SSD性能测试规范详解

5星 · 超过95%的资源需积分: 50 38 浏览量更新于2024-07-18 收藏 3.98MB PDF 举报

"SSD测试介绍 - Solid State Storage (SSS) Performance Test Specification (PTS) Version 2.0.1" SSD（固态存储）是现代计算机系统中广泛使用的高速存储技术，它基于NAND闪存，替代了传统的机械硬盘（HDD）。《Solid State Storage (SSS) Performance Test Specification (PTS)》是一份由SNIA（存储网络行业协会）制定的技术文档，旨在提供一套准确、可重复且可靠的测试方法，用于比较不同形态、协议和接口的NAND闪存基固态存储产品在客户端和企业应用中的性能。该规范的目的是确保不同厂商的SSD产品在性能评估时有统一的标准，使得用户可以公平地对比不同产品的读写速度、延迟、耐用性和能效等关键指标。测试套件包括了一系列针对随机读写、顺序读写、多任务并发处理能力以及耐用性等方面的具体测试项目，这些项目涵盖了日常使用和极端工作负载的各种情况。 PTS的测试方法论涵盖了以下关键方面： 1. **基准测试工具**：文档中可能包含了一组推荐或强制使用的基准软件，如Iometer、FIO等，用于模拟实际工作负载并度量SSD性能。 2. **测试环境**：定义了理想的测试环境，包括硬件平台、操作系统配置、驱动程序版本等，以减少外部因素对测试结果的影响。 3. **性能指标**：详细说明了如何衡量SSD的读写速度、IOPS（每秒输入/输出操作次数）、响应时间、功耗等核心指标。 4. **耐用性测试**：除了性能测试，PTS还关注SSD的寿命，包括写入耐久性（P/E周期）和数据保留能力。 5. **报告格式**：规定了测试结果的呈现方式，以便于用户理解和比较不同产品的性能差异。 6. **一致性与标准化**： PTS强调了测试过程的标准化，确保所有参与测试的SSD在相同的条件下进行，避免了由于测试方法不一致导致的误导性结果。 7. **反馈机制**：SNIA鼓励业界人士对PTS提出改进建议，以持续优化测试规范，保持其与时俱进。通过遵循PTS，SSD制造商和评测机构可以提供公正的性能数据，帮助消费者和企业做出明智的采购决策。同时，该规范也有助于推动SSD技术的发展，促进性能提升和成本优化。对于IT专业人士来说，理解PTS的内容和测试流程是评估SSD性能的关键，也是选择合适存储解决方案的重要依据。

SSS PTS SNIA Technical Position 16

Version 2.0.1

• JEDEC JESD218 – Solid-State Drive (SSD) Requirements and Endurance Test Method

• JEDEC JESD219 – Solid-State Drive (SSD) Endurance Workloads

1.7.2 Other references

• None in this version

SSS PTS SNIA Technical Position 17

Version 2.0.1

2 Definitions, symbols, abbreviations, and conventions

2.1 Definitions

2.1.1 ActiveRange: Specified as ActiveRange(start:end), where “start” and “end” are

percentages. ActiveRange (AR) is the range of LBA’s that may be accessed by the pre-

conditioning and/or test code, where the starting LBA# = start%*MaxUserLBA and the

ending LBA# = end%*MaxUserLBA.

2.1.2 Cache: A volatile or non-volatile data storage area outside the User Capacity that may

contain a subset of the data stored within the User Capacity.

2.1.3 Client: Single user desktop or laptop system used in home or office.

2.1.4 CPU Usage: amount of time for which a central processing unit (CPU) is used for

processing instructions. CPU time is also measured as a percentage of the CPU's

capacity at any given time.

2.1.5 Data Excursion: As used in the definition of Steady State, shall be measured by taking

the absolute value of the difference between each sample and the average.

2.1.6 Enterprise: Servers in data centers, storage arrays, and enterprise wide / multiple user

environments that employ direct attached storage, storage attached networks and tiered

storage architectures.

2.1.7 Client: laptop or desktop computers used in small office, home, mobile, entertainment

and other single user applications.

2.1.8 Fresh Out of the Box (FOB): State of SSS prior to being put into service.

2.1.9 IO Demand: Measured # of OIOs executing in the host.

2.1.10 Logical Block Address (LBA): The address of a logical block, i.e., the offset of the block

from the beginning of the logical device that contains it.

2.1.11 Latency: The time between when the workload generator makes an IO request and when

it receives notification of the request’s completion.

2.1.12 MaxUserLBA: The maximum LBA # addressable in the User Capacity.

2.1.13 Measurement Window: The interval, measured in Rounds, during which test data is

collected, bounded by the Round in which the device has been observed to have

maintained Steady State for the specified number of Rounds (Round x), and five Rounds

previous (Round x-4).

2.1.14 Nonvolatile Cache: A cache that retains data through power cycles.

2.1.15 Outstanding IO (OIO): The number of IO operations issued by a host, or hosts, awaiting

completion.

2.1.16 OIO/Thread: The number of OIO allowed per Thread (Worker, Process)

2.1.17 Over-Provisioned Capacity: LBA range provided by the manufacturer for performance

and endurance considerations, but not accessible by the host file system, operating

system, applications, or user.

2.1.18 Pre-conditioning: The process of writing data to the device to prepare it for Steady State

measurement.

(a) Workload Independent Pre-conditioning (WIPC): The technique of running a

prescribed workload, unrelated to the test workload, as a means to facilitate convergence

to Steady State.

(b) Workload Dependent Pre-conditioning (WDPC): The technique of running the test

workload itself, typically after Workload Independent Pre-conditioning, as a means to put

SSS PTS SNIA Technical Position 18

Version 2.0.1

the device in a Steady State relative to the dependent variable being tested.

