Programming Persistent Memory : A Comprehensive Guide for Developers

Beginning and experienced programmers will use this comprehensive guide to persistent memory programming. You will understand how persistent memory brings together several new software/hardware requirements, and offers great promise for better performance and faster application startup times—a huge leap forward in byte-addressable capacity compared with current DRAM offerings.This revolutionary new technology gives applications significant performance and capacity improvements over existing technologies. It requires a new way of thinking and developing, which makes this highly disruptive to the IT/computing industry. The full spectrum of industry sectors that will benefit from this technology include, but are not limited to, in-memory and traditional databases, AI, analytics, HPC, virtualization, and big data. Programming Persistent Memory describes the technology and why it is exciting the industry. It covers the operating system and hardware requirements as well as how to create development environments using emulated or real persistent memory hardware. The book explains fundamental concepts; provides an introduction to persistent memory programming APIs for C, C++, JavaScript, and other languages; discusses RMDA with persistent memory; reviews security features; and presents many examples. Source code and examples that you can run on your own systems are included.What You’ll Learn* Understand what persistent memory is, what it does, and the value it brings to the industry* Become familiar with the operating system and hardware requirements to use persistent memory* Know the fundamentals of persistent memory programming: why it is different from current programming methods, and what developers need to keep in mind when programming for persistence* Look at persistent memory application development by example using the Persistent Memory Development Kit (PMDK)* Design and optimize data structures for persistent memory* Study how real-world applications are modified to leverage persistent memory* Utilize the tools available for persistent memory programming, application performance profiling, and debuggingWho This Book Is ForC, C++, Java, and Python developers, but will also be useful to software, cloud, and hardware architects across a broad spectrum of sectors, including cloud service providers, independent software vendors, high performance compute, artificial intelligence, data analytics, big data, etc. Table of Contents......Page 4 About the Author......Page 14 About the Technical Reviewer......Page 15 About the Contributors......Page 16 Acknowledgments......Page 20 Preface......Page 21 Chapter 1: Introduction to Persistent Memory Programming......Page 30 A High-Level Example Program......Page 31 What’s Different?......Page 34 The Performance Difference......Page 35 Program Complexity......Page 36 How Does libpmemkv Work?......Page 37 Summary......Page 38 Chapter 2: Persistent Memory Architecture......Page 40 Persistent Memory Characteristics......Page 41 Platform Support for Persistent Memory......Page 42 Cache Hierarchy......Page 43 Power-Fail Protected Domains......Page 45 The Need for Flushing, Ordering, and Fencing......Page 48 Data Visibility......Page 52 Intel Machine Instructions for Persistent Memory......Page 53 Detecting Platform Capabilities......Page 54 Application Startup and Recovery......Page 56 Summary......Page 58 Operating System Support for Memory and Storage......Page 60 Persistent Memory As Block Storage......Page 62 Persistent Memory-Aware File Systems......Page 63 Memory-Mapped Files......Page 64 Persistent Memory Direct Access (DAX)......Page 72 Summary......Page 82 What’s Different?......Page 84 Atomic Updates......Page 85 Atomicity......Page 86 Durability......Page 87 Start-Time Responsibilities......Page 88 Summary......Page 89 Background......Page 91 Choosing the Right Semantics......Page 92 libmemkind......Page 93 libvmemcache......Page 94 libpmem......Page 95 libpmemobj-cpp......Page 96 libpmemblk......Page 97 pmemcheck......Page 98 Summary......Page 99 Chapter 6: libpmem: Low-Level Persistent Memory Support......Page 101 Using the Library......Page 102 Mapping a File......Page 103 Copying to Persistent Memory......Page 104 Separating the Flush Steps......Page 105 Summary......Page 107 What is libpmemobj?......Page 108 Why not malloc()?......Page 109 Creating Memory Pools......Page 110 Pool Object Pointer (POP) and the Root Object......Page 114 Opening and Reading from Memory Pools......Page 115 Concatenated Poolsets......Page 117 Replica Poolsets......Page 118 Typed Object Identifiers (TOIDs)......Page 119 Allocating Memory......Page 120 Atomic Operations......Page 121 Reserve/Publish API......Page 124 Transactional API......Page 127 Persisting Data Summary......Page 131 Guarantees of libpmemobj's APIs......Page 132 Debugging and Error Handling......Page 133 Summary......Page 135 Introduction......Page 137 Persistent Pointers......Page 138 Transactions......Page 139 Snapshotting......Page 141 Allocating......Page 142 C++ Standard limitations......Page 144 An Object’s Lifetime......Page 145 Trivial Types......Page 146 Object Layout......Page 148 Pointers......Page 149 Limitations Summary......Page 151 Persistence Simplified......Page 