Volatile Use of Persistent Memory

Introduction

In this article, we discuss how applications that require a large quantity of volatile memory can leverage high-capacity persistent memory as a complementary solution to dynamic random-access memory (DRAM).

Note: The article is excerpted from the book Programming Persistent Memory: A Comprehensive Guide for Developers, soon to be published by Apress*. Visit the book preview page at pmem.io for more information and a preview of some of the other chapters.

Applications that work with large datasets, like in-memory databases, caching systems, and scientific simulations, are often limited by the amount of volatile memory capacity available in the system or the cost of the DRAM required to load a complete data set. Persistent memory offers a tradeoff between price and performance.

In the memory-storage hierarchy, data is stored in tiers with frequently accessed data placed in DRAM for low latency access, and less frequently accessed data is placed in larger capacity, higher latency storage devices. Examples of such solutions include Redis* on Flash  and Extstore for Memcached*.

Compared with DRAM, persistent memory is relatively inexpensive and offers much higher capacity. Using these large capacities as volatile memory provides a new opportunity for memory-hungry applications that don’t require persistence.

Using persistent memory as a volatile memory solution is advantageous when an application:

• Has control over data placement between DRAM and other storage tiers within the system
• Does not need to persist data
• Can use the native latencies of persistent memory, which may be slower than DRAM but are faster than non-volatile memory express (NVMe)_ NAND solid-state drives (SSDs).

Background

Applications manage different kinds of data structures such as user data, key-value stores, metadata, and working buffers. Architecting a solution that uses tiered memory and storage enhances application performance; for example, placing objects that are accessed frequently and require fast access in DRAM, while storing data that requires larger allocations and is not latency-sensitive on persistent memory in use as volatile memory.

Memory Allocation

Persistent memory is exposed to the application using memory-mapped files on a persistent memory-aware file system that provides direct access to the application. Since malloc() and free() do not operate on files, an interface is needed that provides malloc() and free() semantics through an API for memory-mapped files as a source for memory allocation. This interface is implemented as the memkind library.

How it Works

The memkind library is a user-extensible heap manager built on top of jemalloc, which enables control of memory characteristics and partitioning of the heap between kinds of memory. It was originally created to support different kinds of memory with the introduction of high bandwidth memory (HBM). A PMEM kind was introduced to support persistent memory.

Different “kinds” of memory are defined by the operating system memory policies that were applied to virtual address ranges. Memory characteristics supported by memkind without user extension include control of non-uniform memory access (NUMA) and page size features. The jemalloc non-standard interface was extended so that specialized arenas could make requests for virtual memory from the operating system through the memkind partition interface. Through the other memkind interfaces, developers can control and extend memory partition features and allocate memory while selecting enabled features. Figure 1.1 shows an overview of libmemkind components and hardware support.

Figure 1.1. An overview of the memkind components and hardware support

The memkind library serves as a wrapper that redirects memory allocation requests from an application to an allocator that manages the heap. At the time of publication, only the jemalloc allocator is supported. Future versions may introduce and support multiple allocators. Memkind provides jemalloc with different sources of memory: A static kind is created automatically, whereas a dynamic kind is created by an application using the memkind_create_kind() API.

The PMEM kind, which is dynamic, is best used with memory-addressable persistent storage through a DAX-enabled file system that supports load/store operations without being paged via the system page cache. The PMEM kind supports the traditional malloc/free interfaces on a memory-mapped file. A temporary file is automatically created on a mounted DAX file system and memory-mapped into the application’s virtual address space. The temporary file is deleted when the program terminates, giving the perception of volatility.

Supported “Kinds” of Memory

When the kind of memory is PMEM_KIND, the memory allocation source is a memory-mapped file created as a temporary file on a persistent memory-aware file system.

For allocations from DRAM, the application sets the kind to MEMKIND_DEFAULT with the operating system’s default page size.  Refer to the memkind documentation for large and huge page support.

Figure 1.2. Application using different “kinds” of memory

Figure 1.2 shows memory mappings from two memory sources – DRAM (MEMKIND_DEFAULT) and persistent memory (PMEM_KIND).

When using PMEM_KIND, the key points to understand are:

• Two pools of memory are available to the application from DRAM and persistent memory. Both can be accessed simultaneously by setting the kind type to PMEM_KIND and MEMKIND_DEFAULT.
• Jemalloc is the single memory allocator used to manage all kinds of memory.
• Memkind is a wrapper around jemalloc that provides a unified API for allocations from different kinds of memory.
• Memory allocations are provided by a temporary file created on a persistent memory-aware file system. The file is destroyed when the application exits.
• Allocations are not persistent
• Using libmemkind for persistent memory requires simple modifications to the application.

The Memkind API

The memkind API functions related to persistent memory programming are shown in Listing 1.1 and described in this section. The complete memkind API is available in the memkind man pages.

