perf(levm): improve sstore perfomance further #3657

edg-l · 2025-07-16T14:09:13Z

Motivation
Improves sstore perfomance

Requires #3564

From 1100 to over 2200

The main change is going from Hashmaps to BTreeMaps.

They are more efficient for the type of storages we use, for small datasets (1k~100k i would say) they overperform hashmaps due to avoiding entirely the hashing cost, which seemed to be the biggest factor.

This changes comes with 2 other minor changes, like a more efficient u256 to big endian and a change to backup_storage_slot.

…fe_stack

…ents

github-actions · 2025-07-16T14:11:30Z

Lines of code report

Total lines added: 42
Total lines removed: 6
Total lines changed: 48

Detailed view

+------------------------------------------------------------------------+-------+------+
| File                                                                   | Lines | Diff |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/common/utils.rs                                          | 57    | +28  |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/storage/store_db/libmdbx.rs                              | 1478  | +1   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/bench/revm_comparison/src/levm_bench.rs          | 70    | +1   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/src/call_frame.rs                                | 296   | +3   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/src/db/gen_db.rs                                 | 241   | -6   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/src/memory.rs                                    | 276   | +3   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/src/opcode_handlers/stack_memory_storage_flow.rs | 284   | +5   |
+------------------------------------------------------------------------+-------+------+
| ethrex/crates/vm/levm/src/utils.rs                                     | 445   | +1   |
+------------------------------------------------------------------------+-------+------+

github-actions · 2025-07-16T14:28:52Z

Benchmark Results Comparison

Benchmark Results: SstoreBench_no_opt

Command	Mean [s]	Min [s]	Max [s]	Relative
`main_revm_SstoreBench_no_opt`	161.5 ± 4.3	158.4	172.7	1.59 ± 0.09
`main_levm_SstoreBench_no_opt`	114.8 ± 7.0	104.7	124.1	1.13 ± 0.09
`pr_levm_SstoreBench_no_opt`	101.7 ± 4.7	98.8	114.9	1.00

Detailed Results

Benchmark Results: BubbleSort

Command	Mean [s]	Min [s]	Max [s]	Relative
`main_revm_BubbleSort`	3.204 ± 0.013	3.187	3.228	1.00
`main_levm_BubbleSort`	4.277 ± 0.016	4.253	4.301	1.33 ± 0.01
`pr_revm_BubbleSort`	3.211 ± 0.017	3.193	3.238	1.00 ± 0.01
`pr_levm_BubbleSort`	4.170 ± 0.024	4.149	4.228	1.30 ± 0.01

Benchmark Results: ERC20Approval

Command	Mean [s]	Min [s]	Max [s]	Relative
`main_revm_ERC20Approval`	1.055 ± 0.023	1.038	1.114	1.01 ± 0.02
`main_levm_ERC20Approval`	1.434 ± 0.014	1.416	1.466	1.37 ± 0.02
`pr_revm_ERC20Approval`	1.048 ± 0.011	1.035	1.069	1.00
`pr_levm_ERC20Approval`	1.419 ± 0.016	1.404	1.458	1.35 ± 0.02

Benchmark Results: ERC20Mint

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_ERC20Mint`	138.2 ± 1.1	137.1	140.7	1.00
`main_levm_ERC20Mint`	220.1 ± 2.9	216.2	224.7	1.59 ± 0.03
`pr_revm_ERC20Mint`	138.5 ± 0.6	137.4	139.5	1.00 ± 0.01
`pr_levm_ERC20Mint`	206.4 ± 3.1	204.0	213.6	1.49 ± 0.03

Benchmark Results: ERC20Transfer

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_ERC20Transfer`	244.3 ± 1.1	242.7	246.9	1.00
`main_levm_ERC20Transfer`	359.9 ± 1.3	357.8	362.8	1.47 ± 0.01
`pr_revm_ERC20Transfer`	245.1 ± 3.2	242.6	252.4	1.00 ± 0.01
`pr_levm_ERC20Transfer`	358.5 ± 7.8	354.8	379.7	1.47 ± 0.03

Benchmark Results: Factorial

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_Factorial`	229.6 ± 1.6	228.3	232.8	1.00 ± 0.01
`main_levm_Factorial`	451.7 ± 3.6	443.8	457.4	1.97 ± 0.02
`pr_revm_Factorial`	229.5 ± 1.1	228.0	231.6	1.00
`pr_levm_Factorial`	457.4 ± 11.9	449.3	486.9	1.99 ± 0.05