152 The Ecosystem......Page 159 Persistent Containers......Page 160 Examples of Persistent Containers......Page 161 Summary......Page 164 Chapter 9: pmemkv: A Persistent In-Memory Key-Value Store......Page 166 pmemkv Architecture......Page 168 A Phonebook Example......Page 172 Bringing Persistent Memory Closer to the Cloud......Page 176 Summary......Page 177 Introduction......Page 179 How it Works......Page 180 Supported “Kinds” of Memory......Page 181 Kind Management API......Page 183 Creating a Fixed-Size Heap......Page 184 Automatic Kind Detection......Page 186 Memory Kind Detection......Page 187 Heap Management API......Page 188 Allocating Memory......Page 189 Freeing Allocated Memory......Page 190 Memory Usage Policy......Page 191 C++ Allocator for PMEM Kind......Page 192 Nested Containers......Page 193 Using the pmem::allocator......Page 194 Creating a Vector of Strings......Page 195 Expanding Volatile Memory Using Persistent Memory......Page 197 libvmemcache: An Efficient Volatile Key-Value Cache for Large-Capacity Persistent Memory......Page 201 libvmemcache Overview......Page 202 Extent-Based Allocator......Page 204 Scalable Replacement Policy......Page 206 Using libvmemcache......Page 207 Summary......Page 210 Contiguous Data Structures and Fragmentation......Page 211 Internal and External Fragmentation......Page 212 Transactions......Page 213 Copy-on-Write and Versioning......Page 216 Example Data Structures......Page 217 Hash Table with Transactions......Page 218 Hash Table with Transactions and Selective Persistence......Page 221 Sorted Array with Versioning......Page 226 Summary......Page 230 Chapter 12: Debugging Persistent Memory Applications......Page 231 Stack Overflow Example......Page 232 Memory Leak Example......Page 233 Intel Inspector – Persistence Inspector......Page 234 Stack Overflow Example......Page 235 Memory Leak Example......Page 236 Nonpersistent Stores......Page 238 Stores Not Added into a Transaction......Page 252 Memory Added to Two Different Transactions......Page 257 Memory Overwrites......Page 264 Unnecessary Flushes......Page 266 Out-of-Order Writes......Page 271 Summary......Page 283 Chapter 13: Enabling Persistence Using a Real-World Application......Page 285 Different Persistent Memory Enablement Approaches......Page 286 Developing a Persistent Memory-Aware MariaDB* Storage Engine......Page 287 Understanding the Storage Layer......Page 288 Creating a Storage Engine Class......Page 289 Creating a Database Table......Page 290 Opening a Database Table......Page 291 INSERT Operation......Page 292 UPDATE Operation......Page 294 DELETE Operation......Page 296 SELECT Operation......Page 299 Summary......Page 300 Chapter 14: Concurrency and Persistent Memory......Page 301 Transactions and Multithreading......Page 302 Mutexes on Persistent Memory......Page 306 Lock-Free Algorithms and Persistent Memory......Page 309 Concurrent Data Structures for Persistent Memory......Page 310 Concurrent Ordered Map......Page 311 Find Operation......Page 312 Insert Operation......Page 313 Concurrent Hash Map......Page 315 Insert Operation......Page 316 Summary......Page 317 Compute-Bound vs. Memory-Bound......Page 319 Memory Access Patterns......Page 320 Determining the Suitability of Workloads for Persistent Memory......Page 321 Identifying Workloads That Are Memory-Capacity Bound......Page 322 Identifying the Hot Working Set Size of a Workload......Page 324 Use Cases Requiring Persistence......Page 325 Performance Analysis of Workloads Using Persistent Memory......Page 326 Memory Bandwidth and Latency......Page 327 Non-Uniform Memory Architecture (NUMA) Behavior......Page 329 Addressable Memory Capacity......Page 330 Guided Data Placement......Page 331 Memory Access Optimization......Page 332 Data Allocation vs. First Access......Page 333 Thread Migration......Page 334 Summary......Page 335 A Pool of Persistent Memory: High-Level Architecture Overview......Page 337 The Uncertainty of Memory Mapping: Persistent Memory Object Identifier......Page 339 Persistent Thread Local Storage: Using Lanes......Page 342 Transaction Redo Logging......Page 344 Transaction Undo Logging......Page 345 libpmemobj Unified Logging......Page 346 Persistent Allocations: The Interface of a Transactional Persistent Allocator......Page 347 Persistent Memory Heap Management: Allocator Design for Persistent Memory......Page 348 ACID Transactions: Efficient Low-Level Persistent Transactions......Page 352 Lazy Reinitialization of Variables: Storing the Volatile State on Persistent Memory......Page 354 Summary......Page 355 Dealing with Uncorrectable Errors......Page 356 Consumed Uncorrectable Error Handling......Page 357 Unconsumed Uncorrectable Error Handling......Page 359 Patrol Scrub......Page 360 Address Range Scrub......Page 361 Device Health......Page 362 Vendor-Specific Device Health (_DSMs)......Page 365 Unsafe/Dirty Shutdown......Page 366 Application Utilization of Data Loss Count (DLC)......Page 367 Summary......Page 369 Chapter 18: Remote Persistent Memory......Page 370 RDMA Networking Protocols......Page 371 Guaranteeing Remote Persistence......Page 374 General-Purpose Remote Replication Method......Page 376 How Does the General-Purpose Remote Replication Method Make Data Persistent?......Page 377 Appliance Remote Replication