Document History
Version 1.0 (2017/10/31): Introduced technical data, SPECcpu2006, OLTP-2, vServCon, STREAM, and LINPACK measurements with Intel Xeon Processor Scalable Family.
Version 1.1 (2018/04/10): Updated SPECcpu2006 and vServCon with additional measurements.
Technical Data
The FUJITSU PRIMERGY RX2530 M4 is a rack server. Key specifications include:
Model | PRIMERGY RX2530 M4 |
Model versions | PY RX2530 M4 4x 3.5' expandable, PY RX2530 M4 2x 2.5' expandable, PY RX2530 M4 10x 2.5' |
Form factor | Rack server |
Chipset | Intel® C620 |
Number of sockets | 2 |
Number of processors orderable | 1 or 2 |
Processor type | Intel® Xeon® Processor Scalable Family |
Number of memory slots | 24 (12 per processor) |
Maximum memory configuration | 3,072 GB |
Onboard HDD controller | Controller with RAID 0, RAID 1 or RAID 10 for up to 8 SATA HDDs |
PCI slots | PCI-Express 3.0 x8 × 1, PCI-Express 3.0 x16 × 3 |
Max. number of internal hard disks | 4 (4x 3.5'), 8 (8x 2.5' expandable), 10 (10 x 2.5') |
Note: Decimal prefixes (e.g., 1 GB = 10⁹ bytes) are used according to SI standard, except for cache and memory modules where binary prefixes (e.g., 1 GB = 2³⁰ bytes) are implied.
Processors
The document lists various Intel® Xeon® Processor Scalable Family processors supported by the PRIMERGY RX2530 M4. Key metrics include Cores, Threads, Cache, UPI Speed, Rated Frequency, Max. Turbo Frequency, Max. Memory Frequency, and TDP. Intel® Turbo Boost Technology 2.0 allows higher frequencies, with the maximum achieved depending on active cores, power consumption, and temperature. Intel does not guarantee maximum turbo frequency due to manufacturing tolerances.
Processor | Cores | Threads | Cache [MB] | UPI Speed [GT/s] | Rated Frequency [Ghz] | Max. Turbo Frequency [Ghz] | Max. Memory Frequency [MHz] | TDP [Watt] |
Xeon Bronze 3104 | 6 | 6 | 8.3 | 9.6 | 1.7 | n/a | 2133 | 85 |
Xeon Bronze 3106 | 8 | 8 | 11.0 | 9.6 | 1.7 | n/a | 2133 | 85 |
Xeon Silver 4108 | 8 | 16 | 11.0 | 9.6 | 1.8 | 3.0 | 2400 | 85 |
Xeon Silver 4110 | 8 | 16 | 11.0 | 9.6 | 2.1 | 3.0 | 2400 | 85 |
Xeon Silver 4114 | 10 | 20 | 13.8 | 9.6 | 2.2 | 3.0 | 2400 | 85 |
Xeon Silver 4116 | 12 | 24 | 16.5 | 9.6 | 2.1 | 3.0 | 2400 | 85 |
Xeon Gold 5115 | 10 | 20 | 13.8 | 10.4 | 2.4 | 3.2 | 2400 | 85 |
Xeon Gold 5118 | 12 | 24 | 16.5 | 10.4 | 2.3 | 3.2 | 2400 | 105 |
Xeon Gold 5120 | 14 | 28 | 19.3 | 10.4 | 2.2 | 3.2 | 2400 | 105 |
