* introduce code node
* replace codeMap with code touches
* fix a comment
* fixups to tests
* fix compile error
* fix getnodedata tests
* add tests and test stubs
* add more test stubs
* add test method bodies
* add and fix more tests on trie for new codenode
* add test change code between blocks
* fix crash in stateless
* remove unneded files
* remove comment
* fix deleted account code
* fix resolve set builder for code nodes
* Fix download only
* Fix lint
* Reset references
* Only reset on error
* Potential fixes
* no NPE
* no NPE
* Not use multi-put
* Reduce ideal batch size for download only
* Handle tds == nil
* remove nested mutation
* Return multiput
* Better reporting
* Reduce batch size for download only
* Avoid extra copying
* Avoid extra copying
* IdealBatchSize
* Not write tx lookup entries
* Larger batches
* Go back to normal batch size
* Fix lint
* Gen tx lookup
* print progress
* Add filling up the lookup array
* Show tx count
* Introduce second round
* Add generating tx lookup
* Fix lint
* properly stop at specified block
* measure the duration of the last phase
* not to fail if the bucket is not found
* Fix lint
* Alternative tx generation
* Fix out of memory
* Fix out of memory
* Split in parts to conserve memory
* Copy keys
* Fix lint
* Fix lint
* Enable testGetNodeData by default
* Crude TestHashMapLeak
* nodeFlag ->nodeRef
* linter
* accountNode shouln't be in hashMap since only branch and short nodes are the standart ones
* Finalize hash eviction logic
* small changes to TestHashMapLeak
* Fix for the incorrect hash
Co-authored-by: ledgerwatch <akhounov@gmail.com>
* Use TrieDbState for tx pool
* Not initialise tx pool until state is loaded
* Add preimage
* Fix account
* Print codehash
* Print correct code hash
* Print incarnatin
* Print incarnatin
* Use proper incarnation
* Print dbValue
* Actually fix
* Actually fix
* Fix verifySnapshot
* readAccount to get code hash
* Next incarnation
* Print addrHashes with 0 incarnations
* Print storage history
* Print storage history
* Print storage history
* Print storage history
* Print all storage history
* print change set keys
* print change set keys
* print change set keys
* print change set keys
* Not print codebucket info
* Fixes
* Fix for incarnation
* Fix for storage history bucket
* Try to fix the leak
* Try to fix the leak
* Try to fix the leak
* Try to fix the leak
* Try to fix the leak
* Try to fix the leak
* Try to fix the leak
* Fix embedded nodes
* Hasher
* Fix
* Test fixes
* Add experimental debug flag
* Fix tx_pool_test
* Disable GetNodeData test unless in experiment
* Fix more tests
* Fix lint and revert some changes
* Fix lint
* Fix lint
* add env INTERMEDIATE_TRIE_CACHE
* try to use assert.New() pattern
* Fix "maligned" linter warnings to reduce space consumption of structs:
core/types/accounts/account.go:18:14: struct of size 136 bytes could be of size 128 bytes (maligned)
type Account struct {
--
trie/node.go:44:10: struct of size 80 bytes could be of size 72 bytes (maligned)
duoNode struct {
--
trie/resolve_set.go:28:17: struct of size 56 bytes could be of size 48 bytes (maligned)
type ResolveSet struct {
--
trie/resolver.go:34:15: struct of size 88 bytes could be of size 72 bytes (maligned)
type Resolver struct {
--
trie/visual.go:32:17: struct of size 104 bytes could be of size 96 bytes (maligned)
type VisualOpts struct {
The current trie memory database/cache that we do pruning on stores
trie nodes as binary rlp encoded blobs, and also stores the node
relationships/references for GC purposes. However, most of the trie
nodes (everything apart from a value node) is in essence just a
collection of references.
This PR switches out the RLP encoded trie blobs with the
collapsed-but-not-serialized trie nodes. This permits most of the
references to be recovered from within the node data structure,
avoiding the need to track them a second time (expensive memory wise).
Commit 40cdcf1183 broke the optimisation which kept nodes resolved
during Get in the trie. The decoder assigned cache generation 0
unconditionally, causing resolved nodes to get flushed on Commit.
This commit fixes it and adds two tests.
* trie: store nodes as pointers
This avoids memory copies when unwrapping node interface values.
name old time/op new time/op delta
Get 388ns ± 8% 215ns ± 2% -44.56% (p=0.000 n=15+15)
GetDB 363ns ± 3% 202ns ± 2% -44.21% (p=0.000 n=15+15)
UpdateBE 1.57µs ± 2% 1.29µs ± 3% -17.80% (p=0.000 n=13+15)
UpdateLE 1.92µs ± 2% 1.61µs ± 2% -16.25% (p=0.000 n=14+14)
HashBE 2.16µs ± 6% 2.18µs ± 6% ~ (p=0.436 n=15+15)
HashLE 7.43µs ± 3% 7.21µs ± 3% -2.96% (p=0.000 n=15+13)
* trie: close temporary databases in GetDB benchmark
* trie: don't keep []byte from DB load around
Nodes decoded from a DB load kept hashes and values as sub-slices of
the DB value. This can be a problem because loading from leveldb often
returns []byte with a cap that's larger than necessary, increasing
memory usage.
* trie: unload old cached nodes
* trie, core/state: use cache unloading for account trie
* trie: use explicit private flags (fixes Go 1.5 reflection issue).
* trie: fixup cachegen overflow at request of nick
* core/state: rename journal size constant
