perf: Int::decode small-value fast path by sasa-tomic · Pull Request #717 · dfinity/candid

sasa-tomic · 2026-03-13T14:05:13Z

Summary

Decode SLEB128-encoded Int values into i64 first, only falling back to BigInt for values that exceed 64-bit range. Mirrors the existing Nat::decode fast path from an earlier commit in this series.

Fast path: accumulates SLEB128 bytes into i64 with proper sign extension
Fallback: converts accumulated i64 to BigInt and continues with BigInt arithmetic
Most real-world Int values fit in i64, avoiding heap allocation entirely

Benchmark (canbench, wasm32)

Metric	Before	After	Change
option_list Decoding	26.2M inst	23.2M inst	-11.7%
variant_list Decoding	25.1M inst	22.1M inst	-11.9%
option_list Heap	2 pages	1 page	-50%
variant_list Heap	2 pages	1 page	-50%

No regressions on any of the 9 benchmarks.

Compatibility

Wire format is unchanged. Additional edge-case values added to the round-trip compatibility test (Int(0), Int(i64::MAX), large positive Int).

Relates to #710

github-actions · 2026-03-13T14:10:21Z

Name	Max Mem (Kb)	Encode	Decode
blob	4_224	4_207_487	2_122_465
btreemap	73_856	531_975_943	13_058_092_273
nns	192	2_021_253	5_675_179 ($\textcolor{red}{0.04\%}$)
nns_list_proposal	1_216	7_016_131 ($\textcolor{red}{0.02\%}$)	64_370_369 ($\textcolor{red}{0.11\%}$)
option_list	64	715_981	22_086_226 ($\textcolor{red}{1.31\%}$)
text	6_336	4_204_384	7_877_792
variant_list	64	710_969	20_730_750 ($\textcolor{red}{0.67\%}$)
vec_int16	12_480	8_404_689	408_970_132

Parser cost: 16_179_361
Extra args: 2_838_484 ($\textcolor{red}{0.00\%}$)

Click to see raw report

---------------------------------------------------

Benchmark: blob
  total:
    instructions: 6.33 M (no change)
    heap_increase: 66 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 4.21 M (no change)
    heap_increase: 66 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 2.12 M (no change)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: btreemap
  total:
    instructions: 13.59 B (no change)
    heap_increase: 1154 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 531.98 M (no change)
    heap_increase: 159 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 13.06 B (no change)
    heap_increase: 995 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: extra_args
  total:
    instructions: 2.84 M (0.00%) (change within noise threshold)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: nns
  total:
    instructions: 24.71 M (0.01%) (change within noise threshold)
    heap_increase: 3 pages (no change)
    stable_memory_increase: 0 pages (no change)

  0. Parsing (scope):
    calls: 1 (no change)
    instructions: 16.18 M (no change)
    heap_increase: 3 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 2.02 M (no change)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 5.68 M (0.04%) (change within noise threshold)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: nns_list_proposal
  total:
    instructions: 71.39 M (0.10%) (change within noise threshold)
    heap_increase: 19 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 7.02 M (0.02%) (change within noise threshold)
    heap_increase: 5 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 64.37 M (0.11%) (change within noise threshold)
    heap_increase: 14 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: option_list
  total:
    instructions: 22.80 M (1.27%) (change within noise threshold)
    heap_increase: 1 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 715.98 K (no change)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 22.09 M (1.31%) (change within noise threshold)
    heap_increase: 1 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: text
  total:
    instructions: 12.08 M (no change)
    heap_increase: 99 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 4.20 M (no change)
    heap_increase: 66 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 7.88 M (no change)
    heap_increase: 33 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: variant_list
  total:
    instructions: 21.44 M (0.65%) (change within noise threshold)
    heap_increase: 1 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 710.97 K (no change)
    heap_increase: 0 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 20.73 M (0.67%) (change within noise threshold)
    heap_increase: 1 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Benchmark: vec_int16
  total:
    instructions: 417.38 M (no change)
    heap_increase: 195 pages (no change)
    stable_memory_increase: 0 pages (no change)

  1. Encoding (scope):
    calls: 1 (no change)
    instructions: 8.40 M (no change)
    heap_increase: 130 pages (no change)
    stable_memory_increase: 0 pages (no change)

  2. Decoding (scope):
    calls: 1 (no change)
    instructions: 408.97 M (no change)
    heap_increase: 65 pages (no change)
    stable_memory_increase: 0 pages (no change)

---------------------------------------------------

Summary:
  instructions:
    status:   No significant changes 👍
    counts:   [total 9 | regressed 0 | improved 0 | new 0 | unchanged 9]
    change:   [max +285.95K | p75 +72.95K | median +72 | p25 0 | min 0]
    change %: [max +1.27% | p75 +0.10% | median 0.00% | p25 0.00% | min 0.00%]

  heap_increase:
    status:   No significant changes 👍
    counts:   [total 9 | regressed 0 | improved 0 | new 0 | unchanged 9]
    change:   [max 0 | p75 0 | median 0 | p25 0 | min 0]
    change %: [max 0.00% | p75 0.00% | median 0.00% | p25 0.00% | min 0.00%]

  stable_memory_increase:
    status:   No significant changes 👍
    counts:   [total 9 | regressed 0 | improved 0 | new 0 | unchanged 9]
    change:   [max 0 | p75 0 | median 0 | p25 0 | min 0]
    change %: [max 0.00% | p75 0.00% | median 0.00% | p25 0.00% | min 0.00%]

---------------------------------------------------
Successfully persisted results to canbench_results.yml

Decode SLEB128-encoded Int values into i64 first, only falling back to BigInt for values that exceed 64-bit range. Mirrors the existing Nat::decode fast path. Benchmark: option_list decoding 26.2M → 23.2M instructions (-11.7%), variant_list decoding 25.1M → 22.1M instructions (-11.9%), heap usage halved for both. Wire format is unchanged. Made-with: Cursor

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

sasa-tomic requested a review from a team as a code owner March 13, 2026 14:05

lwshang changed the base branch from sat-perf-improvements to master March 15, 2026 17:49

sasa-tomic and others added 2 commits March 15, 2026 14:05

refactor: inline leb128_fits_u64 into its only call site

db3bdea

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

lwshang force-pushed the sat-perf-int-decode branch from a17141e to db3bdea Compare March 15, 2026 18:09

lwshang approved these changes Mar 15, 2026

View reviewed changes

lwshang merged commit 325b275 into master Mar 15, 2026
11 checks passed

lwshang deleted the sat-perf-int-decode branch March 15, 2026 18:11

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

perf: Int::decode small-value fast path#717

perf: Int::decode small-value fast path#717
lwshang merged 2 commits intomasterfrom
sat-perf-int-decode

sasa-tomic commented Mar 13, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Mar 13, 2026 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

sasa-tomic commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Benchmark (canbench, wasm32)

Compatibility

Uh oh!

github-actions Bot commented Mar 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

sasa-tomic commented Mar 13, 2026 •

edited

Loading

github-actions Bot commented Mar 13, 2026 •

edited

Loading