Xeon Gold 6130 | 16 | 32 | 22.0 | 10.4 | 2.1 | 3.7 | 2666 | 125 |
Xeon Gold 6140 | 18 | 36 | 24.8 | 10.4 | 2.3 | 3.7 | 2666 | 140 |
Xeon Gold 6138 | 20 | 40 | 27.5 | 10.4 | 2.0 | 3.7 | 2666 | 125 |
Xeon Gold 6148 | 20 | 40 | 27.5 | 10.4 | 2.4 | 3.7 | 2666 | 150 |
Xeon Gold 6152 | 22 | 44 | 30.3 | 10.4 | 2.1 | 3.7 | 2666 | 140 |
Xeon Platinum 8153 | 16 | 32 | 22.0 | 10.4 | 2.0 | 2.8 | 2666 | 125 |
Xeon Platinum 8160 | 24 | 48 | 33.0 | 10.4 | 2.1 | 3.7 | 2666 | 150 |
Xeon Platinum 8164 | 26 | 52 | 35.8 | 10.4 | 2.0 | 3.7 | 2666 | 150 |
Xeon Platinum 8170 | 26 | 52 | 35.8 | 10.4 | 2.1 | 3.7 | 2666 | 165 |
Xeon Platinum 8176 | 28 | 56 | 38.5 | 10.4 | 2.1 | 3.8 | 2666 | 165 |
Xeon Platinum 8180 | 28 | 56 | 38.5 | 10.4 | 2.5 | 3.8 | 2666 | 205 |
Xeon Silver 4112 | 4 | 8 | 8.3 | 9.6 | 2.6 | 3.0 | 2400 | 85 |
Xeon Gold 5122 | 4 | 8 | 16.5 | 10.4 | 3.6 | 3.7 | 2666 | 105 |
Xeon Gold 6128 | 6 | 12 | 19.3 | 10.4 | 3.4 | 3.7 | 2666 | 115 |
Xeon Gold 6134 | 8 | 16 | 24.8 | 10.4 | 3.2 | 3.7 | 2666 | 130 |
Xeon Gold 6144 | 8 | 16 | 24.8 | 10.4 | 3.5 | 3.7 | 2666 | 150 |
Xeon Gold 6126 | 12 | 24 | 19.3 | 10.4 | 2.6 | 3.7 | 2666 | 125 |
Xeon Gold 6136 | 12 | 24 | 24.8 | 10.4 | 3.0 | 3.7 | 2666 | 150 |
Xeon Gold 6146 | 12 | 24 | 24.8 | 10.4 | 3.2 | 3.7 | 2666 | 165 |
Xeon Gold 6132 | 14 | 28 | 19.3 | 10.4 | 2.6 | 3.7 | 2666 | 140 |
Xeon Gold 6142 | 16 | 32 | 22.0 | 10.4 | 2.6 | 3.7 | 2666 | 150 |
Xeon Gold 6150 | 18 | 36 | 24.8 | 10.4 | 2.7 | 3.7 | 2666 | 165 |
Xeon Gold 6154 | 18 | 36 | 24.8 | 10.4 | 3.0 | 3.7 | 2666 | 200 |
Xeon Platinum 8168 | 24 | 48 | 33.0 | 10.4 | 2.7 | 3.7 | 2666 | 205 |
Xeon Silver 4114T | 10 | 20 | 13.8 | 9.6 | 2.2 | 3.0 | 2400 | 85 |
Xeon Gold 5119T | 14 | 28 | 19.3 | 10.4 | 1.9 | 3.2 | 2400 | 85 |
Xeon Gold 6134M | 8 | 16 | 24.8 | 10.4 | 3.2 | 3.7 | 2666 | 130 |
Xeon Gold 6140M | 18 | 36 | 24.8 | 10.4 | 2.3 | 3.7 | 2666 | 140 |
Xeon Gold 6142M | 16 | 32 | 22.0 | 10.4 | 2.6 | 3.7 | 2666 | 150 |
Xeon Platinum 8160M | 24 | 48 | 33.0 | 10.4 | 2.1 | 3.7 | 2666 | 150 |
Xeon Platinum 8170M | 26 | 52 | 35.8 | 10.4 | 2.1 | 3.7 | 2666 | 165 |
Xeon Platinum 8176M | 28 | 56 | 38.5 | 10.4 | 2.1 | 3.8 | 2666 | 165 |
Xeon Platinum 8180M | 28 | 56 | 38.5 | 10.4 | 2.5 | 3.8 | 2666 | 205 |
Memory Modules and Power Supplies
The PRIMERGY RX2530 M4 supports various DDR4 memory modules with different capacities, ranks, bit widths, and frequencies. It also offers modular power supplies with capacities ranging from 450W to 1200W.