Method......Page 378 General Software Architecture......Page 380 librpmem Architecture and Its Use in Replication......Page 381 Performance Considerations......Page 385 Say Hello to the Replicated World......Page 387 Execution Example......Page 392 Summary......Page 393 Nonuniform Memory Access (NUMA)......Page 395 NUMACTL Linux Utility......Page 396 NDCTL Linux Utility......Page 398 Intel Memory Latency Checker Utility......Page 400 NUMASTAT Utility......Page 402 IPMCTL Utility......Page 403 Automatic NUMA Balancing......Page 404 Using Volume Managers with Persistent Memory......Page 405 The mmap() MAP_SYNC Flag......Page 407 Summary......Page 408 Prerequisites......Page 410 Installing NDCTL and DAXCTL Using the Linux Distribution Package Repository......Page 411 Searching for Packages Within a Package Repository......Page 412 Installing NDCTL and DAXCTL from the Package Repository......Page 413 Installing PMDK on RHEL and CentOS 7.5 or Later......Page 414 Installing PMDK on Ubuntu 18.04 or Later......Page 415 Installing PMDK Using the Linux Distribution Package Repository......Page 416 Searching for Packages Within a Package Repository......Page 417 Installing PMDK Libraries from the Package Repository......Page 419 Installing PMDK on Fedora 22 or Later......Page 420 Installing PMDK on SLES 12 and OpenSUSE or Later......Page 421 Installing PMDK on Ubuntu 18.04 or Later......Page 422 Installing PMDK on Microsoft Windows......Page 423 Appendix C: How to Install IPMCTL on Linux and Windows......Page 424 IPMCTL for Microsoft Windows......Page 425 Using ipmctl......Page 426 Volatile Use of Persistent Memory......Page 431 Heap Allocation on Alternative Memory Devices......Page 432 Partial Heap Allocation on Alternative Memory Devices......Page 434 Non-volatile Mapped Byte Buffers......Page 435 Persistent Collections for Java (PCJ)......Page 436 Using PCJ in Java Applications......Page 437 Low-Level Persistent Library (LLPL)......Page 438 Summary......Page 439 Appendix E: The Future of Remote Persistent Memory Replication......Page 441 Glossary......Page 445 Index......Page 448 Beginning and experienced programmers will use this comprehensive guide to persistent memory programming. You will understand how persistent memory brings together several new software/hardware requirements, and offers great promise for better performance and faster application startup times—a huge leap forward in byte-addressable capacity compared with current DRAM offerings. This revolutionary new technology gives applications significant performance and capacity improvements over existing technologies. It requires a new way of thinking and developing, which makes this highly disruptive to the IT/computing industry. The full spectrum of industry sectors that will benefit from this technology include, but are not limited to, in-memory and traditional databases, AI, analytics, HPC, virtualization, and big data. Programming Persistent Memory describes the technology and why it is exciting the industry. It covers the operating system and hardware requirements as well as how to create development environments using emulated or real persistent memory hardware. The book explains fundamental concepts; provides an introduction to persistent memory programming APIs for C, C++, JavaScript, and other languages; discusses RMDA with persistent memory; reviews security features; and presents many examples. Source code and examples that you can run on your own systems are included. What You'll Learn Understand what persistent memory is, what it does, and the value it brings to the industry Become familiar with the operating system and hardware requirements to use persistent memory Know the fundamentals of persistent memory programming: why it is different from current programming methods, and what developers need to keep in mind when programming for persistence Look at persistent memory application development by example using the Persistent Memory Development Kit (PMDK) Design and optimize data structures for persistent memory Study how real-world applications are modified to leverage persistent memory Utilize the tools available for persistent memory programming, application performance profiling, and debugging Who This Book Is For C, C++, Java, and Python developers, but will also be useful to software, cloud, and hardware architects across a broad spectrum of sectors, including cloud service providers, independent software vendors, high performance compute, artificial intelligence, data analytics, big data, etc. Beginning and experienced programmers will use this comprehensive guide to persistent memory programming. You will understand how persistent memory brings together several new software/hardware requirements, and offers great promise for better performance and faster application startup times--a huge leap forward in byte-addressable capacity compared with current DRAM offerings. This revolutionary new technology gives applications significant performance and capacity improvements over existing technologies. It requires a new way of thinking and developing, which makes this highly disruptive to the IT/computing industry. The full spectrum of industry sectors that will benefit from this technology include, but are not limited to, in-memory and traditional