Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
be able to render the planet with 32gb of RAM (#618)
* move OutputObjects to mmap store For the planet, we need 1.3B output objects, 12 bytes per, so ~15GB of RAM. * treat objects at low zoom specially For GB, ~0.3% of objects are visible at low zooms. I noticed in previous planet runs that fetching the objects for tiles in the low zooms was quite slow - I think it's because we're scanning 1.3B objects each time, only to discard most of them. Now we'll only be scanning ~4M objects per tile, which is still an absurd number, but should mitigate most of the speed issue without having to properly index things. This will also help us maintain performance for memory-constrained users, as we won't be scanning all 15GB of data on disk, just a smaller ~45MB chunk. * make more explicit that this is unexpected * extend --materialize-geometries to nodes For Points stored via Layer(...) calls, store the node ID in the OSM store, unless `--materialize-geometries` is present. This saves ~200MB of RAM for North America, so perhaps 1 GB for the planet if NA has similar characteristics as the planet. Also fix the OSM_ID(...) macro - it was lopping off many more bits than needed, due to some previous experiments. Now that we want to track nodes, we need at least 34 bits. This may pose a problem down the road when we try to address thrashing. The mechanism I hoped to use was to divide the OSM stores into multiple stores covering different low zoom tiles. Ideally, we'd be able to recall which store to look in -- but we only have 36 bits, we need 34 to store the Node ID, so that leaves us with 1.5 bits => can divide into 3 stores. Since the node store for the planet is 44GB, dividing into 3 stores doesn't give us very much headroom on a 32 GB box. Ah well, we can sort this out later. * rejig AttributePair layout On g++, this reduces the size from 48 bytes to 34 bytes. There aren't _that_ many attribute pairs, even on the planet scale, but this plus a better encoding of string attributes might save us ~2GB at the planet level, which is meaningful for a 32GB box * fix initialization order warning * add PooledString Not used by anything yet. Given Tilemaker's limited needs, we can get away with a stripped-down string class that is less flexible than std::string, in exchange for memory savings. The key benefits - 16 bytes, not 32 bytes (g++) or 24 bytes (clang). When it does allocate (for strings longer than 15 bytes), it allocates from a pool so there's less per-allocation overhead. * add tests for attribute store ...I'm going to replace the string implementation, so let's have some backstop to make sure I don't break things * rejig isHot Break dependency on AttributePair, just work on std::string * teach PooledString to work with std::string ...this will be useful for doing map lookups when testing if an AttributePair has already been created with the given value. * use PooledString in AttributePair AttributePair has now been trimmed from 48 bytes to 18 bytes. There are 40M AttributeSets for the planet. That suggests there's probably ~30M AttributePairs, so hopefully this is a savings of ~900MB at the planet level. Runtime doesn't seem affected. There's a further opportunity for savings if we can make more strings qualify for the short string optimization. Only about 40% of strings fit in the 15 byte short string optimization. Of the remaining 60%, many are Latin-alphabet title cased strings like `Wellington Avenue` -- this could be encoded using 5 bits per letter, saving us an allocation. Even in the most optimistic case where: - there are 30M AttributePairs - of these, 90% are strings (= 27M) - of these, 60% don't fit in SSO (=16m) - of these, we can make 100% fit in SSO ...we only save about 256MB at the planet level, but at some significant complexity cost. So probably not worth pursuing at the moment. * log timings When doing the planet, especially on a box with limited memory, there are long periods with no output. Show some output so the user doesn't think things are hung. This also might be useful in detecting perf regressions more granularly. * AppendVector: an append-only chunked vector When using --store, deque is nice because growing doesn't require invalidating the old storage and copying it to a new location. However, it's also bad, because deque allocates in 512-byte chunks, which causes each 4KB OS page to have data from different z6 tiles. Instead, use our own container that tries to get the best of both worlds. Writing a random access iterator is new for me, so I don't trust this code that much. The saving grace is that the container is very limited, so errors in the iterator impelementation may not get exercised in practice. * fix progress when --store present * mutex on RelationScan progress output * make NodeStore/WayStore shardable This adds three methods to the stores: - `shard()` returns which shard you are - `shards()` returns how many shards total - `contains(shard, id)` returns whether or not shard N has an item with id X SortedNodeStore/SortedWayStore are not implemented yet, that'll come in a future commit. This will allow us to create a `ShardedNodeStore` and `ShardedWayStore` that contain N stores. We will try to ensure that each store has data that is geographically close to each other. Then, when reading, we'll do multiple passes of the PBF to populate each store. This should let us reduce the working set used to populate the stores, at the cost of additional linear scans of the PBF. Linear scans of disk are much less painful than random scans, so that should be a good trade. * add minimal SortedNodeStore test I'm going to rejig the innards of this class, so let's have some tests. * stop using internal linkage for atomics In order to shard the stores, we need to have multiple instances of the class. Two things block this currently: atomics at file-level, and thread-locals. Moving the atomics to the class is easy. Making the thread-locals per-class will require an approach similar to that adopted in