trace count differs between besu and zkgeth #449

FlorianHuc · 2023-11-30T09:23:39Z

We have been generating the trace count using rpc endpoint of besu for block number 2261111

result is (Note that the result could be improved by adding block number and block hash):

{
   "jsonrpc":"2.0",
   "id":1,
   "result":{
      "tracesEngineVersion":"0.0.1-SNAPSHOT",
      "tracesCountersByModule":{
         "Add":0,
         "Ext":0,
         "LogData":0,
         "LogInfo":0,
         "Mod":0,
         "Mxp":0,
         "RlpTxrcpt":0,
         "Mul":1,
         "Wcp":0,
         "RlpTxn":0,
         "TxnData":0,
         "InstructionDecoder":256,
         "Shf":0,
         "ShfRt":2305,
         "Rom":0,
         "RomLex":0,
         "Hub":0,
         "RlpAddr":0
      }
   }
}

It is different from the trace count returned by zkget which is:

{
    "jsonrpc": "2.0",
    "id": 1,
    "result": {
        "tracesEngineVersion": "0.2.0",
        "blockNumber": "2261111",
        "blockL1Size": "4035",
        "tracesCounters": {
            "ADD": 5490,
            "BIN": 2864,
            "BIN_RT": 0,
            "EC_DATA": 12,
            "EXT": 80,
            "HUB": 8439,
            "INSTRUCTION_DECODER": 0,
            "MMIO": 3864,
            "MMU": 2271,
            "MMU_ID": 0,
            "MOD": 256,
            "MUL": 433,
            "MXP": 2032,
            "PHONEY_RLP": 0,
            "PUB_HASH": 80,
            "PUB_HASH_INFO": 18,
            "PUB_LOG": 16,
            "PUB_LOG_INFO": 4,
            "RLP": 8,
            "ROM": 66722,
            "SHF": 1936,
            "SHF_RT": 0,
            "TX_RLP": 1516,
            "WCP": 2548,
            "BLOCK_TX": 9,
            "BLOCK_L2L1LOGS": 0,
            "BLOCK_KECCAK": 70,
            "PRECOMPILE_ECRECOVER": 0,
            "PRECOMPILE_SHA2": 0,
            "PRECOMPILE_RIPEMD": 0,
            "PRECOMPILE_IDENTITY": 0,
            "PRECOMPILE_MODEXP": 0,
            "PRECOMPILE_ECADD": 0,
            "PRECOMPILE_ECMUL": 0,
            "PRECOMPILE_ECPAIRING": 0,
            "PRECOMPILE_BLAKE2F": 0
        }
    }
}

They should be identical.
Note that Besu gives result that 'looks' better for a more recent block (2314707).

The text was updated successfully, but these errors were encountered:

FlorianHuc · 2023-11-30T09:25:09Z

@ahamlat and @jonesho can add more details on how to reproduce the testing.

delehef · 2023-11-30T09:35:58Z

This looks like the tracing has not been triggered; the lines are (i) ShfRT and InstructionDecoder (which are static and do not depend on the block), and (ii) the one line for mul that is static as well and encodes 0×0 (IIRC).

ahamlat · 2023-11-30T10:07:43Z

Could be related to Bonsai because the results I shared were on a Bonsai/Snap Sync node, I will double check with a forest archive node.

jonesho · 2023-12-01T07:48:59Z

As mentioned in the Slack thread here, it would be great if we can get the full mappings of trace counts from besu to zkgeth (e.g. rlpAddr txnData logData rlpTxRcpt, how do they map to their zkgeth counterparts), and also it would be great if we have consistent naming of the trace count fields here, at least not some as CamelCase and some as underscored cc @delehef

jonesho · 2023-12-01T07:50:07Z

E.g. trace counts from Besu (ec2-18-222-142-230.us-east-2.compute.amazonaws.com):

{
    "add": 25856,
    "ext": 72,
    "instruction-decoder": 256,
    "mod": 7240,
    "mxp": 62266,
    "rlpTxRcpt": 31720,
    "mul": 2520,
    "wcp": 20031,
    "logInfo": 2398,
    "shf": 4832,
    "rom": 3929808,
    "txnData": 1340,
    "shfRT": 2305,
    "logData": 2666,
    "romLex": 520,
    "rlpTxn": 80590,
    "rlpAddr": 92,
    "hub_v2_off": 359392
}

ahamlat · 2023-12-01T15:10:40Z

Testing PR #335 on a Besu Forest Archive node on Linea return an output different from what is expected :

 curl -X POST --data '{"jsonrpc":"2.0","method":"rollup_getTracesCountersByBlockNumberV0","params":["1287000", "0.0.1-SNAPSHOT"],"id":1}' http://localhost:8545
 
