jeirewired-1.2.0-omni-1.20.1.jar
Curse Maven Snippet
What's new
JEI Rewired V1.2.0-RELEASE
Big search and recipe perf pass :]
Startup and recipe performance
Compared with the older Rewired build. All timings are milliseconds.
| Operation | Before | Now | Change |
|---|---|---|---|
| Runtime publication | 2,519.14 | 1,982.85 | 21.3% less time |
| All fixture recipes usable | 2,532.60 | 2,009.91 | 20.6% less time |
| First search | 12.26 | 10.64 | 13.2% less time |
| Readiness + first search | 2,544.87 | 2,020.55 | 20.6% less time |
| Repeated 24-query cycle | 24.16 | 9.40 | 61.1% less time |
| Query-cycle P95 | 27.27 | 11.91 | 56.3% less time |
| All custom recipes, 16 lookups | 8.070 | 3.929 | 51.3% less time |
| Focused custom recipes, 64 lookups | 0.829 | 0.507 | 38.8% less time |
| Focused categories, 64 lookups | 0.337 | 0.150 | 55.5% less time |
| Custom layouts, 32 builds | 0.074 | 0.053 | 28.4% less time |
| Open custom recipe screen | 1.761 | 0.975 | 44.6% less time |
| All native recipes, 4 lookups | 29.502 | 8.661 | 70.6% less time |
| Native input focus, 64 lookups | 2.418 | 1.125 | 53.5% less time |
| Native output focus, 64 lookups | 0.309 | 0.209 | 32.4% less time |
| Item categories, 64 lookups | 1.206 | 0.867 | 28.1% less time |
| Native layouts, 32 builds | 0.256 | 0.207 | 19.2% less time |
| Open focused-item screen | 1.759 | 1.270 | 27.8% less time |
| Cold custom recipe registration | 28.72 | 16.18 | 43.7% less time |
Every measured search query
Warmed query timings, in milliseconds.
| Query | Before | Now | Time reduction |
|---|---|---|---|
| Empty filter | 2.24970 | 1.95595 | 13.1% |
alloy |
0.48610 | 0.34370 | 29.3% |
copper |
0.35450 | 0.27560 | 22.3% |
copper plate |
0.65730 | 0.09890 | 85.0% |
steel |
0.70060 | 0.25950 | 63.0% |
steel gear |
0.15700 | 0.08645 | 44.9% |
crystal |
0.73280 | 0.27105 | 63.0% |
crystal shard |
0.14720 | 0.08630 | 41.4% |
gold |
0.70180 | 0.27550 | 60.7% |
gold casing |
0.48670 | 0.08870 | 81.8% |
iron |
0.70940 | 0.28025 | 60.5% |
iron rod |
0.28160 | 0.08885 | 68.4% |
module |
3.48930 | 1.54735 | 55.7% |
module 12 |
0.31900 | 0.26725 | 16.2% |
alloy plate |
0.92090 | 0.09050 | 90.2% |
copper gear |
0.81110 | 0.09755 | 88.0% |
steel shard |
0.79780 | 0.09870 | 87.6% |
crystal casing |
0.56450 | 0.09990 | 82.3% |
gold rod |
0.63730 | 0.08965 | 85.9% |
iron module |
2.22930 | 0.36550 | 83.6% |
@matchedbench |
3.01030 | 1.52130 | 49.5% |
alloy -copper |
1.42640 | 0.42220 | 70.4% |
gear | shard |
1.59150 | 0.53555 | 66.3% |
unmatched987654321 |
0.02990 | 0.01410 | 52.8% |
Memory and allocations
Heap numbers cover the whole test client. Registration allocations are temporary allocations on the calling thread, not retained RAM!!!!!
| Measurement | Before | Now | Reduction |
|---|---|---|---|
| Client heap after GC | 545.34 MiB | 538.71 MiB | 1.22% |
| Client heap at capture end | 545.35 MiB | 540.71 MiB | 0.85% |
| Native registration allocation | 253.72 MiB | 248.18 MiB | 2.18% |
| Custom registration allocation | 14.31 MiB | 10.72 MiB | 25.13% |
| Combined registration allocation | 268.03 MiB | 258.90 MiB | 3.41% |
Sorting and tokenizer work
| Component | Before | After | Reduction |
|---|---|---|---|
| Tooltip tokenization | 3.309 ms | 2.968 ms | 10.31% |
| Sorting, 8 configured mods | 4.515 ms | 2.367 ms | 47.57% |
| Sorting, 256 configured mods | 74.691 ms | 3.428 ms | 95.41% |
| Tokenizer temporary allocation | 13,180,896 B | 8,576,000 B | 34.94% |
This took me way too much pain xd, but hopefully it serves well :DDD
This mod has no additional files

