The pmemkv project will no longer be maintained by Intel.
- Intel has ceased development and contributions including, but not limited to, maintenance, bug fixes, new releases, or updates, to this project.
- Intel no longer accepts patches to this project.
- If you have an ongoing need to use this project, are interested in independently developing it, or would like to maintain patches for the open source software community, please create your own fork of this project.
- You will find more information here.
pmemkv is a local/embedded key-value datastore optimized for persistent memory.
Rather than being tied to a single language or backing implementation, pmemkv
provides different options for language bindings and storage engines.
For more information, including C API and C++ API see: https://pmem.io/pmemkv. Documentation is available for every branch/release. For most recent always see (master branch):
Latest releases can be found on the "releases" tab.
There is also a small helper library pmemkv_json_config provided.
See its manual for details.
Installation guide
provides detailed instructions how to build and install pmemkv from sources,
build rpm and deb packages and explains usage of experimental engines and pool sets.
- Building from Sources
- Installing on Fedora
- Installing on Ubuntu
- Using Experimental Engines
- Building Packages
- Using a Pool Set
pmemkv is written in C/C++ and can be used in other languages - Java, Node.js,
Python, and Ruby.
Examples for C and C++ can be found within this repository in examples directory.
The above-mentioned bindings are maintained in separate GitHub repositories, but are still kept in sync with the main pmemkv distribution.
- Java - https://github.com/pmem/pmemkv-java
- Node.js - https://github.com/pmem/pmemkv-nodejs
- Python - https://github.com/pmem/pmemkv-python
- Ruby - https://github.com/pmem/pmemkv-ruby
pmemkv provides multiple storage engines that share common API, so every engine can be used with
all language bindings and utilities. Engines are loaded by name at runtime.
| Engine Name | Description | Experimental | Concurrent | Sorted | Persistent |
|---|---|---|---|---|---|
| cmap | Concurrent hash map | No | Yes | No | Yes |
| vsmap | Volatile sorted hash map | No | No | Yes | No |
| vcmap | Volatile concurrent hash map | No | Yes | No | No |
| csmap | Concurrent sorted map | Yes | Yes | Yes | Yes |
| radix | Radix tree | Yes | No | Yes | Yes |
| tree3 | Persistent B+ tree | Yes | No | No | Yes |
| stree | Sorted persistent B+ tree | Yes | No | Yes | Yes |
| robinhood | Persistent hash map with Robin Hood hashing | Yes | Yes | No | Yes |
The production quality engines are described in the libpmemkv(7) manual and the experimental ones are described in the ENGINES-experimental.md file.
pmemkv also provides testing engines, which may be used in unit tests or for benchmarking application overhead
| Engine Name | Description | Experimental | Concurrent | Sorted | Persistent |
|---|---|---|---|---|---|
| blackhole | Accepts everything, returns nothing | No | Yes | No | No |
| dram_vcmap | Volatile concurrent hash map placed entirely on DRAM | Yes | Yes | No | No |
Contributing a new engine is easy, so feel encouraged!
Experimental benchmark based on leveldb's db_bench to measure pmemkv's performance is available here: https://github.com/pmem/pmemkv-bench (previously pmemkv-tools).
If you read the blog post and still have some questions (especially about discontinuation of the project), please contact us using the dedicated e-mail: pmdk_support@intel.com.