Memory module | Capacity [GB] | Ranks | Bit width of the memory chips | Frequency [MHz] | Low voltage | Load reduced | Registered | ECC |
8 GB (1x8 GB) 2Rx8 DDR4-2666 R ECC | 8 | 2 | 8 | 2666 | ✓ | ✓ | ||
16 GB (1x16 GB) 2Rx8 DDR4-2666 R ECC | 16 | 2 | 8 | 2666 | ✓ | ✓ | ||
8 GB (1x8 GB) 1Rx4 DDR4-2666 R ECC | 8 | 1 | 4 | 2666 | ✓ | ✓ | ||
16 GB (1x16 GB) 1Rx4 DDR4-2666 R ECC | 16 | 1 | 4 | 2666 | ✓ | ✓ | ||
16 GB (1x16 GB) 2Rx4 DDR4-2666 R ECC | 16 | 2 | 4 | 2666 | ✓ | ✓ | ||
32 GB (1x32 GB) 2Rx4 DDR4-2666 R ECC | 32 | 2 | 4 | 2666 | ✓ | ✓ | ||
64 GB (1x64 GB) 4Rx4 DDR4-2666 3DS ECC | 64 | 4 | 4 | 2666 | ✓ | ✓ | ||
128 GB (1x128 GB) 8Rx4 DDR4-2666 3DS ECC | 128 | 8 | 4 | 2666 | ✓ | ✓ | ||
64 GB (1x64 GB) 4Rx4 DDR4-2666 LR ECC | 64 | 4 | 4 | 2666 | ✓ | ✓ |
Power supplies (since system release) | Max. number |
Modular PSU 450 W platinum hp | 2 |
Modular PSU 800 W platinum hp | 2 |
Modular PSU 800 W titanium hp | 2 |
Modular PSU 1200 W platinum hp | 2 |
SPECcpu2006 Benchmark
SPECcpu2006 measures system efficiency for integer and floating-point operations, focusing on CPU and memory. It includes SPECint2006 (12 applications) and SPECfp2006 (17 applications). Results are based on 'base' and 'peak' runs, with 'base' values used for publication. The benchmark measures speed (time per task) and rate (tasks per second, parallel execution).
Benchmark | Arithmetic | Type | Compiler optimization | Measurement result | Application |
SPECint2006 | integer | peak | aggressive | Speed | single-threaded |
SPECint_base2006 | integer | base | conservative | ||
SPECint_rate2006 | integer | peak | aggressive | Throughput | multi-threaded |
SPECint_rate_base2006 | integer | base | conservative | ||
SPECfp2006 | floating point | peak | aggressive | Speed | single-threaded |
SPECfp_base2006 | floating point | base | conservative | ||
SPECfp_rate2006 | floating point | peak | aggressive | Throughput | multi-threaded |
SPECfp_rate_base2006 | floating point | base | conservative |
Results are geometric averages of normalized ratio values, where '1' represents the reference system. A value of '2' means twice as fast as the reference.
SPECcpu2006 Benchmark Environment
The System Under Test (SUT) for SPECcpu2006 benchmarks was a PRIMERGY RX2530 M4 with dual Intel® Xeon® Processor Scalable Family CPUs and 16 GB memory. The BIOS settings were configured with specific options like Hyper-Threading disabled, Intel Virtualization Technology disabled, and Sub NUMA Clustering enabled. The operating system was SUSE Linux Enterprise Server 12 SP2 (x86_64), with various kernel and CPU settings optimized for performance. Compilers used included Intel C/C++ and Fortran compilers.
SPECcpu2006 Benchmark Results
The tables below show the SPECint_rate_base2006 and SPECfp_rate_base2006 results for various Intel Xeon processors in the PRIMERGY RX2530 M4. These results indicate performance differences based on processor core count, frequency, and other features.
