Commit graph

4877 commits

Author SHA1 Message Date
Roman Khimov
932a57e1e4 keys: reuse coordLen where appropriate 2021-08-26 17:30:04 +03:00
Roman Khimov
2585713f31 *: use Go 1.17 by default, test with 1.15, 1.16 and 1.17 2021-08-26 17:30:04 +03:00
Roman Khimov
6d074a96e9 *: make tests use TempDir(), fix #1319
Simplify things, drop TempFile at the same time (refs. #1764)
2021-08-26 17:29:40 +03:00
Roman Khimov
f4ba21a41a keys: use (*Int).FillBytes where appropriate
Allows to avoid some allocations. Refs. #1319.
2021-08-25 22:35:39 +03:00
Roman Khimov
76eca07961 keys: simplify NewPrivateKeyFrom* functions
Avoid allocating a slice and doing double calculations.
2021-08-25 22:35:39 +03:00
Roman Khimov
61ea42c570 keys: simplify end of buffer check 2021-08-25 22:35:39 +03:00
Roman Khimov
4803cc15c7 keys: add (*PublicKey).DecodeBytes benchmark
Attempts to reuse elliptic.Unmarshal() and elliptic.UnmarshalCompressed() lead
to this:
name                 old time/op    new time/op    delta
PublicDecodeBytes-8    59.5µs ± 2%    61.8µs ± 1%  +3.78%  (p=0.000 n=10+9)

name                 old alloc/op   new alloc/op   delta
PublicDecodeBytes-8    3.99kB ± 0%    4.27kB ± 0%  +6.81%  (p=0.000 n=9+10)

name                 old allocs/op  new allocs/op  delta
PublicDecodeBytes-8       136 ± 0%       135 ± 0%  -0.74%  (p=0.000 n=10+10)

So it makes no sense. Refs. #1319.
2021-08-25 22:35:39 +03:00
Roman Khimov
a1d96a7d7d keys: use elliptic package marshalling functions, #1319
name                       old time/op    new time/op    delta
PublicBytes-8                81.4ns ± 6%    71.2ns ± 8%  -12.56%  (p=0.000 n=10+10)
PublicUncompressedBytes-8    93.2ns ±17%    72.5ns ±14%  -22.25%  (p=0.000 n=10+10)

name                       old alloc/op   new alloc/op   delta
PublicBytes-8                 80.0B ± 0%     48.0B ± 0%  -40.00%  (p=0.000 n=10+10)
PublicUncompressedBytes-8     80.0B ± 0%     48.0B ± 0%  -40.00%  (p=0.000 n=10+10)

name                       old allocs/op  new allocs/op  delta
PublicBytes-8                  2.00 ± 0%      1.00 ± 0%  -50.00%  (p=0.000 n=10+10)
PublicUncompressedBytes-8      2.00 ± 0%      1.00 ± 0%  -50.00%  (p=0.000 n=10+10)
2021-08-25 22:35:39 +03:00
Roman Khimov
217d7bdf44 keys: add equality benchmark
Go 1.15 provides native (*ecdsa.PublicKey).Equal method, but we can't drop our
own Equal because the types are different and there is still code using our
Equal (forcing it to convert types is counterproductive), while changing
(*PublicKey).Equal to use (*ecdsa.PublicKey).Equal internally with some kind of

  (*ecdsa.PublicKey)(p).Equal((*ecdsa.PublicKey)(key))

slows it down:

name           old time/op    new time/op    delta
PublicEqual-8    14.9ns ± 1%    18.4ns ± 2%  +23.55%  (p=0.000 n=9+10)

name           old alloc/op   new alloc/op   delta
PublicEqual-8     0.00B          0.00B          ~     (all equal)

name           old allocs/op  new allocs/op  delta
PublicEqual-8      0.00           0.00          ~     (all equal)

