Top Related Projects
Quick Overview
snmalloc is a high-performance memory allocator developed by Microsoft Research. It is designed to be fast, memory-efficient, and scalable for multi-threaded applications. snmalloc uses novel techniques to minimize contention and improve cache locality, making it particularly suitable for large-scale server applications.
Pros
- Excellent performance in multi-threaded environments
- Low memory fragmentation and efficient memory usage
- Thread-local caching for improved speed and reduced contention
- Easy to integrate into existing projects
Cons
- May not be optimal for small, single-threaded applications
- Requires C++17 or later, which might limit compatibility with older codebases
- Documentation could be more comprehensive for advanced usage scenarios
- Relatively new compared to some other established allocators
Code Examples
- Basic usage of snmalloc:
#include <snmalloc.h>
int main() {
void* ptr = snmalloc::ThreadAlloc::get().alloc(1024);
// Use the allocated memory
snmalloc::ThreadAlloc::get().dealloc(ptr);
return 0;
}
- Using snmalloc with custom alignment:
#include <snmalloc.h>
int main() {
constexpr size_t alignment = 64;
void* ptr = snmalloc::ThreadAlloc::get().alloc(1024, alignment);
// Use the aligned memory
snmalloc::ThreadAlloc::get().dealloc(ptr);
return 0;
}
- Allocating an array of objects:
#include <snmalloc.h>
#include <new>
struct MyStruct { int x; double y; };
int main() {
size_t count = 100;
MyStruct* arr = static_cast<MyStruct*>(snmalloc::ThreadAlloc::get().alloc(sizeof(MyStruct) * count));
for (size_t i = 0; i < count; i++) {
new (&arr[i]) MyStruct();
}
// Use the array
for (size_t i = 0; i < count; i++) {
arr[i].~MyStruct();
}
snmalloc::ThreadAlloc::get().dealloc(arr);
return 0;
}
Getting Started
To use snmalloc in your project:
-
Clone the repository:
git clone https://github.com/microsoft/snmalloc.git
-
Add snmalloc as a subdirectory in your CMakeLists.txt:
add_subdirectory(path/to/snmalloc)
-
Link your target with snmalloc:
target_link_libraries(your_target snmalloc)
-
Include snmalloc in your C++ code:
#include <snmalloc.h>
-
Compile your project with C++17 or later:
g++ -std=c++17 your_code.cpp -o your_program
Competitor Comparisons
Public domain cross platform lock free thread caching 16-byte aligned memory allocator implemented in C
Pros of rpmalloc
- Simpler implementation, making it easier to understand and maintain
- Lower memory overhead for small allocations
- Better performance on some benchmarks, particularly for small allocations
Cons of rpmalloc
- Less focus on security features compared to snmalloc
- May not perform as well for large allocations or high-concurrency scenarios
- Less extensive documentation and community support
Code Comparison
rpmalloc:
static void* _rpmalloc_allocate_large(size_t size) {
size_t total_size = size + SPAN_HEADER_SIZE;
size_t num_spans = total_size >> _memory_span_size_shift;
if (total_size & (_memory_span_size - 1))
++num_spans;
size_t span_count = (size_t)num_spans;
span_t* span = _rpmalloc_heap_allocate_spans(heap, span_count);
return pointer_offset(span, SPAN_HEADER_SIZE);
}
snmalloc:
void* Alloc::alloc_large(size_t size)
{
size = round_size(size);
auto sizeclass = size_to_sizeclass(size);
auto rsize = sizeclass_to_size(sizeclass);
auto span = large_allocator.alloc(rsize);
if (span == nullptr)
return nullptr;
return span->start;
}
Both allocators use similar approaches for large allocations, but snmalloc's implementation is more abstracted and uses higher-level constructs.
Pros of jemalloc
- Widely adopted and battle-tested in large-scale production environments
- Extensive performance tuning options and customization capabilities
- Strong support for multi-threaded applications and scalability
Cons of jemalloc
- More complex codebase, potentially harder to maintain or modify
- Higher memory overhead in some scenarios due to its sophisticated management structures
- May require more fine-tuning to achieve optimal performance in specific use cases
Code Comparison
jemalloc:
void *ptr = malloc(size);
free(ptr);
snmalloc:
void *ptr = snmalloc::ThreadAlloc::get().alloc(size);
snmalloc::ThreadAlloc::get().dealloc(ptr);
Key Differences
- snmalloc is designed with a focus on security and memory isolation
- jemalloc offers more extensive statistics and debugging features
- snmalloc has a simpler codebase, potentially easier to integrate and maintain
- jemalloc provides better support for fragmentation reduction in long-running applications
Both allocators aim to improve performance and memory efficiency, but they take different approaches. snmalloc emphasizes security and simplicity, while jemalloc focuses on scalability and extensive customization options. The choice between them depends on specific project requirements and use cases.