2.1.19 Pre-conditioning Code: Refers to the Pre-conditioning steps set forth in this

Specification.

2.1.20 Purge: The process of returning an SSS device to a state in which subsequent writes

execute, as closely as possible, as if the device had never been used and does not contain

any valid data.

2.1.21 Round: A complete pass through all the prescribed test points for any given test.

2.1.22 Queue Depth: Interchangeably refers to the OIO/Thread produced by the Workload

Generator.

2.1.23 Slope: As used in the definition of Steady State, shall mean the slope of the “Best Linear

Fit Line.”

2.1.24 Steady State: A device is said to be in Steady State when, for the dependent variable (y)

being tracked:

a) Range(y) is less than 20% of Ave(y): Max(y)-Min(y) within the Measurement Window

is no more than 20% of the Ave(y) within the Measurement Window; and

b) Slope(y) is less than 10%: Max(y)-Min(y), where Max(y) and Min(y) are the maximum

and minimum values on the best linear curve fit of the y-values within the Measurement

Window, is within 10% of Ave(y) value within the Measurement Window.

2.1.25 Test Code: Refers to the measurement steps set forth in the test sections contained in

this Specification.

2.1.26 Transition Zone: A performance state where the device’s performance is changing as it

goes from one state to another (such as from FOB to Steady State).

2.1.27 Thread: Execution context defined by host OS/CPU (also: Process, Worker)

2.1.28 Thread Count (TC): Number of Threads (or Workers or Processes) specified by a test.

2.1.29 Total OIO: Total outstanding IO Operations specified by a test = (OIO/Thread) * (TC)

2.1.30 User Capacity: LBA range directly accessible by the file system, operating system and

applications, not including Over-Provisioned Capacity.

2.1.31 Volatile Cache: A cache that does not retain data through power cycles.

2.2 Acronyms and Abbreviations

2.2.1 IOPS: I/O Operations per Second

2.2.2 DUT: Device Under Test

2.2.3 DI: Demand Intensity (aka Total OIO)

2.2.4 FOB: Fresh Out of Box

2.2.5 OIO: Outstanding IO

2.2.6 TOIO: Total Outstanding IO

2.2.7 R/W: Read/Write

2.2.8 SSSI: Solid State Storage Initiative

2.2.9 SSS TWG: Solid State Storage Technical Working Group

2.2.10 TC: Thread Count

2.2.11 ART: Average Response Time

2.2.12 QD: Queue Depth

2.2.13 TP: Throughput

SSS PTS SNIA Technical Position 19

Version 2.0.1

2.2.14 LAT: Latency

2.2.15 WSAT: Write Saturation

2.2.16 HIR: Host Idle Recovery

2.2.17 XSR: Cross Stimulus Recovery

2.2.18 ECW: Enterprise Composite Workload

2.2.19 DIRTH: Demand Intensity / Response Time Histogram

2.3 Keywords

The key words “shall”, “required”, “shall not”, “should”, “recommended”, “should not”, “may”, and

“optional” in this document are to be interpreted as:

2.3.1 Shall: This word, or the term "required", means that the definition is an absolute

requirement of the specification.

2.3.2 Shall Not: This phrase means that the definition is an absolute prohibition of the

specification.

2.3.3 Should: This word, or the adjective "recommended", means that there may be valid

reasons in particular circumstances to ignore a particular item, but the full implications

must be understood and weighed before choosing a different course.

2.3.4 Should Not: This phrase, or the phrase "not recommended", means that there may

exist valid reasons in particular circumstances when the particular behavior is

acceptable or even useful, but the full implications should be understood and the case

carefully weighed before implementing any behavior described with this label.

2.3.5 May: This word, or term “optional”, indicates flexibility, with no implied preference.

2.4 Conventions

2.4.1 Number Conventions

Numbers that are not immediately followed by lower-case b or h are decimal values.

Numbers immediately followed by lower-case b (xxb) are binary values.

Numbers immediately followed by lower-case h (xxh) are hexadecimal values.

Hexadecimal digits that are alphabetic characters are upper case (i.e., ABCDEF, not abcdef).

Hexadecimal numbers may be separated into groups of four digits by spaces. If the number is not

a multiple of four digits, the first group may have fewer than four digits (e.g., AB CDEF 1234

5678h).

Storage capacities and data transfer rates and amounts shall be reported in Base-10. IO transfer

sizes and offsets shall be reported in Base-2. The associated units and abbreviations used in this

specification are:

•

A kilobyte (KB) is equal to 1,000 (10

) bytes.

•

A megabyte (MB) is equal to 1,000,000 (10

) bytes.

•

A gigabyte (GB) is equal to 1,000,000,000 (10

) bytes.

•

A terabyte (TB) is equal to 1,000,000,000,000 (10

) bytes.

•

A petabyte (PB) is equal to 1,000,000,000,000,000 (10

) bytes

•

A kibibyte (KiB) is equal to 2

bytes.

•

A mebibyte (MiB) is equal to 2

bytes.

•

A gibibyte (GiB) is equal to 2

bytes.

•

A tebibyte (TiB) is equal to 2

bytes.

•

A pebibyte (PiB) is equal to 2

bytes

剩余101页未读，继续阅读

weixin_43216647

粉丝: 0
资源: 1

SSD性能测试规范详解

性能测试PTS脚本调试指南

SSS_PTS_Enterprise_v1.1_Solid_State_Storage_Performance_Test_Specification.pdf

三星SSD硬盘检测程序

ssd测试工具

ssd 测试工具

SSD深入浅出之SSD测试

ssd测试工具汇总

ssd算法测试程序

CrystalDiskMark5.0.2(SSD测试)中文版.rar

SSD测试详解：DevSlp支持与FIO性能测试

最新资源