Listing 1.1. Persistent memory related memkind API functions

KIND CREATION MANAGEMENT: 

int memkind_create_pmem(const char *dir, size_t max_size, memkind_t *kind); 
int memkind_create_pmem_with_config(struct memkind_config *cfg, memkind_t *kind);
memkind_t memkind_detect_kind(void *ptr);
int memkind_destroy_kind(memkind_t kind);

KIND HEAP MANAGEMENT:

void *memkind_malloc(memkind_t kind, size_t size); 
void *memkind_calloc(memkind_t kind, size_t num, size_t size); 
void *memkind_realloc(memkind_t kind, void *ptr, size_t size); 
void memkind_free(memkind_t kind, void *ptr); 
size_t memkind_malloc_usable_size(memkind_t kind, void *ptr); 
memkind_t memkind_detect_kind(void *ptr);

KIND CONFIGURATION MANAGEMENT:

struct memkind_config *memkind_config_new(); 
void memkind_config_delete(struct memkind_config *cfg); 
void memkind_config_set_path(struct memkind_config *cfg, const char *pmem_dir); 
void memkind_config_set_size(struct memkind_config *cfg, size_t pmem_size); 
void memkind_config_set_memory_usage_policy(struct memkind_config *cfg, memkind_mem_usage_policy policy);

Kind Management API

The memkind library supports a plugin architecture to incorporate new memory kinds, which are referred to as dynamic kinds. The memkind library provides the API to create and manage the heap for the new kind.

Kind Creation

Use the memkind_create_pmem() function to create a PMEM kind of memory from a file-backed source. This file is created as a tmpfile(3) in a specified directory and is unlinked so the filename is not listed under the directory and is automatically removed when the program terminates.

Use memkind_create_pmem() to create a fixed or dynamic heap size depending on the application requirement. Additionally, configurations can be created and supplied rather than passing in configuration options to the *_create_* function.

Creating a Fixed Size Heap

Applications that require a fixed amount of memory can specify a non-zero value for the PMEM_MAX_SIZE argument to memkind_create_pmem(). This defines the size of the memory pool to be created for the specified kind of memory. The value of PMEM_MAX_SIZE should be less than the available capacity of the file system specified in PMEM_DIR to avoid ENOMEM or ENOSPC errors. An internal data structure struct memkind is populated internally by the library and used by the memory management functions.

int memkind_create_pmem(PMEM_DIR, PMEM_MAX_SIZE, &pmem_kind)

The arguments to memkind_create_pmem() are:

• PMEM_DIR is the directory where the temp file is created
• PMEM_MAX_SIZE is the size of the memory region to be passed to jemalloc
• &pmem_kind is the address of a memkind data structure.

If successful, memkind_create_pmem() returns a value of zero. On failure, an error number is returned that memkind_error_message() can convert to an error message. Listing 1.1 shows how a 32 MiB PMEM kind is created on a /pmemfs file system. 

Listing 1.2. Creating a 32 MiB PMEM kind

#define PMEM_MAX_SIZE (1024 * 1024 * 32)

struct memkind *pmem_kind = NULL;
int err = 0;

// Create first PMEM partition with specific size
err = memkind_create_pmem("/pmemfs/", PMEM_MAX_SIZE, &pmem_kind);
if (err) {
    print_err_message(err);
    return 1;
}

You can also create a heap with a specific configuration using the function memkind_create_pmem_with_config(). This function requires completing a memkind_config structure with optional parameters such as size, path to file, and memory usage policy. Listing 1.3 shows how to build a test_cfg using meknind_config_new(), then passing that configuration to memkind_create_pmem_with_config() to create a PMEM kind. We use the same path and size parameters from the Listing 1.1 example for comparison.

Listing 1.3. Creating PMEM kind with configuration

struct memkind_config *test_cfg = memkind_config_new();
memkind_config_set_path(test_cfg, “/pmemfs/”);
memkind_config_set_size(test_cfg, 1024 * 1024 * 32);
memkind_config_set_memory_usage_policy(test_cfg, MEMKIND_MEM_USAGE_POLICY_CONSERVATIVE);

// create FPMEM partition with specific configuration
err = memkind_create_pmem_with_config(test_cfg, &pmem_kind); 
if (err) {
    print_err_message(err);
    return 1;
}

Creating a Variable Size Heap

When PMEM_MAX_SIZE is set to zero, allocations are satisfied as long as the temporary file can grow. The maximum heap size growth is limited by the capacity of the file system mounted under the PMEM_DIR argument. 

memkind_create_pmem(PMEM_DIR, 0, &pmem_kind)

The arguments to memkind_create_pmem() are:

• PMEM_DIR is the directory where the temp file is created
• PMEM_MAX_SIZE is 0
• &pmem_kind is the address of a memkind data structure

If the PMEM kind is created successfully, memkind_create_pmem() returns zero. On failure, memkind_error_message() can be used to convert an error number returned by memkind_create_pmem() to an error message.
 

Listing 1.4. shows how to create a PMEM kind with variable size.

Listing 1.4. Creating a PMEM kind with variable size

struct memkind *pmem_kind = NULL;
int err = 0;
err = memkind_create_pmem(“/pmemfs/”,0,&pmem_kind);
if (err) {
    print_err_message(err);
    return 1;
}

Detecting the Memory Kind

Memkind supports both automatic detection of a kind as well as a function to detect a kind associated with a memory referenced by a pointer. 

Automatic Kind Detection

Support for automatically detecting the kind of memory was added to simplify code changes when adopting libmemkind. Thus, the memkind library will automatically retrieve the kind of memory pool where the allocation was done so that the heap management functions listed in Table 1.1 can be called without specifying the kind.

