How Does libc Implement Short String Optimization (SSO)?

Mechanics of Short String Optimization in libc

Short String Optimization (SSO) is a technique used to improve the performance of string operations by storing small strings directly in the string object's memory instead of allocating separate storage on the heap. This optimization reduces the overhead of dynamic memory allocation and improves memory locality.

libc SSO Implementation

• Size Threshold: The maximum size for SSO varies depending on the architecture. On 32-bit machines, strings with up to 10 characters qualify for SSO. On 64-bit machines, it's up to 22 characters.
• Data Storage: In SSO strings, the first byte of the string object contains a flag indicating whether the string is short or long. The remaining bytes represent the string data.
• Size Extraction: Since the size field is stored in only 7 bits, it needs to be shifted to get the actual size:

size_type __get_short_size() const {
    return __r_.first().__s.__size_ >> 1;
}

Long Strings

• Layout: Long strings use a separate capacity member to store the string's actual capacity.
• Capacity Access: The getter and setter for the capacity of a long string use a mask (__long_mask) to work around the is_long bit.

_LIBCPP_ABI_ALTERNATE_STRING_LAYOUT

The _LIBCPP_ABI_ALTERNATE_STRING_LAYOUT flag rearranges the data members of the string object, placing the data pointer first. This change is intended to improve alignment and potentially enhance performance, but should be used with caution as it creates a different ABI.

The above is the detailed content of How Does libc Implement Short String Optimization (SSO)?. For more information, please follow other related articles on the PHP Chinese website!