: Digi International Inc. Model: Watchport/H Driver Version: 5.10.26.0 Connectivity: USB PTDaemon Host System: same as CCS PTDaemon Host OS: same as CCS Setup Description: 25 mm in front of SUT main airflow intake Notes None Aggregate Electrical and Environmental Data Target Load | Average | Minimum | Active | Ambient | Power (W) | Temperature | | (Degrees C) ----------- | ---------- | ------------- 100% | 861 | 22.3 90% | 812 | 22.4 80% | 771 | 22.4 70% | 710 | 22.5 60% | 661 | 22.5 50% | 593 | 22.6 40% | 515 | 22.6 30% | 438 | 22.7 20% | 367 | 22.8 10% | 297 | 22.8 Active Idle | 138 | 22.8 Line Standard | Minimum | Elevation (m) | Temperature | | (Degrees C) | ------------------------- | --------------- | --------------- 208V / 60 Hz / 1 phase / | 22.3 | 255 2 wires | | See the Power/Temperature Details Report (power_ssj2008-20241007-01464-power.txt) for additional details. Aggregate Performance Data Target Load | Actual Load | ssj_ops | | Target | Actual ----------------- | ----------------- | ----------------- | ----------------- Calibration 1 | | | 40,798,914 Calibration 2 | | | 40,326,382 Calibration 3 | | | 40,260,960 ssj_ops@calibrated=40,293,671 100% | 99.3% | 40,293,671 | 40,017,802 90% | 90.0% | 36,264,303 | 36,252,933 80% | 80.0% | 32,234,936 | 32,246,773 70% | 70.0% | 28,205,569 | 28,199,831 60% | 60.0% | 24,176,202 | 24,183,842 50% | 50.0% | 20,146,835 | 20,153,061 40% | 40.0% | 16,117,468 | 16,123,742 30% | 30.0% | 12,088,101 | 12,092,032 20% | 20.0% | 8,058,734 | 8,050,595 10% | 10.0% | 4,029,367 | 4,024,453 Active Idle | | 0 | 0 See the Host Performance Report (power_ssj2008-20241007-01464.WIN-CKU6IPL9I0A.txt) for additional details. Copyright (C) 2007-2024 Standard Performance Evaluation Corporation http://www.spec.org - info@spec.org SPECpower_ssj2008 Reporter Version: [SSJ 1.2.10, May 9, 2012]