Processor | Number of processors | SPECint_rate_base2006 | SPECint_rate2006 | SPECint_rate_base2006 Version 18.0.0.128 of Intel C++ Compiler |
Xeon Bronze 3104 | 2 | 330 | ||
Xeon Bronze 3106 | 2 | 440 | ||
Xeon Silver 4108 | 2 | 641 | ||
Xeon Silver 4110 | 2 | 711 | ||
Xeon Silver 4114 | 2 | 911 | ||
Xeon Silver 4116 | 2 | 1060 | ||
Xeon Gold 5115 | 2 | 980 | ||
Xeon Gold 5118 | 2 | 1160 | ||
Xeon Gold 5120 | 2 | 1310 | ||
Xeon Gold 6130 | 2 | 1560 | ||
Xeon Gold 6140 | 2 | 1760 | ||
Xeon Gold 6138 | 2 | 1770 | ||
Xeon Gold 6148 | 2 | 1940 | ||
Xeon Gold 6152 | 2 | 1990 | 2090 | |
Xeon Platinum 8153 | 2 | 1370 | ||
Xeon Platinum 8160 | 2 | 2170 | ||
Xeon Platinum 8164 | 2 | 2220 | ||
Xeon Platinum 8170 | 2 | 2310 | ||
Xeon Platinum 8176 | 2 | 2440 | ||
Xeon Platinum 8180 | 2 | 2670 | 2780 | 2870 |
Xeon Silver 4112 | 2 | 426 | ||
Xeon Gold 5122 | 2 | 547 | ||
Xeon Gold 6128 | 2 | 822 | ||
Xeon Gold 6134 | 2 | 1060 | ||
Xeon Gold 6144 | 2 | 1120 | ||
Xeon Gold 6126 | 2 | 1310 | ||
Xeon Gold 6136 | 2 | 1480 | ||
Xeon Gold 6146 | 2 | 1540 | ||
Xeon Gold 6132 | 2 | 1540 | ||
Xeon Gold 6142 | 2 | 1710 | ||
Xeon Gold 6150 | 2 | 1900 | ||
Xeon Gold 6154 | 2 | 2090 | ||
Xeon Platinum 8168 | 2 | 2460 | ||
Xeon Silver 4114T | 2 | 910 | ||
Xeon Gold 5119T | 2 | 1190 | ||
Xeon Gold 6134M | 2 | 1060 | ||
Xeon Gold 6140M | 2 | 1540 | ||
Xeon Gold 6142M | 2 | 1710 | ||
Xeon Platinum 8160M | 2 | 2170 | ||
Xeon Platinum 8170M | 2 | 2310 | ||
Xeon Platinum 8176M | 2 | 2440 | ||
Xeon Platinum 8180M | 2 | 2670 | 2780 |
Processor | Number of processors | SPECfp_rate_base2006 | SPECfp_rate2006 | SPECfp_rate_base2006 Version 18.0.0.128 of Intel C++ Compiler |
Xeon Bronze 3104 | 2 | 364 | ||
Xeon Bronze 3106 | 2 | 481 | ||
Xeon Silver 4108 | 2 | 640 | ||
Xeon Silver 4110 | 2 | 690 | ||
Xeon Silver 4114 | 2 | 838 | ||
Xeon Silver 4116 | 2 | 938 | ||
Xeon Gold 5115 | 2 | 874 | ||
Xeon Gold 5118 | 2 | 993 | ||
Xeon Gold 5120 | 2 | 1080 | ||
Xeon Gold 6130 | 2 | 1270 | ||
Xeon Gold 6140 | 2 | 1380 | ||
Xeon Gold 6138 | 2 | 1390 | ||
Xeon Gold 6148 | 2 | 1470 | ||
Xeon Gold 6152 | 2 | 1490 | 1520 | |
Xeon Platinum 8153 | 2 | 1170 | ||
Xeon Platinum 8160 | 2 | 1560 | ||
Xeon Platinum 8164 | 2 | 1590 | ||
Xeon Platinum 8170 | 2 | 1630 | ||
Xeon Platinum 8176 | 2 | 1680 | ||
Xeon Platinum 8180 | 2 | 1770 | 1800 | |
Xeon Silver 4112 | 2 | 430 | ||
Xeon Gold 5122 | 2 | 534 | ||
Xeon Gold 6128 | 2 | 780 | ||
Xeon Gold 6134 | 2 | 968 | ||
Xeon Gold 6144 | 2 | 996 | ||
Xeon Gold 6126 | 2 | 1130 | ||
Xeon Gold 6136 | 2 | 1230 | ||
Xeon Gold 6146 | 2 | 1260 | ||
Xeon Gold 6132 | 2 | 1250 | ||
Xeon Gold 6142 | 2 | 1350 | ||
Xeon Gold 6150 | 2 | 1430 | ||
Xeon Gold 6154 | 2 | 1520 | ||
Xeon Platinum 8168 | 2 | 1690 | ||
Xeon Silver 4114T | 2 | 839 | ||
Xeon Gold 5119T | 2 | 1020 | ||
Xeon Gold 6134M | 2 | 968 | ||
Xeon Gold 6140M | 2 | 1380 | ||
Xeon Gold 6142M | 2 | 1350 | ||
Xeon Platinum 8160M | 2 | 1560 | ||