Pros of tcmalloc
- Highly optimized for multi-threaded applications
- Extensive performance profiling and debugging tools
- Wider adoption and longer history of use in production environments
Cons of tcmalloc
- More complex implementation, potentially harder to maintain
- May have higher memory overhead for small allocations
- Less focus on security features compared to snmalloc
Code Comparison
snmalloc:
void* alloc = snmalloc::ThreadAlloc::get().alloc(size);
snmalloc::ThreadAlloc::get().dealloc(alloc);
tcmalloc:
void* alloc = tc_malloc(size);
tc_free(alloc);
Both libraries provide similar APIs for allocation and deallocation, but snmalloc uses a thread-local allocator object, while tcmalloc uses global functions.
snmalloc focuses on simplicity and security, with features like randomized allocation and memory zeroing. tcmalloc emphasizes performance and scalability, particularly for multi-threaded applications.
tcmalloc offers more advanced profiling and debugging tools, which can be beneficial for large-scale projects. However, snmalloc's simpler design may make it easier to integrate and maintain in smaller projects or those with specific security requirements.
mimalloc is a compact general purpose allocator with excellent performance.
Pros of mimalloc
- Higher performance in multi-threaded scenarios
- More extensive documentation and benchmarks
- Wider adoption and community support
Cons of mimalloc
- Slightly larger memory footprint
- More complex implementation, potentially harder to maintain
Code comparison
mimalloc:
#include <mimalloc.h>
void* p = mi_malloc(sizeof(int));
mi_free(p);
snmalloc:
#include <snmalloc.h>
void* p = snmalloc::malloc(sizeof(int));
snmalloc::free(p);
Key differences
- mimalloc uses a global API, while snmalloc requires namespace usage
- mimalloc focuses on performance optimizations for common allocation patterns
- snmalloc emphasizes memory isolation and security features
Performance
- mimalloc generally outperforms snmalloc in multi-threaded scenarios
- snmalloc may have an edge in single-threaded applications
Memory usage
- snmalloc typically has a smaller memory footprint
- mimalloc trades some memory overhead for improved performance
Compatibility
- Both allocators can be used as drop-in replacements for standard malloc
- mimalloc has broader platform support and more extensive testing
Community and development
- mimalloc has a larger user base and more frequent updates
- snmalloc focuses on specific use cases and security features
Convert designs to code with AI
Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.
Try Visual CopilotREADME
snmalloc
snmalloc is a high-performance allocator.
snmalloc can be used directly in a project as a header-only C++ library,
it can be LD_PRELOAD
ed on Elf platforms (e.g. Linux, BSD),
and there is a crate to use it from Rust.
Its key design features are:
- Memory that is freed by the same thread that allocated it does not require any synchronising operations.
- Freeing memory in a different thread to initially allocated it, does not take any locks and instead uses a novel message passing scheme to return the memory to the original allocator, where it is recycled. This enables 1000s of remote deallocations to be performed with only a single atomic operation enabling great scaling with core count.
- The allocator uses large ranges of pages to reduce the amount of meta-data required.
- The fast paths are highly optimised with just two branches on the fast path for malloc (On Linux compiled with Clang).
- The platform dependencies are abstracted away to enable porting to other platforms.
snmalloc's design is particular well suited to the following two difficult scenarios that can be problematic for other allocators:
- Allocations on one thread are freed by a different thread
- Deallocations occur in large batches
Both of these can cause massive reductions in performance of other allocators, but do not for snmalloc.
The implementation of snmalloc has evolved significantly since the initial paper. The mechanism for returning memory to remote threads has remained, but most of the meta-data layout has changed. We recommend you read docs/security to find out about the current design, and if you want to dive into the code docs/AddressSpace.md provides a good overview of the allocation and deallocation paths.
Hardening
There is a hardened version of snmalloc, it contains
- Randomisation of the allocations' relative locations,
- Most meta-data is stored separately from allocations, and is protected with guard pages,
- All in-band meta-data is protected with a novel encoding that can detect corruption, and
- Provides a
memcpy
that automatically checks the bounds relative to the underlying malloc.
A more comprehensive write up is in docs/security.
Further documentation
Contributing
This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com.
When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.
This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.
Top Related Projects
Convert designs to code with AI
Introducing Visual Copilot: A new AI model to turn Figma designs to high quality code using your components.
Try Visual Copilot