So leave it as is, but add this micro-bench. Refs. #1319.
2021-08-25 15:18:26 +03:00
Roman Khimov
ad5acf4873 *: drop Go 1.14 2021-08-25 14:59:00 +03:00
Roman Khimov
581132727b
Merge pull request #2141 from nspcc-dev/decoding-speedup
Decoding speedup
2021-08-24 13:33:09 +03:00
Roman Khimov
7808762ba0 transaction: avoid reencoding and reading what can't be read
name               old time/op    new time/op    delta
DecodeFromBytes-8    1.79µs ± 2%    1.46µs ± 4%  -18.44%  (p=0.000 n=10+10)

name               old alloc/op   new alloc/op   delta
DecodeFromBytes-8      800B ± 0%      624B ± 0%  -22.00%  (p=0.000 n=10+10)

name               old allocs/op  new allocs/op  delta
DecodeFromBytes-8      10.0 ± 0%       8.0 ± 0%  -20.00%  (p=0.000 n=10+10)
2021-08-23 21:41:38 +03:00
Roman Khimov
d0620b24ec io: simplify BinReader uint buffer
Similar to c69670c85b, allows to eliminate one
allocation and reduce memory footprint a bit (tested on tx decoding):

name               old time/op    new time/op    delta
DecodeFromBytes-8    1.78µs ± 3%    1.79µs ± 2%    ~     (p=1.000 n=10+10)

name               old alloc/op   new alloc/op   delta
DecodeFromBytes-8      888B ± 0%      800B ± 0%  -9.91%  (p=0.000 n=10+10)

name               old allocs/op  new allocs/op  delta
DecodeFromBytes-8      11.0 ± 0%      10.0 ± 0%  -9.09%  (p=0.000 n=10+10)
2021-08-23 21:18:07 +03:00
Roman Khimov
1a733ca456
Merge pull request #2135 from nspcc-dev/fix-test
cli: fix tests for go1.17
2021-08-20 16:48:22 +03:00
Roman Khimov
37c92f5c2f
Merge pull request #2136 from nspcc-dev/micro-reduce-allocs
Drop some useless allocations
2021-08-20 16:48:14 +03:00
Roman Khimov
2808f6857d interop: don't allocate for Functions and Notifications in New
Functions are usually immediately replaced (and it's OK for them to be nil,
searching through an array with length of zero is fine), Notifications are
usually appended to (and are absolutely useless in verification contexts).
2021-08-20 11:56:28 +03:00
Roman Khimov
2e39f1a1e3 io: drop one allocation from NewBufBinWriter 2021-08-20 11:38:42 +03:00
Roman Khimov
a68a8aa8fc core: simplify and correct notification handling
* both 'to' and 'from' are either Null or Hash160, there is no other
   possibility for valid NEP-17. So returning util.Uint160{} in case of
   parsing error is wrong.
 * but this is what allowed burns/mints to work at the expense of error
   allocation inside of util.Uint160DecodeBytesBE()
 * Uint160 can technically fit into regular VM integer, so even though it'd be
   quite surprising to see it there, TryBytes() is more correct (and easier!)
   to use
 * same thing with `amount`, we have `TryInteger()` that easily covers all
   possible cases and does appropriate error checking inside
2021-08-20 11:26:16 +03:00
Roman Khimov
abc48229a3 block: Grow buffer on Trim, avoid reallocations 2021-08-20 11:05:46 +03:00
Roman Khimov
b8dd284d3d io: don't allocate new error on every call to Bytes()
It makes no sense and we're using Bytes() pretty often.
2021-08-20 10:58:51 +03:00
Evgeniy Stratonikov
0b681d40e8 cli: fix tests for go1.17
`getPath` returns paths without leading dot.

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-19 18:16:56 +03:00
Roman Khimov
894cfe29be CHANGELOG: minor fix for 0.97.2 2021-08-18 15:15:59 +03:00
Roman Khimov
42e69c12ee CHANGELOG: release 0.97.2 2021-08-18 15:06:12 +03:00
Roman Khimov
eb37f982a1
Merge pull request #2134 from nspcc-dev/remove-untraceable-conflicts
core: remove untraceable dummy conflicting transactions
2021-08-18 14:45:05 +03:00
Anna Shaleva
5f9d38f640 core: refactor (*DAO).StoreAsTransaction
Squash (*DAO).StoreAsTransaction and
(*DAO).StoreConflictingTransactions. It's better to keep them this way,
because StoreAsTransaction is always followed by
StoreConflictingTransactions, so it's an atomic operation.

