Updated July 2024
The Reference Test Platform (RTP) is comprised of hardware and software that was used by the Solid State Storage Technical Work Group (SSS TWG) to validate the SSS Performance Test Specification (SSS PTS). The RTP is not required to run the SSS PTS tests; it is the platform that was used to validate the PTS and is thus known to be appropriate for PTS SSD testing. The RTP is recommended both because it is known to minimize performance bottlenecks as well as provide a standardized and normalized test environment for comparative performance measurement.
RTP Configurations
The RTP is designed to enable testing of SSDs. The table below shows the RTP 4.0 and the newly approved (July 2016) RTP 5.0 refresh. RTP 4.0 is a DDR3 dual socket 8 core CPU and is suitable for testing Client and Enterprise SSDs including mSATA, SATA, SAS and M.2 SSDs. RTP 5.0 is a higher performance DDR4 dual socket 12 core CPU system that is intended to be used for faster Enterprise PCIe AIC and NVDIMM-N SSDs. The RTP 5.0 is a preliminary listing. Users may select either the DDR3 or DDR4 platform/CPU for use with the RTP 5.0. The SSSI is solicting feedback from users of the RTP 5.0. Please send comments or questions to
[email protected].
|
Component
|
RTP 4.0
|
Notes
|
RTP 5.0
|
Notes
|
|
Chassis
|
Intel P4308XXMHJC |
4u
|
Intel P4308XXMHJC
|
4u
|
|
Motherboard
|
Gen 3 Intel S2600 COE
|
dual socket
|
Gen 3 Intel S2600 CWR2 or
Gen 3 Intel S2600 COE
|
dual socket
|
|
Processor
|
Intel 8 core XEON E5 2670 v2 2.3 Ghz
|
2 CPU's
|
Intel 12 core E5-2670 v3 3.4 Ghz or
Intel 8 core XEON E5 2670 v2 2.3 Ghz
|
2 CPU's
|
|
Main Memory
|
1600Mhz DDR3 ECC
|
32GB
|
2133Mhz DDR4 ECC or
1600 Mhz DDR3 ECC
|
32GB to 64GB
|
|
Boot Drive(s)
|
User Selectable
|
|
User Selectable
|
|
|
HBA
|
LSI 6Gb/s
9212-4i4e
|
HBA
HBA
M.2
U.2
PCIe
|
LSI 6Gb/s 9212-4e4i
LSI 12Gb/s 9300-4i
DT-119
SC PCIe x4-SFF-8639
QTL1789
|
User choice
User choice
Passive
Passive w/ power tap
Passive w/ power tap
|
|
Operating System
|
Linux
|
CentOS 6.x
|
|
CentOS 6.x - 7.x
|
|
Test Software
|
Calypso CTS
|
CTS 6.5
|
Calypso CTS
|
CTS 6.5
|
RTP Components
Chassis:
The chassis can effect performance test measurements depending on the type of device being tested (see below) and should have an adequate power supply and adequate air flow over all test SSDs - or Device Under Test (or DUTs).
The RTP 4.0 uses an Intel P4308XXMHJC 4U. The chassis has more impact on the testing of higher performance PCIe and SAS DUTs due to the increased power usage and consequent heat dissipation. The RTP 4.0 chassis has an internal airflow specification for use with larger power suppplies and higher performance SAS, U.2 and PCIe SSDs that draw 12W to 25W per DUT or more.
Chassis and motherboard specs list minimum airflow in Linear Feet per Minute (LFM) or Cubic Feet per Minute (CFM). Associated DUT specs may also list minimum DUT airflow specs in LFM or CFM over the DUT surface or at server entry or exit points and should be consulted by the test operator. The same chassis, or its equivalent, is recommended for the RTP 5.0.
Motherboard:
The RTP 4.0 motherboard is a Gen 3 Intel S2600 COE DDR3 with dual CPU sockets. A similar motherboard can be used, and if so, must be disclosed. To support the testing of higher performance DUTs, it is recommended to populate two processor sockets for dual CPU usage.
The RTP 5.0 motherboard is a Gen 3 Intel S2600 CWR2 DDR4 with dual CPU sockets. While a similar motherboard can be used, it is recommended to use this specific motherboard for comparative performance test as the motherboard can have significant impact on performance.
CPU:
The processor used in the RTP 4.0 is an eight core Intel E5-2670 v2. The Standard configuration has two CPUs.
The processors used in the RTP 5.0 are twelve core Intel E5-2670 v3. The Standard configuration has two CPUs.
Memory:
The RTP 4.0 main memory utilizes 1600MHz DDR3 DIMMs with ECC. The Standard configuration has 32GB of RAM.
The RTP 5.0 main memory utilizes 2133MHz DDR4 DIMMs with ECC. The Standard configuration has 32GB to 64GB RAM.
It is advisable to install the memory in multiples of three DIMMs, to optimize performance.
Boot Drive(s):
Boot drives are deemed not to affect operation of the RTP nor have an impact on performance testing. Accordingly, the boot drive(s) is user selectable.
HBA:
The RTP 4.0 uses an LSI 9212-4i4e Host Bus Adaptor (HBA) containing the IT firmware set. Other sufficiently high performance HBAs are available, but should be disclosed if used. Performance can vary significantly depending on the type of HBA used. The RTP 5.0 is listed with both the 6Gb/s LSI 9212-4e4i and the LSI 12Gb/s 9300-4i.
U.2 Interposer Card: RTP 5.0
U.2 SSDs are four lane 2.5" SSDs that plug in to the PCIe bus via an interposer card. RTP 5.0 recommends use of the Serial Cable PCIe x4-SFF-8639 passive (no switch) interposer card. This interposer card also provides passive power supply interrupt pins (taps) for connection to power supply/power measurement devices.
M.2 Interposer Card: RTP 5.0
M.2 SSDs are two to four lane SSDs that plug in to the PCIe bus or SATA port via an interposer card. RTP 5.0 recommends use of the mFactors DT-119 or equivalent passive (no switch) interposer card for insertion into the PCIe bus.
PCIe x8x16 Interposer Card: RTP 5.0
The Quarch QTL1789 passive PCIe x8x16 interposer cards with interrupt pins (taps) for power source tap or connection to power measurement devices are recommended in RTP 5.0.
DUT Power Supply:
The RTP hardware platform has a dedicated DUT power supply for all SAS/SATA test bays separate from the main power supply. The DUT power supply must provide adequate power to simultaneously run all test DUTs with a surplus of 25% of the rated peak voltage draw for all DUTs that may be tested - i.e. cumulative potential draw for each DUT test slot.
Operating System (OS):
Both RTP 4.0 and 5.0 utilize Linux Community Enterprise OS (CentOS) ver 6.x or 7.x. The OS kernel revision used must be disclosed in the PTS report. Users may select to run the PTS using other OSes but must disclose the OS that is used (such as Windows Server 2012 R2).
Test Software:
The test software used to validate the RTP / PTS-E v1.1 and PTS-C v1.2 is CTS 6.5 made by Calypso Systems, Inc. (
www.calypsotesters.com).
The Calypso test software requires CentOS 6.x Linux OS. It is also available in Windows Server 2012 R2.