 {
   "jsonrpc":"2.0",
   "id":1,
   "result":{
      "tracesEngineVersion":"0.0.1-SNAPSHOT",
      "tracesCountersByModule":{
         "Add":0,
         "Ext":0,
         "LogData":0,
         "LogInfo":0,
         "Mod":0,
         "Mxp":0,
         "RlpTxrcpt":0,
         "Mul":1,
         "Wcp":0,
         "RlpTxn":0,
         "TxnData":0,
         "InstructionDecoder":256,
         "Shf":0,
         "ShfRt":2305,
         "Rom":0,
         "RomLex":0,
         "Hub":0,
         "RlpAddr":0
      }
   }
}

I also see this error when I first test the RPC endpoint

{"@timestamp":"2023-12-01T14:42:19,890","level":"ERROR","thread":"vert.x-worker-thread-7","class":"BlockchainQueries","message":"failed worldstate query for 0x654ae22b7cadb8d85f46b205d1ebdf3233b6321f30ed252835e5c1eaa52d6a2f","throwable":" java.lang.NullPointerException: Cannot invoke \"net.consensys.linea.zktracer.module.hub.DeploymentInfo.number(org.hyperledger.besu.datatypes.Address)\" because the return value of \"net.consensys.linea.zktracer.runtime.stack.ConflationInfo.deploymentInfo()\" is null
at net.consensys.linea.zktracer.module.romLex.RomLex.lambda$traceStartTx$0(RomLex.java:142)
at java.base/java.util.Optional.ifPresent(Optional.java:178)
at net.consensys.linea.zktracer.module.romLex.RomLex.traceStartTx(RomLex.java:139)
at net.consensys.linea.zktracer.module.hub.Hub.traceStartTx(Hub.java:588)
at net.consensys.linea.zktracer.ZkTracer.traceStartTransaction(ZkTracer.java:156)
at org.hyperledger.besu.services.TraceServiceImpl.lambda$trace$6(TraceServiceImpl.java:215)
at java.base/java.util.ArrayList.forEach(ArrayList.java:1511)
at org.hyperledger.besu.services.TraceServiceImpl.trace(TraceServiceImpl.java:202)
at org.hyperledger.besu.services.TraceServiceImpl.lambda$trace$4(TraceServiceImpl.java:182)
at org.hyperledger.besu.ethereum.api.jsonrpc.internal.processor.Tracer.lambda$processTracing$0(Tracer.java:52)
at org.hyperledger.besu.ethereum.api.query.BlockchainQueries.lambda$getAndMapWorldState$24(BlockchainQueries.java:918)
at java.base/java.util.Optional.flatMap(Optional.java:289)
at org.hyperledger.besu.ethereum.api.query.BlockchainQueries.getAndMapWorldState(BlockchainQueries.java:914)
at org.hyperledger.besu.ethereum.api.jsonrpc.internal.processor.Tracer.lambda$processTracing$1(Tracer.java:50)
at java.base/java.util.Optional.flatMap(Optional.java:289)
at org.hyperledger.besu.ethereum.api.jsonrpc.internal.processor.Tracer.processTracing(Tracer.java:48)
at org.hyperledger.besu.ethereum.api.jsonrpc.internal.processor.Tracer.processTracing(Tracer.java:38)
at org.hyperledger.besu.services.TraceServiceImpl.trace(TraceServiceImpl.java:178)
at org.hyperledger.besu.services.TraceServiceImpl.traceBlock(TraceServiceImpl.java:101)
at org.hyperledger.besu.services.TraceServiceImpl.traceBlock(TraceServiceImpl.java:80)
at net.consensys.linea.tracegeneration.LineCountsByBlockCache.getBlockTraces(LineCountsByBlockCache.java:41)
at net.consensys.linea.tracegeneration.rpc.RollupGetTracesCountersByBlockNumberV0.getTracesCounters(RollupGetTracesCountersByBlockNumberV0.java:68)
at net.consensys.linea.tracegeneration.rpc.RollupGetTracesCountersByBlockNumberV0.execute(RollupGetTracesCountersByBlockNumberV0.java:55)
at net.consensys.linea.tracegeneration.rpc.RollupGetTracesCountersByBlockNumberV0.execute(RollupGetTracesCountersByBlockNumberV0.java:26)
at org.hyperledger.besu.ethereum.api.jsonrpc.internal.methods.PluginJsonRpcMethod.response(PluginJsonRpcMethod.java:53)
at org.hyperledger.besu.ethereum.api.jsonrpc.execution.BaseJsonRpcProcessor.process(BaseJsonRpcProcessor.java:44)
at org.hyperledger.besu.ethereum.api.jsonrpc.execution.TracedJsonRpcProcessor.process(TracedJsonRpcProcessor.java:41)
at org.hyperledger.besu.ethereum.api.jsonrpc.execution.TimedJsonRpcProcessor.process(TimedJsonRpcProcessor.java:45)
at org.hyperledger.besu.ethereum.api.jsonrpc.execution.JsonRpcExecutor.execute(JsonRpcExecutor.java:92)
at org.hyperledger.besu.ethereum.api.handlers.AbstractJsonRpcExecutor.executeRequest(AbstractJsonRpcExecutor.java:87)
at org.hyperledger.besu.ethereum.api.handlers.JsonRpcObjectExecutor.execute(JsonRpcObjectExecutor.java:57)
at org.hyperledger.besu.ethereum.api.handlers.JsonRpcExecutorHandler.lambda$handler$0(JsonRpcExecutorHandler.java:57)
at java.base/java.util.Optional.ifPresentOrElse(Optional.java:196)
at org.hyperledger.besu.ethereum.api.handlers.JsonRpcExecutorHandler.lambda$handler$2(JsonRpcExecutorHandler.java:54)
at io.vertx.ext.web.impl.BlockingHandlerDecorator.lambda$handle$0(BlockingHandlerDecorator.java:48)
at io.vertx.core.impl.ContextBase.lambda$null$0(ContextBase.java:137)
at io.vertx.core.impl.ContextInternal.dispatch(ContextInternal.java:264)
at io.vertx.core.impl.ContextBase.lambda$executeBlocking$1(ContextBase.java:135)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
at io.netty.util.concurrent.FastThreadLocalRunnable.run(FastThreadLocalRunnable.java:30)
at java.base/java.lang.Thread.run(Thread.java:833)\n"}