Table 1.1. Automatic kind detection functions and their equivalent specified kind functions and operations

Operation	Memkind API with Kind	Memkind API using automatic detection
free	memkind_free(kind, ptr)	memkind_free(NULL, ptr)
realloc	memkind_realloc(kind, ptr, size)	memkind_realloc(NULL, ptr, size)
Get size of allocated memory	memkind_malloc_usable_size(kind, ptr)	memkind_malloc_usable_size(NULL, ptr)

The memkind library internally tracks the kind of a given object from the allocator metadata. However, to get this information some of the operations may need to acquire a lock to prevent accesses from other threads, which may negatively affect the performance in a multithreaded environment.

Memory Kind Detection API

Memkind also provides the memkind_detect_kind() function to query and return the kind of memory associated with the memory referenced by the pointer passed into the function. If the input pointer argument is NULL, it returns NULL. The input pointer argument that gets passed into memkind_detect_kind() must have been returned by a previous call to memkind_malloc(), memkind_calloc(), memkind_realloc() or memkind_posix_memalign().

memkind_t memkind_detect_kind(void *ptr)

Similar to the automatic detection approach, this function has non-trivial performance overhead.

Listing 1.5. pmem_kind_detection.c – How to automatically detect the 'kind' type

37  #include <memkind.h>
38
39  #include <limits.h>
40  #include <stdio.h>
41  #include <stdlib.h>
42
43  static char path[PATH_MAX]="/pmemfs/";
44
45  #define MALLOC_SIZE 512U
46  #define REALLOC_SIZE 2048U
47  #define ALLOC_LIMIT  1000U
48
49  static void *alloc_buffer[ALLOC_LIMIT];
50
51  static void print_err_message(int err)
52  {
53      char error_message[MEMKIND_ERROR_MESSAGE_SIZE];
54      memkind_error_message(err, error_message, MEMKIND_ERROR_MESSAGE_SIZE);
55      fprintf(stderr, "%s\n", error_message);
56  }
57
58  static int allocate_pmem_and_default_kind(struct memkind *pmem_kind)
59  {
60      unsigned i;
61      for(i = 0; i < ALLOC_LIMIT; i++) {
62          if (i%2)
63              alloc_buffer[i] = memkind_malloc(pmem_kind, MALLOC_SIZE);
64          else
65              alloc_buffer[i] = memkind_malloc(MEMKIND_DEFAULT, MALLOC_SIZE);
66
67          if (!alloc_buffer[i]) {
68              return 1;
69          }
70      }
71
72      return 0;
73  }
74
75  static int realloc_using_get_kind_only_on_pmem()
76  {
77      unsigned i;
78      for(i = 0; i < ALLOC_LIMIT; i++) {
79          if (memkind_detect_kind(alloc_buffer[i]) != MEMKIND_DEFAULT) {
80              void *temp = memkind_realloc(NULL, alloc_buffer[i], REALLOC_SIZE);
81              if (!temp) {
82                  return 1;
83              }
84              alloc_buffer[i] = temp;
85          }
86      }
87
88      return 0;
89  }
90
91
92  static int verify_allocation_size(struct memkind *pmem_kind, size_t pmem_size)
93  {
94      unsigned i;
95      for(i = 0; i < ALLOC_LIMIT; i++) {
96          void *val = alloc_buffer[i];
97          if (i%2) {
98              if (memkind_malloc_usable_size(pmem_kind, val) != pmem_size ) {
99                  return 1;
100              }
101          } else {
102              if (memkind_malloc_usable_size(MEMKIND_DEFAULT, val) != MALLOC_SIZE) {
103                  return 1;
104              }
105          }
106      }
107
108      return 0;
109  }
110
111  int main(int argc, char *argv[])
112  {
113
114      struct memkind *pmem_kind = NULL;
115      int err = 0;
116
117      if (argc > 2) {
118          fprintf(stderr, "Usage: %s [pmem_kind_dir_path]\n", argv[0]);
119          return 1;
120      } else if (argc == 2 && (realpath(argv[1], path) == NULL)) {
121          fprintf(stderr, "Incorrect pmem_kind_dir_path %s\n", argv[1]);
122          return 1;
123      }
124
125      fprintf(stdout,
126              "This example shows how to distinguish allocation from different kinds using detect kind function"
127              "\nPMEM kind directory: %s\n", path);
128
129      err = memkind_create_pmem(path, 0, &pmem_kind);
130      if (err) {
131          print_err_message(err);
132          return 1;
133      }
134
135      fprintf(stdout, "Allocate to PMEM and DEFAULT kind.\n");
136
137      if (allocate_pmem_and_default_kind(pmem_kind)) {
138          fprintf(stderr, "allocate_pmem_and_default_kind().\n");
139          return 1;
140      }
141
142      if (verify_allocation_size(pmem_kind, MALLOC_SIZE)) {
143          fprintf(stderr, "verify_allocation_size() before resize.\n");
144          return 1;
145      }
146
147      fprintf(stdout,
148              "Reallocate memory only on PMEM kind using memkind_detect_kind().\n");
149
150      if (realloc_using_get_kind_only_on_pmem()) {
151          fprintf(stderr, "realloc_using_get_kind_only_on_pmem().\n");
152          return 1;
153      }
154
155      if (verify_allocation_size(pmem_kind, REALLOC_SIZE)) {
156          fprintf(stderr, "verify_allocation_size() after resize.\n");
157          return 1;
158      }
...

Destroying Kind Objects

Use the memkind_destroy_kind() function to delete the kind object that was previously created using the memkind_create_pmem() or memkind_create_pmem_with_config() function. The memkind_destroy_kind() is defined as:

int memkind_destroy_kind(memkind_t kind);

Using the same pmem_detect_kind.c code from Listing 1.5, Listing 1.6 shows how the kind is destroyed before the program exits.