The logic wasn't changed.
2021-08-18 13:39:28 +03:00
Anna Shaleva
4b35a1cf92 core: remove conflicting transactions wrt MaxTraceableBlocks 2021-08-18 13:31:47 +03:00
Roman Khimov
483934d3a6
Merge pull request #2133 from nspcc-dev/optimize-util
util: reduce allocations in `util.Uint256DecodeStringLE`
2021-08-17 19:20:16 +03:00
Roman Khimov
a323beb03f
Merge pull request #2132 from nspcc-dev/block-contracts
contract: block calls to contracts via Policy contracts
2021-08-17 19:18:59 +03:00
Evgeniy Stratonikov
8c31831626 util: reduce allocations in util.Uint256DecodeStringLE
It is used a lot in clients (including our benchmark).
`Uint160` is already optimized.

```
name                     old time/op    new time/op    delta
Uint256DecodeStringLE-8     150ns ±15%     112ns ± 3%  -25.23%  (p=0.000 n=10+10)

name                     old alloc/op   new alloc/op   delta
Uint256DecodeStringLE-8     96.0B ± 0%     64.0B ± 0%  -33.33%  (p=0.000 n=10+10)

name                     old allocs/op  new allocs/op  delta
Uint256DecodeStringLE-8      2.00 ± 0%      1.00 ± 0%  -50.00%  (p=0.000 n=10+10)
```

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-17 16:53:39 +03:00
Roman Khimov
f477a48758 contract: block calls to contracts via Policy contract
See neo-project/neo#2567.
2021-08-17 15:24:06 +03:00
Roman Khimov
11351b9702
Merge pull request #2114 from nspcc-dev/optimize-rpc
rpc/request: delay parameter unmarshaling
2021-08-13 16:30:42 +03:00
Evgeniy Stratonikov
3c34e6fa21 rpc/request: delay parameter unmarshaling
It is rather costly to try to unmarshal many structs in order.

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 16:22:54 +03:00
Roman Khimov
5b12dd2025
Merge pull request #2128 from nspcc-dev/vm-update-int
Some VM optimizations
2021-08-13 16:16:01 +03:00
Evgeniy Stratonikov
6879f76a13 stackitem: make Buffer an alias to []byte
Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 14:41:26 +03:00
Evgeniy Stratonikov
1dfef4ba26 stackitem: make ByteArray an alias to []byte
Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 14:41:26 +03:00
Evgeniy Stratonikov
4f98ec2f53 vm: embed reference counter in compound items
```
name              old time/op  new time/op  delta
RefCounter_Add-8  44.8ns ± 4%  11.7ns ± 3%  -73.94%  (p=0.000 n=10+10)
```

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 14:41:26 +03:00
Roman Khimov
adc660c3e0
Merge pull request #2123 from nspcc-dev/store-better
Store better
2021-08-13 12:50:24 +03:00
Evgeniy Stratonikov
dc9287bf5c compiler: use parameter directly in writeJumps
`Next` doesn't longer copy parameter.

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 11:59:04 +03:00
Evgeniy Stratonikov
f5d1277bfd vm: do not copy parameter
Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 11:52:38 +03:00
Roman Khimov
4a714425bf
Merge pull request #2125 from nspcc-dev/cli-convert-public
vm/cli: add public key -> address conversion, fix #2121
2021-08-13 11:49:49 +03:00
Evgeniy Stratonikov
e2910a7cb4 vm/cli: add public key -> address conversion, fix #2121
Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 10:43:49 +03:00
Evgeniy Stratonikov
bb137abb03 crypto/keys: enforce length in PublicKey.DecodeBytes()
Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-13 10:38:09 +03:00
Evgeniy Stratonikov
a5516e8c96 stackitem: make BigInteger alias to big.Int
Remove one indirection step.
``
name       old time/op    new time/op    delta
MakeInt-8    79.7ns ± 8%    56.2ns ± 8%  -29.44%  (p=0.000 n=10+10)