Benchmark Results: FactorialRecursive

Command	Mean [s]	Min [s]	Max [s]	Relative
`main_revm_FactorialRecursive`	1.596 ± 0.020	1.550	1.624	1.01 ± 0.06
`main_levm_FactorialRecursive`	8.208 ± 0.052	8.140	8.311	5.21 ± 0.32
`pr_revm_FactorialRecursive`	1.576 ± 0.097	1.308	1.648	1.00
`pr_levm_FactorialRecursive`	8.246 ± 0.027	8.209	8.296	5.23 ± 0.32

Benchmark Results: Fibonacci

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_Fibonacci`	204.0 ± 1.8	202.6	208.7	1.00
`main_levm_Fibonacci`	467.5 ± 18.8	452.2	507.3	2.29 ± 0.09
`pr_revm_Fibonacci`	206.9 ± 10.2	202.9	235.9	1.01 ± 0.05
`pr_levm_Fibonacci`	469.7 ± 15.7	449.1	499.2	2.30 ± 0.08

Benchmark Results: FibonacciRecursive

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_FibonacciRecursive`	850.0 ± 13.1	825.0	866.6	1.00
`main_levm_FibonacciRecursive`	1315.8 ± 15.2	1288.1	1338.7	1.55 ± 0.03
`pr_revm_FibonacciRecursive`	860.0 ± 17.6	844.1	902.4	1.01 ± 0.03
`pr_levm_FibonacciRecursive`	1327.1 ± 18.7	1302.0	1351.2	1.56 ± 0.03

Benchmark Results: ManyHashes

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_ManyHashes`	8.7 ± 0.1	8.6	8.9	1.00
`main_levm_ManyHashes`	12.4 ± 0.1	12.3	12.6	1.42 ± 0.01
`pr_revm_ManyHashes`	8.8 ± 0.1	8.7	8.9	1.00 ± 0.01
`pr_levm_ManyHashes`	12.0 ± 0.1	11.9	12.3	1.38 ± 0.02

Benchmark Results: MstoreBench

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_MstoreBench`	265.8 ± 2.3	263.8	271.3	1.00
`main_levm_MstoreBench`	639.1 ± 4.4	636.2	651.3	2.40 ± 0.03
`pr_revm_MstoreBench`	265.8 ± 2.4	263.0	271.6	1.00 ± 0.01
`pr_levm_MstoreBench`	640.5 ± 6.8	636.1	656.8	2.41 ± 0.03

Benchmark Results: Push

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_Push`	291.5 ± 2.0	289.5	296.3	1.00 ± 0.01
`main_levm_Push`	819.4 ± 10.2	810.6	838.8	2.81 ± 0.04
`pr_revm_Push`	291.2 ± 0.9	290.0	292.6	1.00
`pr_levm_Push`	815.1 ± 7.5	809.1	834.7	2.80 ± 0.03

Benchmark Results: SstoreBench_no_opt

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
`main_revm_SstoreBench_no_opt`	161.5 ± 4.3	158.4	172.7	1.59 ± 0.09
`main_levm_SstoreBench_no_opt`	114.8 ± 7.0	104.7	124.1	1.13 ± 0.09
`pr_revm_SstoreBench_no_opt`	159.8 ± 1.1	158.5	161.8	1.57 ± 0.07
`pr_levm_SstoreBench_no_opt`	101.7 ± 4.7	98.8	114.9	1.00

Arkenan

LGTM

crates/common/utils.rs

github-actions · 2025-07-17T16:23:55Z

Benchmark Block Execution Results Comparison Against Main

Command	Mean [s]	Min [s]	Max [s]	Relative
`base`	207.215 ± 0.407	206.393	207.752	1.00
`head`	209.377 ± 0.433	208.682	209.989	1.01 ± 0.00