Listing 1.6. Destroying a kind object

160      err = memkind_destroy_kind(pmem_kind);
161      if (err) {
162          print_err_message(err);
163          return 1;
164      }
...
168      return 0;
169  }

When the kind returned by memkind_create_pmem() or memkind_create_pmem_with_config() is successfully destroyed, all the allocated memory for the kind object is freed.

Heap Management API

The heap management functions described in this section have an interface modeled on the ISO C standard API, with an additional “kind” parameter to specify the memory type used for allocation.

Allocating Memory

The memkind library provides memkind_malloc(), memkind_calloc() and memkind_realloc() functions for allocating memory, defined as follows:

void *memkind_malloc(memkind_t kind, size_t size); 
void *memkind_calloc(memkind_t kind, size_t num, size_t size); 
void *memkind_realloc(memkind_t kind, void *ptr, size_t size);

memkind_malloc() allocates size bytes of uninitialized memory of the specified kind. The allocated space is suitably aligned (after possible pointer coercion) for storage of any object type. If size is 0, then memkind_malloc() returns NULL.

memkind_calloc() allocates space for num objects, each are size bytes in length. The result is identical to calling memkind_malloc() with an argument of num * size. The exception is that the allocated memory is explicitly initialized to zero bytes. If num or size is 0, then memkind_calloc() returns NULL.

memkind_realloc() changes the size of the previously allocated memory referenced by ptr to size bytes of the specified kind. The contents of the memory remain unchanged, up to the lesser of the new and old sizes. If the new size is larger, the contents of the newly allocated portion of the memory are undefined. If successful, the memory referenced by ptr is freed and a pointer to the newly allocated memory is returned.

The examples in Listing 1.7 show how to allocate memory from DRAM and persistent memory (pmem_kind) using memkind_malloc(). Rather than using the common C library malloc() for DRAM memory and memkind_malloc() for persistent memory, we recommend using a single library to simplify code.

Listing 1.7. An example of allocating memory from both DRAM and persistent memory.

/* 
 * Allocates 100 bytes using appropriate "kind" 
 * of volatile memory 
 */

// Create first PMEM partition with a specific size
  err = memkind_create_pmem(path, PMEM_MAX_SIZE, &pmem_kind);
  if (err) {
      print_err_message(err);
      return 1;
  } 
  char *pstring = memkind_malloc(pmem_kind, 100);
  char *dstring = memkind_malloc(MEMKIND_DEFAULT, 100);

Freeing Allocated Memory

To avoid memory leaks, allocated memory can be freed using the memkind_free() function, defined as:

void memkind_free(memkind_t kind, void *ptr);

memkind_free() causes the allocated memory referenced by ptr to be made available for future allocations. This pointer must be returned by a previous call to memkind_malloc(), memkind_calloc(), memkind_realloc() or memkind_posix_memalign().  Otherwise, if memkind_free(kind, ptr) was previously called, undefined behavior occurs. If ptr is NULL, no operation is performed. In cases where the kind is unknown in the context of the call to memkind_free(), NULL can be given as the kind specified to memkind_free(), but this will require an internal lookup for the correct kind. Always specify the correct kind because the lookup for kind could result in serious performance penalty.

Listing 1.8 shows four examples of memkind_free() being used. The first two specify the kind, and the second two use NULL.

Listing 1.8. Examples of memkind_free() usage.

/* Free the memory by specifying the ‘kind’ */
memkind_free(MEMKIND_DEFAULT, dstring);
memkind_free(PMEM_KIND, pstring);

/* Free the memory using automatic ‘kind’ detection */
memkind_free(NULL, dstring);
memkind_free(NULL, pstring);

Kind Configuration Management

Memory Usage Policy

A tunable run time option set by the dirty_decay_ms in jemalloc determines how fast it returns unused memory back to the operating system. A shorter decay time purges unused memory pages faster but the purging costs CPU cycles. Trade-offs between memory and CPU cycles needs to be careful thought out before setting this parameter.

A new implementation was introduced in memkind release v1.9 to improve memory utilization and reduce fragmentation. The first implementation supports two policies:

1. MEMKIND_MEM_USAGE_POLICY_DEFAULT
2. MEMKIND_MEM_USAGE_POLICY_CONSERVATIVE

The minimum and maximum values for dirty_decay_ms using the MEMKIND_MEM_USAGE_POLICY_DEFAULT are 0 ms to 10,000 ms for arenas assigned to a PMEM kind. Setting MEMKIND_MEM_USAGE_POLICY_CONSERVATIVE sets shorter decay times to purge unused memory faster, resulting in reducing memory usage. To define the memory usage policy, use memkind_config_set_memory_usage_policy(), defined below:

void memkind_config_set_memory_usage_policy (struct memkind_config *cfg, memkind_mem_usage_policy policy );

MEMKIND_MEM_USAGE_POLICY_DEFAULT is the default memory usage policy.

MEMKIND_MEM_USAGE_POLICY_CONSERVATIVE allows changing the dirty_decay_ms parameter.

Range of dirty_decay_ms is 10,000 ms to 0 ms for arenas assigned to PMEM kind.

Listing 1.9 shows how to use memkind_config_set_memory_usage_policy() with a custom configuration.