name       old alloc/op   new alloc/op   delta
MakeInt-8     48.0B ± 0%     40.0B ± 0%  -16.67%  (p=0.000 n=10+10)

name       old allocs/op  new allocs/op  delta
MakeInt-8      3.00 ± 0%      2.00 ± 0%  -33.33%  (p=0.000 n=10+10)
```

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-12 17:53:36 +03:00
Evgeniy Stratonikov
cff8b1c24e stackitem: use Bool item directly
It is always copied.

Signed-off-by: Evgeniy Stratonikov <evgeniy@nspcc.ru>
2021-08-12 17:53:36 +03:00
Roman Khimov
ae071d4542 storage: introduce PutChangeSet and use it for Persist
We're using batches in wrong way during persist, we already have all changes
accumulated in two maps and then we move them to batch and then this is
applied. For some DBs like BoltDB this batch is just another MemoryStore, so
we essentially just shuffle the changeset from one map to another, for others
like LevelDB batch is just a serialized set of KV pairs, it doesn't help much
on subsequent PutBatch, we just duplicate the changeset again.

So introduce PutChangeSet that allows to take two maps with sets and deletes
directly. It also allows to simplify MemCachedStore logic.

neo-bench for single node with 10 workers, LevelDB:

  Reference:

  RPS    30189.132 30556.448 30390.482 ≈ 30379    ±  0.61%
  TPS    29427.344 29418.687 29434.273 ≈ 29427    ±  0.03%
  CPU %     33.304    27.179    33.860 ≈    31.45 ± 11.79%
  Mem MB   800.677   798.389   715.042 ≈   771    ±  6.33%

  Patched:

  RPS    30264.326 30386.364 30166.231 ≈ 30272    ± 0.36% ⇅
  TPS    29444.673 29407.440 29452.478 ≈ 29435    ± 0.08% ⇅
  CPU %     34.012    32.597    33.467 ≈   33.36  ± 2.14% ⇅
  Mem MB   549.126   523.656   517.684 ≈  530     ± 3.15% ↓ 31.26%

BoltDB:

  Reference:

  RPS    31937.647 31551.684 31850.408 ≈ 31780    ±  0.64%
  TPS    31292.049 30368.368 31307.724 ≈ 30989    ±  1.74%
  CPU %     33.792    22.339    35.887 ≈    30.67 ± 23.78%
  Mem MB  1271.687  1254.472  1215.639 ≈  1247    ±  2.30%

  Patched:

  RPS    31746.818 30859.485 31689.761 ≈ 31432    ± 1.58% ⇅
  TPS    31271.499 30340.726 30342.568 ≈ 30652    ± 1.75% ⇅
  CPU %     34.611    34.414    31.553 ≈    33.53 ± 5.11% ⇅
  Mem MB  1262.960  1231.389  1335.569 ≈  1277    ± 4.18% ⇅
2021-08-12 17:42:16 +03:00
Roman Khimov
5aff82aef4
Merge pull request #2119 from nspcc-dev/states-exchange/insole
core, network: prepare basis for Insole module
2021-08-12 10:35:02 +03:00
Roman Khimov
3360c4a0d8
Merge pull request #2122 from nspcc-dev/vm-microopt
VM-related microoptimizations
2021-08-12 10:34:49 +03:00
Roman Khimov
47f0f4c45f dao: completely drop Cached
It was very useful in 2.0 days, but today it only serves one purpose that
could easily (and more effectively!) be solved in another way.
2021-08-11 23:06:17 +03:00
Roman Khimov
3e60771175 core: deduplicate and simplify processNEP17Transfer a bit
Just refactoring, no functional changes.
2021-08-11 22:36:26 +03:00
Roman Khimov
50ee1a1f91 *: don't use dao.Cached in tests
There is no need to use it.
2021-08-11 21:02:50 +03:00