cerc-io/lighthouse
16
0
Fork 0
Code Issues 1 Pull Requests Packages Projects Releases Wiki Activity
73764d0dd2
lighthouse/common/malloc_utils/Cargo.toml

25 lines
767 B
TOML
Raw Normal View History

Tune GNU malloc (#2299) ## Issue Addressed NA ## Proposed Changes Modify the configuration of [GNU malloc](https://www.gnu.org/software/libc/manual/html_node/The-GNU-Allocator.html) to reduce memory footprint. - Set `M_ARENA_MAX` to 4. - This reduces memory fragmentation at the cost of contention between threads. - Set `M_MMAP_THRESHOLD` to 2mb - This means that any allocation >= 2mb is allocated via an anonymous mmap, instead of on the heap/arena. This reduces memory fragmentation since we don't need to keep growing the heap to find big contiguous slabs of free memory. - ~~Run `malloc_trim` every 60 seconds.~~ - ~~This shaves unused memory from the top of the heap, preventing the heap from constantly growing.~~ - Removed, see: https://github.com/sigp/lighthouse/pull/2299#issuecomment-825322646 *Note: this only provides memory savings on the Linux (glibc) platform.* ## Additional Info I'm going to close #2288 in favor of this for the following reasons: - I've managed to get the memory footprint *smaller* here than with jemalloc. - This PR seems to be less of a dramatic change than bringing in the jemalloc dep. - The changes in this PR are strictly runtime changes, so we can create CLI flags which disable them completely. Since this change is wide-reaching and complex, it's nice to have an easy "escape hatch" if there are undesired consequences. ## TODO - [x] Allow configuration via CLI flags - [x] Test on Mac - [x] Test on RasPi. - [x] Determine if GNU malloc is present? - I'm not quite sure how to detect for glibc.. This issue suggests we can't really: https://github.com/rust-lang/rust/issues/33244 - [x] Make a clear argument regarding the affect of this on CPU utilization. - [x] Test with higher `M_ARENA_MAX` values. - [x] Test with longer trim intervals - [x] Add some stats about memory savings - [x] Remove `malloc_trim` calls & code
2021-05-28 05:59:45 +00:00
[package]
name = "malloc_utils"
version = "0.1.0"
authors = ["Paul Hauner <paul@paulhauner.com>"]
Update to Rust 1.59 and 2021 edition (#3038) ## Proposed Changes Lots of lint updates related to `flat_map`, `unwrap_or_else` and string patterns. I did a little more creative refactoring in the op pool, but otherwise followed Clippy's suggestions. ## Additional Info We need this PR to unblock CI.
2022-02-25 00:10:17 +00:00
edition = "2021"
Tune GNU malloc (#2299) ## Issue Addressed NA ## Proposed Changes Modify the configuration of [GNU malloc](https://www.gnu.org/software/libc/manual/html_node/The-GNU-Allocator.html) to reduce memory footprint. - Set `M_ARENA_MAX` to 4. - This reduces memory fragmentation at the cost of contention between threads. - Set `M_MMAP_THRESHOLD` to 2mb - This means that any allocation >= 2mb is allocated via an anonymous mmap, instead of on the heap/arena. This reduces memory fragmentation since we don't need to keep growing the heap to find big contiguous slabs of free memory. - ~~Run `malloc_trim` every 60 seconds.~~ - ~~This shaves unused memory from the top of the heap, preventing the heap from constantly growing.~~ - Removed, see: https://github.com/sigp/lighthouse/pull/2299#issuecomment-825322646 *Note: this only provides memory savings on the Linux (glibc) platform.* ## Additional Info I'm going to close #2288 in favor of this for the following reasons: - I've managed to get the memory footprint *smaller* here than with jemalloc. - This PR seems to be less of a dramatic change than bringing in the jemalloc dep. - The changes in this PR are strictly runtime changes, so we can create CLI flags which disable them completely. Since this change is wide-reaching and complex, it's nice to have an easy "escape hatch" if there are undesired consequences. ## TODO - [x] Allow configuration via CLI flags - [x] Test on Mac - [x] Test on RasPi. - [x] Determine if GNU malloc is present? - I'm not quite sure how to detect for glibc.. This issue suggests we can't really: https://github.com/rust-lang/rust/issues/33244 - [x] Make a clear argument regarding the affect of this on CPU utilization. - [x] Test with higher `M_ARENA_MAX` values. - [x] Test with longer trim intervals - [x] Add some stats about memory savings - [x] Remove `malloc_trim` calls & code