Listing 1.9. An example of a custom configuration and memory policy use

33  /*
34   * pmem_config.c - Demonstrates the use of several configuration
35   *                 functions within libmemkind.
36   */
37
38  #include <memkind.h>
39
40  #include <limits.h>
41  #include <stdio.h>
42  #include <stdlib.h>
43
44  #define PMEM_MAX_SIZE (1024 * 1024 * 32)
45
46  static char path[PATH_MAX] = "pmemfs//";
..
55  int main(int argc, char *argv[])
56  {
57      struct memkind *pmem_kind = NULL;
58      int err = 0;
59
60      if (argc > 2) {
61          fprintf(stderr, "Usage: %s [pmem_kind_dir_path]\n", argv[0]);
62          return 1;
63      } else if (argc == 2 && (realpath(argv[1], path) == NULL)) {
64          fprintf(stderr, "Incorrect pmem_kind_dir_path %s\n", argv[1]);
65          return 1;
66      }
67
68      fprintf(stdout,
69              "This example shows how to use custom configuration to create pmem kind."
70              "\nPMEM kind directory: %s\n", path);
71
72      struct memkind_config *test_cfg = memkind_config_new();
73      if (!test_cfg) {
74          fprintf(stderr, "Unable to create memkind cfg.\n");
75          return 1;
76      }
77
78      memkind_config_set_path(test_cfg, path);
79      memkind_config_set_size(test_cfg, PMEM_MAX_SIZE);
80      memkind_config_set_memory_usage_policy(test_cfg,
81                MEMKIND_MEM_USAGE_POLICY_CONSERVATIVE);
82
83
84      // Create PMEM partition with specific configuration
85      err =  memkind_create_pmem_with_config(test_cfg, &pmem_kind);
86      if (err) {
87          print_err_message(err);
88          return 1;
89      }
90
91      err = memkind_destroy_kind(pmem_kind);
92      if (err) {
93          print_err_message(err);
94          return 1;
95      }
96
97      memkind_config_delete(test_cfg);
98
99      fprintf(stdout,
100              "PMEM kind and configuration was successfully created and destroyed.\n");
101
102      return 0;
103  }

C++ Allocator for PMEM Kind

To enable C++ developers to allocate from a PMEM kind of memory, the pmem::allocator class template, which conforms to C++11 allocator requirements, was developed. It  can be used with C++ compliant data structures from:

• Standard Template Library (STL)
• Intel® Threading Building Blocks (Intel® TBB) library

The pmem::allocator class template uses the memkind_create_pmem() function described previously. This allocator is stateful and has no default constructor. Table 1.3 describes the available allocator methods.

Table 1.3. pmem::allocator methods

pmem::allocator(const char *dir, size_t max_size)

pmem::allocator(const std::string& dir, size_t max_size)

template <typename U> pmem::allocator<T>::allocator(const pmem::allocator<U>&)

template <typename U> pmem::allocator(allocator<U>&& other)

pmem::allocator<T>::~allocator()

T* pmem::allocator<T>::allocate(std::size_t n) const

void pmem::allocator<T>::deallocate(T* p, std::size_t n) const

template <class U, class... Args> void pmem::allocator<T>::construct(U* p, Args... args) const

void pmem::allocator<T>::destroy(T* p) const

For more information about the pmem::allocator class template, refer to the pmem allocator(3) man page.

Nested Containers

Challenges occur while working with multilevel containers such as a vector of sets of lists, tuples, maps, strings, and so on. When the outermost container is constructed, an instance of pmem::allocator is passed as a parameter to the constructor. How should you handle nested objects stored in the outermost container?

Imagine you need to create a vector of strings and store it in persistent memory. The challenges—and their solutions—for this task include:

1. You cannot use std::string for this purpose because it is an alias of the std::basic_string. The std::allocator requires a new alias that uses pmem:allocator.
   
   Solution: A new alias called pmem_string is defined as a typedef of std::basic_string when created with pmem::allocator.

2. How to ensure that an outermost vector will properly construct nested pmem_string with a proper instance of pmem::allocator.
   
   Solution: From C++11 and later, the std::scoped_allocator_adaptor class template can be used with multilevel containers. The purpose of this adaptor is to correctly initialize stateful allocators in nested containers, such as when all levels of a nested container must be placed in the same memory segment.

C++ Examples

This section presents several full-code examples demonstrating the use of libmemkind using C and C++. 

Using the pmem::allocator

As mentioned earlier, you can use pmem::allocator with any STL-like data structure. The code sample in Listing 1.10 includes a pmem_allocator.h header file to use pmem::allocator.

Listing 1.10. Using pmem::allocator with std:vector

33  /*
34   * pmem_allocator.cpp - Demonstrates using the pmem::allocator
35   *                 with std:vector.
36   */
37
38  #include <pmem_allocator.h>
39  #include <vector>
40  #include <cassert>
41
42  int main(int argc, char *argv[]) {
43          const size_t pmem_max_size = 64*1024*1024; // 64 MB
44          const std::string pmem_dir("/pmemfs/");
45
46          // Create allocator object
47          pmem::allocator<int> alc(pmem_dir, pmem_max_size);
48          // Create std::vector with our allocator.
49          std::vector<int, pmem::allocator<int> > v(alc);
50
51          for(int i = 0; i < 100; ++i)
52          v.push_back(i);
53
54          for(int i = 0; i < 100; ++i)
55          assert(v[i] == i);
56
57          return 0;
58  }

• Line 43: We define a persistent memory mapping of 64 MiB.
• Line 47: We create an allocator object alc of type pmem::allocator<int>.
• Line 49: We create a vector object v of type std::vector<int, pmem::allocator<int> > and pass in the alc from line 47 object as an argument. The pmem::allocator is stateful and has no default constructor. This requires passing the allocator object to the vector constructor; otherwise, a compilation error occurs if the default constructor of std::vector<int, pmem::allocator<int> > is called because the vector constructor will try to call the default constructor of pmem::allocator, which does not exist yet.