Xeon Platinum 8170M | 2 | 1630 | ||
Xeon Platinum 8176M | 2 | 1680 | ||
Xeon Platinum 8180M | 2 | 1770 | 1800 |
Performance Comparison (SPECcpu2006): Two bar charts illustrate the performance improvement of the PRIMERGY RX2530 M4 over its predecessor, the RX2530 M2. The M4 shows significant gains in both integer (SPECint_rate2006) and floating-point (SPECfp_rate2006) performance, attributed to its advanced Intel Xeon processors.
OLTP-2 Benchmark
OLTP (Online Transaction Processing) measures database performance by simulating typical application scenarios and reporting transactions per second (tps). OLTP-2 is an internal Fujitsu benchmark based on TPC-E, designed to test system scaling with CPU and memory configurations. The benchmark setup involves a client, application server, database server, and disk subsystem.
Benchmark Environment: The OLTP-2 benchmark was conducted using a specific configuration involving a PRIMERGY RX2530 M4 as the database server, a PRIMERGY RX2530 M2 as the application server, and another PRIMERGY RX2530 M2 as the client. The database server utilized Microsoft SQL Server 2017 Enterprise, while the application and client servers ran Microsoft Windows Server 2012 R2 Standard.
Benchmark Results: Database performance is highly dependent on CPU, memory, and disk subsystem configuration. For optimal memory selection, sufficient quantity is prioritized over speed. Configurations with 1536 GB memory for dual-processor systems and 768 GB for single-processor systems were used, both with 2666 MHz memory access speed.
The comparison of current system generations (RX2530 M4/RX2540 M4) with predecessor systems (RX2530 M2/RX2540 M2) shows an approximate 40% increase in maximum achievable OLTP-2 values.
vServCon Benchmark
vServCon is a Fujitsu benchmark for comparing server consolidation suitability, evaluating hypervisors, virtualization forms, and I/O technologies. It combines existing benchmarks for database, application server, and web server workloads within virtual machines (VMs). The benchmark measures performance scaling by increasing the number of VMs or 'tiles' (groups of VMs).
Benchmark Environment: The vServCon benchmark environment uses a System Under Test (SUT) consisting of a PRIMERGY RX2530 M4 with dual Intel® Xeon® Processor Scalable Family CPUs and 24 x 32 GB memory. Load generators were PRIMERGY RX2530 M2 servers. The operating system for the SUT was VMware ESXi 6.5.0b.
Benchmark Results: The results show significant performance improvements across different Intel Xeon processors. The document highlights that increased physical cores, Hyper-Threading, and CPU clock frequency contribute to higher VM density and performance. Performance increases from 2008 to 2017 are detailed, showing advancements in processor technology leading to higher vServCon scores. The comparison of system generations shows substantial improvements in virtualization performance.
