Our current EVM loop splits out cost, execution and halt evaluation. Flattening these three into one method results in a significant speedup in reference tests, which are EVM heavy.
Mostly the cost, exceptionalHaltCondition, and execute are merged into one method, sharing calculations between the three.
In fixed cost operations the merger is very simple AddOperation is a representative fixed cost operation.
Check the gas, check other exceptional halts, do the work, return the result.
In variable cost operations there is some value reads to be done before the cost is calculated. This is where a lot of the de-duplication occurs. CodeCopyOperation is a representative variable cost operation.
JumpIOperation shows where this merger pays off. If the condition is zero then some exceptional halts don't need to be considered. But with the three way split each step couldn't consider such optimizations because the local data was lost between each call.
Some cleanup was enabled by this. The old exceptional halt predicates were deleted and moved into each operation. Gas costs must be checked by the operation instead of globally, or we would lose state if we had to split into two methods and do the gas check shared.
The OperandStack was flattened into a single class instead of an interface plus a single implementation. stack underflow and overflow are signaled via named exceptions and handled via catches instead of pre-checking the stack height. Since overflow/underflow is exceedingly rare in mainnet transactions java exceptions are the more performant means.
Some of the APIs had lingering impacts on how some tests were run and the EVMTool (we cache operation cost in the message frame now.)
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
Internally we used to use an enum set to track halt reasons and we would
track multiple halt reasons. However, what the halt reason is does not
matter to reference tests and tracing, only that a halt occurred.
Repalace the EnumSet with an Optional and trace only one revert reason.
This saves us time in enumset management and also allows us to return
quicker once any halt is detected.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
ProtocolContext uses a generic for the consensus state, which has a very
large footprint across the code to accomplish what it intends to
accomplish. For every call there are about 61 other lines per call that
need to be updated, over 1300 lines total.
Instead replace it with java.lang.Class#cast, which provides runtime
security, and use generics to provide the compile time sugar that
allows for chained methods of the appropriate type. Then remove its
(quite large) footprint from the rest of the code.
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* Support clique genesis (needs a nodekey)
* Support nested calls
* Write genesis allocations to world state
* Report total gas (EVM operations + intrinsic cost of TX)
Signed-off-by: Danno Ferrin <danno.ferrin@gmail.com>
* [minor] distribution tweaks
* the 'pantheon:client' project is empty, and results in an empty jar, delete
* clique hard codes a version, delete so the parent takes effect
* clique does not describe it's jar, add description
* evmtools does not describe it's jar, add description.
results in
* clique-1.0.0-SNAPSHOT -> pantheon-clique-0.8.0-RC
* evmtools-0.8.0-SNAPSHOT -> pantheon-evmtools-0.8.0-RC
* client-0.8.0-SNAPSHOT -> *deleted*
* spotless formatting on the gradle file
Signed-off-by: Adrian Sutton <adrian.sutton@consensys.net>
Danno Ferrin
Lucas Saldanha
Danno Ferrin