Creating a Vector of Strings

Listing 1.11 shows how to create a vector of strings that resides in persistent memory. We define pmem_string as a typedef of std::basic_string with pmem::allocator. In this example, std::scoped_allocator_adaptor allows the vector to propagate the pmem::allocator instance to all pmem_string objects stored in the vector object.

Listing 1.11. Creating a vector of strings

33  /*
34   * vector_of_strings.cpp - Demonstrated how to create a vector
35   *              of strings residing on persistent memory.
36   */
37
38  #include <pmem_allocator.h>
39  #include <vector>
40  #include <string>
41  #include <scoped_allocator>
42  #include <cassert>
43
44  typedef pmem::allocator<char> str_alloc_type;
45
46  typedef std::basic_string<char, std::char_traits<char>, str_alloc_type> pmem_string;
47
48  typedef pmem::allocator<pmem_string> vec_alloc_type;
49
50  typedef std::vector<pmem_string, std::scoped_allocator_adaptor<vec_alloc_type> > vector_type;
51
52  int main(int argc, char *argv[]) {
53          const size_t pmem_max_size = 64*1024*1024; // 64 MB
54          const std::string pmem_dir("/tmp");
55
56          // Create allocator object
57          vec_alloc_type alc(pmem_dir, pmem_max_size);
58          // Create std::vector with our allocator.
59          vector_type v(alc);
60
61          v.emplace_back(“Foo”);
62          v.emplace_back(“Bar”);
63
64          for(auto str : v) {
65          std::cout << str << std::endl;
66          }
67
68          return 0;
69  }

• Line 46: We define pmem_string as a typedef of std::basic_string.
• Line 48: We define the pmem::allocator using the pmem_string type. 
• Line 50: Using std::scoped_allocator_adaptor allows the vector to propagate the pmem::allocator instance to all pmem_string objects stored in the vector object.

See more examples in the memkind examples directory on GitHub.

More Memkind Code Examples

Table 1.2 lists the code examples available on GitHub*.

Table 1.2. Source code examples using libmemkind

File Name	Description
pmem_kinds.c	Creating and destroying PMEM kind with defined or unlimited size.
pmem_malloc.c	Allocating memory and the possibility to exceed PMEM kind size.
pmem_malloc_unlimited.c	Allocating memory with unlimited kind size.
pmem_usable_size.c	Viewing the difference between the expected and the actual allocation size.
pmem_alignment.c	Using memkind alignment and how it affects allocations.
pmem_multithreads.c	Using multithreading with independent PMEM kinds.
pmem_multithreads_onekind.c	Using multithreading with one main PMEM kind.
pmem_and_default_kind.c	Allocating in standard memory and file-backed memory (PMEM kind).
pmem_detect_kind.c:	Distinguishing allocation from different kinds using the detect kind function.
pmem_config.c	Using custom configuration to create PMEM kind.
pmem_free_with_unknown_kind.c	Allocating in-standard memory, file-backed memory (PMEM kind), and free memory without needing to remember which kind it belongs to.
pmem_cpp_allocator.cpp	Shows usage of C++ allocator mechanism designed for file-backed memory kind with different data structures like vector, list, and map.

libvmemcache: An Efficient Volatile Key-Value Cache for Large-Capacity Persistent Memory

Some existing in-memory databases (IMDB) rely on manual dynamic memory allocations (malloc, jemalloc, tcmalloc), which can exhibit memory fragmentation (external and internal) when run for a long period leaving large amounts of memory un-allocatable. Internal and external fragmentation is briefly explained as follows:

• Internal fragmentation occurs when more than the needed memory is allocated, and the unused memory is contained within the allocated region. For example, if the requested allocation size is 200 bytes, a chunk of 256 bytes is allocated.
   
• External fragmentation occurs when variable memory sizes are allocated dynamically, resulting in a failure to allocate a contiguous chunk of memory, although the requested chunk of memory remains available in the system. This problem is more pronounced when large capacities of persistent memory are being used as volatile memory. Applications with substantially long runtimes need to resolve this problem, especially if the allocated sizes have considerable variation.  Applications and runtime environments handle this problem in different ways:
  • Java* and .NET use compacting garbage collection
  • Redis and Apache Ignite* use defragmentation algorithms
  • Memcached uses a slab allocator

Each of the above allocator mechanisms has pros and cons. Garbage and defragmentation algorithms require processing to occur on the heap to free unused allocations or move data to create contiguous space. Slab allocators usually define a fixed set of different sized buckets at initialization without knowing how many of each bucket the application will need. If the slab allocator depletes a certain bucket size, it allocates from larger sized buckets, which reduces the amount of free space. These three mechanisms can potentially block the application’s processing and reduce its performance.

libvmemcache Overview

libvmemcache is an embeddable and lightweight in-memory caching solution with a key-value store at its core. It is designed to take full advantage of large-capacity memory, such as persistent memory, efficiently using memory mapping in a scalable way. It is optimized for use with memory-addressable persistent storage through a DAX-enabled file system that supports load/store operations. libvmemcache has these unique characteristics:

• The extent-based memory allocator sidesteps the fragmentation problem that affects most in-memory databases, and it allows the cache to achieve very high space utilization for most workloads.
• Buffered LRU (least recently used) combines a traditional LRU doubly linked list with a non-blocking ring buffer to deliver high scalability on modern multi-core CPUs.
• A unique indexing critnib data structure delivers high performance and is very space-efficient.