Diagrams:
- A bar chart compares vServCon scores for various Intel Xeon processors, showing scores and the number of tiles used.
- Diagrams illustrate virtualization performance for increasing numbers of VMs with Xeon Gold 6130 processors.
- Another diagram shows the scaling factor of 1.94 for dual-processor configurations compared to single-processor configurations, using Xeon Platinum 8180 as an example.
STREAM Benchmark
STREAM is a synthetic benchmark developed by John McCalpin to measure memory throughput for sequential data accesses. It outputs results in GB/s, with the TRIAD value commonly used for comparison. The benchmark's results are influenced by memory clock frequency and processor arithmetic capabilities.
Type | Execution | Bytes per step | Floating-point calculation per step |
COPY | a(i) = b(i) | 16 | 0 |
SCALE | a(i) = q × b(i) | 16 | 1 |
SUM | a(i) = b(i) + c(i) | 24 | 1 |
TRIAD | a(i) = b(i) + q × c(i) | 24 | 2 |
Benchmark Environment: The STREAM benchmark was run on a PRIMERGY RX2530 M4 with dual Intel® Xeon® Processor Scalable Family CPUs and 16 GB memory. The operating system was SUSE Linux Enterprise Server 12 SP2 (x86_64), with optimized system settings. Specific processors were run with AVX512 or AVX2 instruction sets.
Benchmark Results: The results table shows Memory Frequency, Memory Bandwidth, Processor Cores, Processor Frequency, Number of Processors, and TRIAD values for various Intel Xeon processors. A bar chart compares the STREAM TRIAD throughput of the PRIMERGY RX2530 M4 against its predecessor, the PRIMERGY RX2530 M2, demonstrating significant performance gains for the M4 across different processor configurations.
LINPACK Benchmark
LINPACK, developed in the 1970s, measures computer performance by solving a system of linear equations. It is a key benchmark in High-Performance Computing (HPC). The result is reported as Rmax (GFlops), representing the measured floating-point operations per second. Rpeak is the theoretical maximum performance.
Benchmark Environment: The LINPACK benchmark was executed on a PRIMERGY RX2530 M4 with dual Intel® Xeon® Processor Scalable Family CPUs and 16 GB memory. The setup utilized SUSE Linux Enterprise Server 12 SP2 (x86_64) and Intel® Math Kernel Library for benchmarks.
Benchmark Results: The results table provides Processor, Frequency, Cores, Number of Processors, Rpeak, Rmax, and Efficiency for various Intel Xeon processors. It notes that Intel does not guarantee maximum turbo frequency and that AVX instruction usage can affect performance.
Literature and Contact
The document provides links to further information on PRIMERGY Servers, specific benchmarks (OLTP-2, SAP SD, SPECcpu2006, STREAM, vServCon, LINPACK), and contact details for FUJITSU PRIMERGY Product Marketing and Performance and Benchmarks teams.
- PRIMERGY Servers: primergy.com
- PRIMERGY RX2530 M4 White Paper: Link 1, Link 2
- Data Sheet: Link
- PRIMERGY Performance: fujitsu.com/fts/x86-server-benchmarks
- OLTP-2 Benchmark Overview: Link
- SAP SD Benchmark Overview: sap.com/benchmark, Link
- SPECcpu2006: spec.org/osg/cpu2006, Link
- STREAM: cs.virginia.edu/stream/
- vServCon Benchmark Overview: Link
- LINPACK Benchmark: netlib.org/utk/people/JackDongarra/PAPERS/hplpaper.pdf
- TOP500: top500.org/
- HPL: netlib.org/benchmark/hpl/
- Intel Math Kernel Library: software.intel.com/en-us/articles/intel-math-kernel-library-linpack-download/
Contact:
- FUJITSU Website: www.fujitsu.com/
- PRIMERGY Product Marketing: Primergy-PM@ts.fujitsu.com
- PRIMERGY Performance and Benchmarks: primergy.benchmark@ts.fujitsu.com
Copyright © 2018 Fujitsu Technology Solutions. All rights reserved.