Whe testing with PR #451 with the same block number on the same node, I have this result

{
   "jsonrpc":"2.0",
   "id":1,
   "result":{
      "tracesEngineVersion":"0.0.1-SNAPSHOT",
      "blockNumber":1287000,
      "tracesCounters":{
         "EXT":0,
         "TX_RLP":650,
         "ADD":4768,
         "MXP":3168,
         "PRECOMPILE_ECMUL":0,
         "ROM_LEX":30,
         "PUB_LOG":56,
         "TRM":96,
         "SHF":384,
         "HUB":27711,
         "rlpAddr":0,
         "PRECOMPILE_ECPAIRING":0,
         "PRECOMPILE_MODEXP":0,
         "BLOCK_TX":2,
         "MOD":1904,
         "rlpTxRcpt":522,
         "MUL":933,
         "wcp":2881,
         "PRECOMPILE_ECADD":0,
         "PRECOMPILE_ECRECOVER":0,
         "ROM":128048,
         "txnData":14,
         "PRECOMPILE_SHA2":0,
         "ecpairingWeightedCall":0,
         "PRECOMPILE_BLAKE2F":0,
         "MMU":0,
         "PUB_LOG_INFO":48,
         "PRECOMPILE_RIPEMD":0
      }
   }
}

letypequividelespoubelles · 2023-12-05T04:09:22Z

it would be great if we can get the full mappings of trace counts from besu to zkgeth (e.g. rlpAddr txnData logData rlpTxRcpt, how do they map to their zkgeth counterparts)

rlpAddr (previously called rlp): it changed from 16 rows per CREATE to 8 rows for CREATE and 6 rows for CREATE2.
rlpTxn should be almost the same
logData, logInfo, txnData, rlpTxRcpt were not traced in Geth AFAIK (were specified in august / september)
rom should be a little bit bigger in Besu (<5% I think)
romLex should be the same
for the stateless module (wcp, mul, mod, etc) it should be way bigger in Besu as these modules are called by other modules. So the more implemented module, the bigger they are.
for mxp I think it's stricly the same @lorenzogentile404
for mmu and mmio idk : @powerslider

delehef · 2023-12-11T17:43:30Z

Superseded by #467

delehef closed this as completed Nov 30, 2023

delehef reopened this Nov 30, 2023

delehef mentioned this issue Nov 30, 2023

feat: implement RPC for line counting #451

Merged

ahamlat mentioned this issue Dec 1, 2023

RPC endpoint to get traces counters from ZkTracer #335

Closed

powerslider assigned delehef Dec 4, 2023

delehef mentioned this issue Dec 6, 2023

fix: line count discrepancy #468

Merged

delehef closed this as completed Dec 11, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

trace count differs between besu and zkgeth #449

trace count differs between besu and zkgeth #449

FlorianHuc commented Nov 30, 2023

FlorianHuc commented Nov 30, 2023

delehef commented Nov 30, 2023

ahamlat commented Nov 30, 2023 •

edited

Loading

jonesho commented Dec 1, 2023

jonesho commented Dec 1, 2023 •

edited

Loading

ahamlat commented Dec 1, 2023 •

edited

Loading

letypequividelespoubelles commented Dec 5, 2023

delehef commented Dec 11, 2023

trace count differs between besu and zkgeth #449

trace count differs between besu and zkgeth #449

Comments

FlorianHuc commented Nov 30, 2023

FlorianHuc commented Nov 30, 2023

delehef commented Nov 30, 2023

ahamlat commented Nov 30, 2023 • edited Loading

jonesho commented Dec 1, 2023

jonesho commented Dec 1, 2023 • edited Loading

ahamlat commented Dec 1, 2023 • edited Loading

letypequividelespoubelles commented Dec 5, 2023

delehef commented Dec 11, 2023

ahamlat commented Nov 30, 2023 •

edited

Loading

jonesho commented Dec 1, 2023 •

edited

Loading

ahamlat commented Dec 1, 2023 •

edited

Loading