marvinlanhenke commented on code in PR #10748: URL: https://github.com/apache/datafusion/pull/10748#discussion_r1623275095
########## datafusion/common/src/utils/memory.rs: ########## @@ -0,0 +1,123 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +//! This module provides a function to estimate the memory size of a HashTable prior to alloaction + +/// Estimates the memory size required for a hash table prior to allocation. +/// +/// # Parameters +/// - `num_elements`: The number of elements expected in the hash table. +/// - `fixed_size`: A fixed overhead size associated with the collection +/// (e.g., HashSet or HashTable). +/// - `T`: The type of elements stored in the hash table. +/// +/// # Details +/// This function calculates the estimated memory size by considering: +/// - An overestimation of buckets to keep approximately 1/8 of them empty. +/// - The total memory size is computed as: +/// - The size of each entry (`T`) multiplied by the estimated number of +/// buckets. +/// - One byte overhead for each bucket. +/// - The fixed size overhead of the collection. +/// - Returns `usize::MAX` if an overflow occurs. +/// +/// # Examples +/// --- +/// +/// ## From within a struct +/// +/// ```rust +/// use datafusion_common::utils::memory::estimate_memory_size; +/// +/// struct MyStruct<T> { +/// values: Vec<T>, +/// other_data: usize, +/// } +/// +/// impl<T> MyStruct<T> { +/// fn size(&self) -> usize { +/// let num_elements = self.values.len(); +/// let fixed_size = std::mem::size_of_val(self) + +/// std::mem::size_of_val(&self.values); +/// +/// estimate_memory_size::<T>(num_elements, fixed_size) +/// } +/// } +/// ``` +/// --- +/// ## With a simple collection +/// +/// ```rust +/// use datafusion_common::utils::memory::estimate_memory_size; +/// use std::collections::HashMap; +/// +/// let num_rows = 100; +/// let fixed_size = std::mem::size_of::<HashMap<u64, u64>>(); +/// let estimated_hashtable_size = +/// estimate_memory_size::<(u64, u64)>(num_rows,fixed_size); +/// ``` +pub fn estimate_memory_size<T>(num_elements: usize, fixed_size: usize) -> usize { + // For the majority of cases hashbrown overestimates the bucket quantity + // to keep ~1/8 of them empty. We take this factor into account by + // multiplying the number of elements with a fixed ratio of 8/7 (~1.14). + // This formula leads to overallocation for small tables (< 8 elements) + // but should be fine overall. + num_elements + .checked_mul(8) + .and_then(|estimate| estimate.checked_div(7)) + .map(|buckets| buckets.next_power_of_two()) + .and_then(|estimated_buckets| { + // + size of entry * number of buckets + // + 1 byte for each bucket + // + fixed size of collection (HashSet/HashTable) + std::mem::size_of::<T>() + .checked_mul(estimated_buckets) + .and_then(|size_of_entries| { + estimated_buckets + .checked_add(fixed_size) + .and_then(|overhead| size_of_entries.checked_add(overhead)) + }) + }) + .unwrap_or(usize::MAX) +} + +#[cfg(test)] +mod tests { + use std::collections::HashSet; + + use super::estimate_memory_size; + + #[test] + fn test_estimate_memory() { + let num_elements = 8; + // size: 48 + let fixed_size = std::mem::size_of::<HashSet<u32>>(); + // size: 128 = 16 * 4 + 16 + 48 Review Comment: Ah sorry it's not well documented...16 is the number of estimated buckets -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