The cache for libvmemcache is tuned to work optimally with relatively large value sizes. While the smallest possible size is 256 bytes, libvmemcache performs best if the expected value sizes are above 1 kilobyte.

libvmemcache has more control over the allocation because it implements a custom memory-allocation scheme using an extents-based approach (like that of file system extents). libvmemcache can, therefore, concatenate and achieve substantial space efficiency. Additionally, because it is a cache, it can evict data to allocate new entries in a worst-case scenario.  libvmemcache will always allocate exactly as much memory as it freed, minus metadata overhead. This is not true for caches based on common memory allocators such as memkind. libvmemcache is designed to work with terabyte-sized in-memory workloads, with very high space utilization.

Libvmemcache works by automatically creating a temporary file on a DAX-enabled file system and memory-mapping it into the application’s virtual address space. The temporary file is deleted when the program terminates and gives the perception of volatility. Figure 1.3 shows the application using traditional malloc() to allocate memory from DRAM and using libvmemcache to memory map a temporary file residing on a DAX-enabled file system from persistent memory.

Figure 1.3. An application using libvmemcache memory maps a temporary file from a DAX-enabled file system.

Although libmemkind supports different kinds of memory and memory consumption policies, the underlying allocator is jemalloc, which uses dynamic memory allocation. Table 1.4 compares the implementation details of libvmemcache and libmemkind.

Table 1.4. Design aspects of libmemkind and libvmemcache

	libmemkind (PMEM)	libvmemcache
Allocation Scheme	Dynamic allocator	Extent based (not restricted to sector, page, etc.)
Purpose	General purpose	Lightweight in-memory cache
Fragmentation	Apps with multiple size allocations/deallocations that run for a long period	Minimized

libvmemcache Design

libvmemcache has two main design aspects: 

1. Allocator design to improve/resolve fragmentation issues
2. A scalable and efficient LRU policy

Extent-Based Allocator

libvmemcache can solve fragmentation issues when working with terabyte-sized in-memory workloads and provide high space utilization. Figure 1.4 shows a workload example that creates many small objects, and over time, the allocator stops due to fragmentation.

Figure 1.4. An example of a workload that creates many small objects, and the allocator stops due to fragmentation.

libvmemcache uses an extent-based allocator, where extent is a contiguous set of blocks allocated for storing the data in a database. Extents are typically used with large blocks supported by file systems (sectors, pages, and so on), but such restrictions do not apply when working with persistent memory that supports smaller block sizes (cache-line). Figure 1.5 shows that if a single contiguous free block is not available to allocate an object, multiple, non-contiguous blocks are used to satisfy the allocation request. The non-contiguous allocations appear as a single allocation to the application.

Figure 1.5. Using non-contiguous free blocks to fulfill a larger allocation request

Scalable Replacement Policy

An LRU cache is traditionally implemented as a doubly-linked list. When an item is retrieved from this list, it gets moved from the middle to the front of the list so it is not evicted. In a multithreaded environment, multiple threads may contend with the front element, all trying to move elements being retrieved to the front element. Therefore, the front element is always locked (along with other locks) before moving the element being retrieved, which results in a few round trips into the kernel. This method is not scalable and is inefficient.

A buffer-based LRU policy creates a scalable and efficient replacement policy. A non-blocking ring buffer is placed in front of the LRU linked list to track the elements being retrieved. When an element is retrieved, it is added to this buffer, and only when the buffer is full (or the element is being evicted), the linked-list is locked and the elements in that buffer are processed and moved to the front of the list. This method preserves the LRU policy and provides a scalable LRU mechanism with minimal performance impact. Figure 1.6 shows a ring buffer-based design for the LRU algorithm.

Figure 1.6. A ring buffer-based LRU design

Using libvmemcache

Table 1.5 lists the basic functions that libvmemcache provides. For a complete list, see the libvmemcache man pages.

Table 1.5. The libvmemcache functions

Function Name	Description
vmemcache_new	Creates an empty unconfigured vmemcache instance with default values: Eviction_policy=VMEMCACHE_REPLACEMENT_LRU Extent_size = VMEMCAHE_MIN_EXTENT VMEMCACHE_MIN_POOL
vmemcache_add	Associates the cache with a path
vmemcache_set_size	Sets the size of the cache
vmemcache_set_extent_size	Sets the block size of the cache (256 bytes minimum)
vmemcache_set_eviction_policy	Sets the eviction policy: 1. VMEMCACHE_REPLACEMENT_NONE 2. VMEMCACHE_REPLACEMENT_LRU
vmemcache_add	Associates the cache with a given path on a DAX-enabled file system or non-DAX enabled file system
vmemcache_delete	Frees any structures associated with the cache
vmemcache_get	Searches for an entry with the given key and if found, the entry’s value is copied to vbuf
vmemcache_put	Inserts the given key:value pair into the cache
vmemcache_evict	Removes the given key from the cache
vmemcache_callback_on_evict	Called when an entry is being removed from the cache
vmemcache_callback_on_miss	Called when a get query fails to provide an opportunity to insert the missing key

To illustrate how libvmemcache is used, Listing 1.12 shows how to create an instance of vmemcache using default values. This example uses a temporary file on a DAX-enabled file system and shows how a callback is registered after a cache miss for a key “meow.”