databases, AI, analytics, HPC, virtualization, and big data. Programming Persistent Memory describes the technology and why it is exciting the industry. It covers the operating system and hardware requirements as well as how to create development environments using emulated or real persistent memory hardware. The book explains fundamental concepts; provides an introduction to persistent memory programming APIs for C, C++, JavaScript, and other languages; discusses RMDA with persistent memory; reviews security features; and presents many examples. Source code and examples that you can run on your own systems are included. You will: Understand what persistent memory is, what it does, and the value it brings to the industry Become familiar with the operating system and hardware requirements to use persistent memory Know the fundamentals of persistent memory programming: why it is different from current programming methods, and what developers need to keep in mind when programming for persistence Look at persistent memory application development by example using the Persistent Memory Development Kit (PMDK) Design and optimize data structures for persistent memory Study how real-world applications are modified to leverage persistent memory Utilize the tools available for persistent memory programming, application performance profiling, and debugging This open access book is a modern guide for all C++ programmers to learn Threading Building Blocks (TBB). Written by TBB and parallel programming experts, this book reflects their collective decades of experience in developing and teaching parallel programming with TBB, offering their insights in an approachable manner. Throughout the book the authors present numerous examples and best practices to help you become an effective TBB programmer and leverage the power of parallel systems. Pro TBB starts with the basics, explaining parallel algorithms and C++'s built-in standard template library for parallelism. You'll learn the key concepts of managing memory, working with data structures and how to handle typical issues with synchronization. Later chapters apply these ideas to complex systems to explain performance tradeoffs, mapping common parallel patterns, controlling threads and overhead, and extending TBB to program heterogeneous systems or system-on-chips. What You'll LearnUse Threading Building Blocks to produce code that is portable, simple, scalable, and more understandableReview best practices for parallelizing computationally intensive tasks in your applicationsIntegrate TBB with other threading packagesCreate scalable, high performance data-parallel programsWork with generic programming to write efficient algorithmsWho This Book Is ForC++ programmers learning to run applications on multicore systems, as well as C or C++ programmers without much experience with templates. No previous experience with parallel programming or multicore processors is required.

Beginning and experienced programmers will use this comprehensive guide to persistent memory programming. You will understand how persistent memory brings together several new software/hardware requirements, and offers great promise for better performance and faster application startup times—a huge leap forward in byte-addressable capacity compared with current DRAM offerings.This revolutionary new technology gives applications significant performance and capacity improvements over existing technologies. It requires a new way of thinking and developing, which makes this highly disruptive to the IT/computing industry. The full spectrum of industry sectors that will benefit from this technology include, but are not limited to, in-memory and traditional databases, AI, analytics, HPC, virtualization, and big data. Programming Persistent Memory describes the technology and why it is exciting the industry. It covers the operating system and hardware requirements as well as how to create development environments using emulated or real persistent memory hardware. The book explains fundamental concepts; provides an introduction to persistent memory programming APIs for C, C++, JavaScript, and other languages; discusses RMDA with persistent memory; reviews security features; and presents many examples. Source code and examples that you can run on your own systems are included.

This open access book is a modern guide for all C++ programmers to learn Intel Threading Building Blocks (TBB). Written by TBB and parallel programming experts, this book reflects their collective decades of experience in developing and teaching parallel programming with TBB, offering their insights in an approachable manner. Throughout the book the authors present numerous examples and best practices to help you become an effective TBB programmer and leverage the power of parallel systems. Pro Intel Threading Building Blocks starts with the basics, explaining parallel algorithms and C++'s built-in standard template library for parallelism. You'll learn the key concepts of managing memory, working with data structures and how to handle typical issues with synchronization. Later chapters apply these ideas to complex systems to explain performance tradeoffs, mapping common parallel patterns, controlling threads and overhead, and extending TBB to program heterogeneous systems or system-on-chips.