edg-l added 30 commits July 8, 2025 17:39

new memory progress

2680471

done

4a01b9e

Merge branch 'main' into new_memory

e7c80b7

fix

4387a0c

fix

2fa2694

improve

a7c23fe

Merge branch 'main' into new_memory

ee33f25

better

62ef32f

fix mcopy

8f5ed9d

fill instead of truncate

b44d5bc

remove dbg

f253bab

fix

72e08b6

better handling

74a289a

improve

7a4f252

fix and better perf

cb33cee

Merge remote-tracking branch 'origin/main' into new_memory

de4520d

Merge branch 'main' into new_memory

87a255b

fixes

37d71b8

changelog

f8a0c49

remove unused

a710c04

rename

1ba46a3

docs

9de7221

remove unwraps

b11af48

better

ace118c

improve

5e7efd3

force inline to improve perfomance

34d4474

unsafe

2a97377

more unsafe

10df4fb

unsafe

47cfc11

fmt

b025e3d

edg-l added 11 commits July 15, 2025 11:06

experimental

981da2c

Merge branch 'main' into new_memory

4d67f60

Merge remote-tracking branch 'origin/new_memory' into new_memory_unsa…

d54da5b

…fe_stack

Merge branch 'main' into new_memory

b0aad27

Merge branch 'main' into new_memory

4299eb0

Merge branch 'main' into new_memory

44a4a3f

Merge branch 'main' into new_memory

f332dd1

small improvement

b80c981

Merge remote-tracking branch 'origin/new_memory' into sstore_improvem…

40dc4c5

…ents

add custom u256 to big endian

0e9e449

use btreemaps for better perfomance

a0679b3

edg-l requested a review from a team as a code owner July 16, 2025 14:09

github-actions bot assigned edg-l Jul 16, 2025

github-actions bot added the performance label Jul 16, 2025

github-project-automation bot added this to ethrex_performance Jul 16, 2025

github-project-automation bot moved this to Todo in ethrex_performance Jul 16, 2025

edg-l moved this from Todo to In review in ethrex_performance Jul 17, 2025

Arkenan approved these changes Jul 17, 2025

View reviewed changes

crates/common/utils.rs Show resolved Hide resolved

Base automatically changed from new_memory to main July 17, 2025 14:19

edg-l added 2 commits July 17, 2025 16:29

Merge remote-tracking branch 'origin/main' into sstore_improvements

699f02d

changelog

b369fb4

Oppen approved these changes Jul 17, 2025

View reviewed changes

Arkenan added this pull request to the merge queue Jul 17, 2025

Merged via the queue into main with commit 23191af Jul 17, 2025
48 checks passed

Arkenan deleted the sstore_improvements branch July 17, 2025 17:43

github-project-automation bot moved this from In review to Done in ethrex_performance Jul 17, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

perf(levm): improve sstore perfomance further #3657

perf(levm): improve sstore perfomance further #3657

Uh oh!

edg-l commented Jul 16, 2025

Uh oh!

github-actions bot commented Jul 16, 2025 •

edited

Loading

Uh oh!

github-actions bot commented Jul 16, 2025 •

edited

Loading

Benchmark Results: BubbleSort

Benchmark Results: ERC20Approval

Benchmark Results: ERC20Mint

Benchmark Results: ERC20Transfer

Benchmark Results: Factorial

Benchmark Results: FactorialRecursive

Benchmark Results: Fibonacci

Benchmark Results: FibonacciRecursive

Benchmark Results: ManyHashes

Benchmark Results: MstoreBench

Benchmark Results: Push

Benchmark Results: SstoreBench_no_opt

Uh oh!

Arkenan left a comment

Uh oh!

Uh oh!

github-actions bot commented Jul 17, 2025 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

perf(levm): improve sstore perfomance further #3657

perf(levm): improve sstore perfomance further #3657

Uh oh!

Conversation

edg-l commented Jul 16, 2025

Uh oh!

github-actions bot commented Jul 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Lines of code report

Uh oh!

github-actions bot commented Jul 16, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmark Results Comparison

Benchmark Results: SstoreBench_no_opt

Benchmark Results: BubbleSort

Benchmark Results: ERC20Approval

Benchmark Results: ERC20Mint

Benchmark Results: ERC20Transfer

Benchmark Results: Factorial

Benchmark Results: FactorialRecursive

Benchmark Results: Fibonacci

Benchmark Results: FibonacciRecursive

Benchmark Results: ManyHashes

Benchmark Results: MstoreBench

Benchmark Results: Push

Benchmark Results: SstoreBench_no_opt

Uh oh!

Arkenan left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions bot commented Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Benchmark Block Execution Results Comparison Against Main

Uh oh!

Uh oh!

Uh oh!

github-actions bot commented Jul 16, 2025 •

edited

Loading

github-actions bot commented Jul 16, 2025 •

edited

Loading

github-actions bot commented Jul 17, 2025 •

edited

Loading