Listing 1.12. An example program using libvmemcache

1  #include <libvmemcache.h>
2  #include <stdio.h>
3  #include <string.h>
4
5  #define STR_AND_LEN(x) (x), strlen(x)
6
7  static VMEMcache *cache;
8 
9  static void on_miss(VMEMcache *cache, const void *key, size_t key_size, void *arg)
10
11  {
12  	vmemcache_put(cache, STR_AND_LEN("meow"),
13  		STR_AND_LEN("Cthulhu fthagn"));
14  }
15
16  static void
17  get(const char *key)
18 {
19  	char buf[128];
20  	ssize_t len = vmemcache_get(cache, STR_AND_LEN(key),
21  		buf, sizeof(buf), 0, NULL);
22  	if (len >= 0)
23   		printf("%.*s\n", (int)len, buf);
24  	else
25   		printf("(key not found: %s)\n", key);
26  }
27
28  	int main()
29 	{
30  		cache = vmemcache_new();
31  		if (vmemcache_add(cache, "/pmemfs")) {
32  			fprintf(stderr, "error: vmemcache_add: %s\n",
33  				vmemcache_errormsg());
34  			return 1;
35  		}
36
37 		/* Query a non-existent key. */
38 		get("meow");
39
40 		/* Insert then query. */
41  		vmemcache_put(cache, STR_AND_LEN("bark"), STR_AND_LEN("Lorem ipsum"));
42  		get("bark");
43
44		/* Install an on-miss handler. */
45  		vmemcache_callback_on_miss(cache, on_miss, 0);
46  		get("meow");
47
48  		vmemcache_delete(cache);
49  		return 0;
50 }

• Line 30: Creates a new instance of vmemcache with default values for eviction_policy and extent_size.
• Line 31: Calls the vmemcache_add() function to associate cache with a given path.
• Line 38: Calls the get() function to query on an existing key. This function calls the vmemcache_get() function with error checking for success/failure of the function.
• Line 42: Calls vmemcache_put() to insert a new key.
• Line 45: Adds an on-miss callback handler to insert the key “meow” into the cache.
• Line 46: Retrieves the key “meow” using the get() function.
• Line 48: Deletes the vmemcache instance.

Expanding Volatile Memory Using Persistent Memory

Persistent memory is treated by the kernel as a device. In a typical usage, a persistent memory-aware file system is created, and files are memory-mapped into the virtual address space of a process to give applications direct load/store access to persistent memory regions.

A new feature was added to Linux* kernel v5.1 so that persistent memory can be used more broadly as RAM. This is done by binding a persistent memory device to the kernel, and the kernel manages it as DRAM. Since persistent memory has different characteristics than DRAM, memory provided by this device is visible as a separate NUMA node on its corresponding socket.

To programmatically allocate memory from a NUMA node created for persistent memory, a new static kind, called MEMKIND_DAX_KMEM, was added to libmemkind.

memkind_malloc(MEMKIND_DAX_KMEM, size_t size)

Using MEMKIND_DAX_KMEM, you can use both DRAM and persistent memory as separate NUMA nodes in a single application, similar to the logic used with file-based PMEM_KIND. Figure 1.3 shows an application that created two static kind objects: MEMKIND_DEFAULT and MEMKIND_DAX_PMEM.

The difference between the two kinds of memory-mapped to the application in Figure 1.3 is that MEMKIND_DAX_KMEM uses a memory-mapped file with the MAP_PRIVATE flag, while the dynamic MEMKIND_DEFAULT created with memkind_create_kind() uses MAP_SHARED when memory-mapping files on a DAX-enabled file system. The MAP_SHARED and MAP_PRIVATE definitions from the mmap() system call are defined  in the man pages as follows:

MAP_SHARED

Share this mapping. Updates to the mapping are visible to other processes mapping the same region and (in the case of file-backed mappings) are carried through to the underlying file. (To precisely control when updates are carried through to the underlying file requires the use of msync(2).)

MAP_PRIVATE

Create a private copy-on-write mapping. Updates to the mapping is not visible to other processes mapping the same file and are not carried through to the underlying file. It is unspecified whether changes made to the file after the mmap() call is visible in the mapped region.

Child processes created using the fork(2) system call inherit the same MAP_PRIVATE mappings from the parent process. When memory pages are modified by the parent process, a copy-on-write mechanism is triggered by the kernel to create an unmodified copy for child process. These pages are allocated on the same NUMA node as the original page.

Summary

In this article, we showed how persistent memory’s large capacity can be used to hold volatile application data. Applications can choose to allocate and access data from DRAM or persistent memory, or both. 

Memkind is a very flexible and easy-to-use library with semantics that are similar to the libc malloc/free APIs that developers frequently use. 

Libvmemcache is an embeddable and lightweight in-memory caching solution that allows applications to efficiently use persistent memory’s large capacity in a scalable way. Libvmemcache is an open-source project available on GitHub.

Resources

Persistent Memory Resources at Intel Developer Zone

PMEM.io

Persistent Memory Programming on GitHub