https://github.com/systemed/tilemaker/blob/52b62dfbd5b6f8e4feb6cad4e3de86ba27874b3a/include/leased_store.h#L48, where we have a container that tracks the per-class data. * SortedNodeStore: abstract TLS behind storage() Still only supports 1 class, but this is a step along the path. * SortedWayStore: abstract TLS behind storage() * SortedNodeStore: support multiple instances * SortedWayStorage: support multiple instances * actually fix the low zoom object collection D'oh, this "worked" due to two bugs cancelling each other: (a) the code to find things in the low zoom list never found anything, because it assumed a base z6 tile of 0/0 (b) we weren't returning early, so the normal code still ran Rejigged to actually do what I was intending * AppendVector tweaks * more low zoom fixes * implement SortedNodeStore::contains * implement SortedWayStore::contains * use TileCoordinatesSet * faster covered tile enumeration Do a single pass, rather than one pass per zoom. * add ShardedNodeStore This distributes nodes into one of 8 shards, trying to roughly group parts of the globe by complexity. This should help with locality when writing tiles. A future commit will add a ShardedWayStore and teach read_pbf to read in a locality-aware manner, which should help when reading ways. * add ShardedWayStore Add `--shard-stores` flag. It's not clear yet this'll be a win, will need to benchmark. The cost of reading the PBF blocks repeatedly is a bit higher than I was expecting. It might be worth seeing if we can index the blocks to skip fruitless reads. * fewer, more balanced shards * skip ReadPhase::Ways passes if node store is empty * support multiple passes for ReadPhase::Relations * fix check for first way * adjust shards With this distribution, no node shard is more than ~8.5GB. * Relations: fix effectiveShards > 1 check Oops, bug that very moderately affected performance in the non `--shard-stores` case * extend --materialize-geometries to LayerAsCentroid It turns out that about 20% of LayerAsCentroid calls are for nodes, which this branch could already do. The remaining calls are predominantly ways, e.g. housenumbers. We always materialize relation centroids, as they're expensive to compute. In GB, this saves about 6.4M points, ~102M. Scaled to the planet, it's perhaps a 4.5GB savings, which should let us use a more aggressive shard strategy. It seems to add 3-4 seconds to the time to process GB. * add `DequeMap`, change AttributeStore to use it This implements the idea in #622 (comment) Rather than storing a `deque<T>` and a `flat_map<T*, uint32_t>`, store a `deque<T>` and `vector<uint32_t>`, to save 8 bytes per AttributePair and AttributeSet. * capture s(this) Seems to save ~1.5 seconds on GB * fix warning * fix warning, really * fewer shards Shard 1 (North America) is ~4.8GB of nodes, shard 4 (some of Europe) is 3.7GB. Even ignoring the memory savings in the recent commits, these could be merged. * extract option parsing to own file We'd like to have different defaults based on whether `--store` is present. Now that option parsing will have some more complex logic, let's pull it into its own class so it can be more easily tested. * use sensible defaults based on presence of --store * improve test coverage * fixes * update number of shards to 6 This has no performance impact as we never put anything in the 7th shard, and so we skip doing the 7th pass in the ReadPhase::Ways and ReadPhase::Relations phase. The benefit is only to avoid emitting a noisy log about how the 7th store has 0 entries in it. Timings with 6 shards on Vultr's 16-core machine here: https://gist.github.com/cldellow/77991eb4074f6a0f31766cf901659efb The new peak memory is ~12.2GB. I am a little perplexed -- the runtime on a 16-core server was previously: ``` $ time tilemaker --store /tmp/store --input planet-latest.osm.pbf --output tiles.mbtiles --shard-stores real 195m7.819s user 2473m52.322s sys 73m13.116s ``` But with the most recent commits on this branch, it was: ``` real 118m50.098s user 1531m13.026s sys 34m7.252s ``` This is incredibly suspicious. I also tried re-running commit bbf0957, and got: ``` real 123m15.534s user 1546m25.196s sys 38m17.093s ``` ...so I can't explain why the earlier runs took 195 min. Ideas: - the planet changed between runs, and a horribly broken geometry was fixed - Vultr gives quite different machines for the same class of server - perhaps most likely: I failed to click "CPU-optimized" when picking the earlier server, and got a slow machine the first time, and a fast machine the second time. I'm pretty sure I paid the same $, so I'm not sure I believe this. I don't think I really believe that a 33% reduction in runtime is explained by any of those, though. Anyway, just another thing to be befuddled by. * --store uses lazy geometries; permit overriding I did some experiments on a Hetzner 48-core box with 192GB of RAM: --store, materialize geometries: real 65m34.327s user 2297m50.204s sys 65m0.901s The process often failed to use 100% of CPU--if you naively divide user+sys/real you get ~36, whereas the ideal would be ~48. Looking at stack traces, it seemed to coincide with calls to Boost's rbtree_best_fit allocator. Maybe: - we're doing disk I/O, and it's just slower than recomputing the geometries - we're using the Boost mmap library suboptimally -- maybe there's some other allocator we could be using. I think we use the mmap allocator like a simple bump allocator, so I don't know why we'd need a red-black tree --store, lazy geometries: real 55m33.979s user 2386m27.294s sys 23m58.973s Faster, but still some overhead (user+sys/real => ~43) no --store, materialize geometries: OOM no --store, lazy geometries (used 175GB): real 51m27.779s user 2306m25.309s sys 16m34.289s This was almost 100% CPU - user+sys/real => ~45) From this, I infer: - `--store` should always default to lazy geometries in order to minimize the I/O burden - `--materialize-geometries` is a good default for non-store usage, but it's still useful to be able to override and use lazy geometries, if it then means you can fit the data entirely in memory
- Loading branch information