2021-05-28 05:59:45 +00:00
[dependencies]
lighthouse_metrics = { path = "../lighthouse_metrics" }
lazy_static = "1.4.0"
libc = "0.2.79"
Update and consolidate dependencies (#3136) ## Proposed Changes I did some gardening :deciduous_tree: in our dependency tree: - Remove duplicate versions of `warp` (git vs patch) - Remove duplicate versions of lots of small deps: `cpufeatures`, `ethabi`, `ethereum-types`, `bitvec`, `nix`, `libsecp256k1`. - Update MDBX (should resolve #3028). I tested and Lighthouse compiles on Windows 11 now. - Restore `psutil` back to upstream - Make some progress updating everything to rand 0.8. There are a few crates stuck on 0.7. Hopefully this puts us on a better footing for future `cargo audit` issues, and improves compile times slightly. ## Additional Info Some crates are held back by issues with `zeroize`. libp2p-noise depends on [`chacha20poly1305`](https://crates.io/crates/chacha20poly1305) which depends on zeroize < v1.5, and we can only have one version of zeroize because it's post 1.0 (see https://github.com/rust-lang/cargo/issues/6584). The latest version of `zeroize` is v1.5.4, which is used by the new versions of many other crates (e.g. `num-bigint-dig`). Once a new version of chacha20poly1305 is released we can update libp2p-noise and upgrade everything to the latest `zeroize` version. I've also opened a PR to `blst` related to zeroize: https://github.com/supranational/blst/pull/111
2022-04-04 00:26:16 +00:00
parking_lot = "0.12.0"
Switch allocator to jemalloc (#3697) ## Proposed Changes Another `tree-states` motivated PR, this adds `jemalloc` as the default allocator, with an option to use the system allocator by compiling with `FEATURES="" make`. - [x] Metrics - [x] Test on Windows - [x] Test on macOS - [x] Test with `musl` - [x] Metrics dashboard on `lighthouse-metrics` (https://github.com/sigp/lighthouse-metrics/pull/37) Co-authored-by: Michael Sproul <micsproul@gmail.com>
2023-01-20 04:19:29 +00:00
jemalloc-ctl = { version = "0.5.0", optional = true }
# Jemalloc's background_threads feature requires Linux (pthreads).
[target.'cfg(target_os = "linux")'.dependencies]
jemallocator = { version = "0.5.0", optional = true, features = ["stats", "background_threads"] }
[target.'cfg(not(target_os = "linux"))'.dependencies]
jemallocator = { version = "0.5.0", optional = true, features = ["stats"] }
Enable `mallinfo2` behind feature flag (#2864) ## Proposed Changes Add `mallinfo2` behind a feature flag so that we can get accurate memory metrics during debugging. It can be enabled when building Lighthouse like so (so long as the platform supports it): ``` cargo install --path lighthouse --features "malloc_utils/mallinfo2" ```
2021-12-15 20:39:50 +00:00
[features]
mallinfo2 = []
Switch allocator to jemalloc (#3697) ## Proposed Changes Another `tree-states` motivated PR, this adds `jemalloc` as the default allocator, with an option to use the system allocator by compiling with `FEATURES="" make`. - [x] Metrics - [x] Test on Windows - [x] Test on macOS - [x] Test with `musl` - [x] Metrics dashboard on `lighthouse-metrics` (https://github.com/sigp/lighthouse-metrics/pull/37) Co-authored-by: Michael Sproul <micsproul@gmail.com>
2023-01-20 04:19:29 +00:00
jemalloc = ["jemallocator", "jemalloc-ctl"]
jemalloc-profiling = ["jemallocator/profiling"]
Reference in New Issue Copy Permalink