Optimize divrem with non-zero immediate rhs values

[Robbepop]

Closes #807.

[paritytech-cicd-pr]

BENCHMARKS

	NATIVE			WASMTIME
BENCHMARK	MASTER	PR	DIFF	MASTER	PR	DIFF	WASMTIME OVERHEAD
execute/ br_table	1.51ms	1.49ms	⚪ -1.35%	1.34ms	1.29ms	🟢 -3.45%	🟢 -13%
execute/ call/host/1	44.99µs	45.02µs	⚪ -0.37%	66.55µs	65.10µs	🟢 -2.25%	🟢 45%
execute/ call/rec	168.54µs	167.69µs	⚪ -0.33%	351.45µs	355.22µs	⚪ 0.87%	🔴 112%
execute/ count_until	7.47ms	6.59ms	🟢 -10.25%	7.50ms	7.49ms	⚪ -0.17%	🟢 14%
execute/ divrem	6.43ms	6.21ms	🟢 -3.33%	6.81ms	6.97ms	🔴 2.34%	🟢 12%
execute/ factorial/iter	231.80µs	237.10µs	🔴 2.95%	323.14µs	316.27µs	🟢 -2.18%	🟢 33%
execute/ factorial/rec	697.94µs	681.03µs	🟢 -2.39%	1.38ms	1.25ms	🟢 -9.34%	🟡 84%
execute/ fibonacci/iter	1.36ms	1.36ms	⚪ 0.20%	1.19ms	1.27ms	🔴 6.65%	🟢 -7%
execute/ fibonacci/rec	6.25ms	6.07ms	🟢 -2.83%	13.28ms	13.22ms	⚪ -0.68%	🔴 118%
execute/ fibonacci/tail	1.63ms	1.50ms	🟢 -8.07%	3.79ms	4.53ms	🔴 19.33%	🔴 202%
execute/ fuse	8.25ms	7.35ms	🟢 -11.11%	11.50ms	11.64ms	⚪ 1.16%	🟡 58%
execute/ global/bump	1.33ms	1.32ms	⚪ -0.54%	1.59ms	1.54ms	🟢 -3.35%	🟢 17%
execute/ global/get_const	735.74µs	711.99µs	🟢 -3.23%	761.80µs	750.85µs	🟢 -1.46%	🟢 5%
execute/ is_even/rec	1.10ms	1.08ms	🟢 -1.95%	2.21ms	2.27ms	🔴 2.75%	🔴 111%
execute/ memory/fill_bytes	1.13ms	1.08ms	🟢 -4.68%	1.23ms	1.31ms	🔴 6.64%	🟢 22%
execute/ memory/sum_bytes	1.13ms	1.10ms	🟢 -3.39%	1.24ms	1.23ms	⚪ -0.95%	🟢 12%
execute/ memory/vec_add	3.01ms	2.96ms	⚪ -3.83%	3.41ms	3.62ms	🔴 6.50%	🟢 23%
execute/ recursive_scan	186.81µs	187.47µs	⚪ 0.37%	373.02µs	390.69µs	🔴 4.75%	🔴 108%
execute/ recursive_trap	16.11µs	15.32µs	🟢 -4.86%	35.34µs	35.33µs	⚪ -0.17%	🔴 131%
execute/ regex_redux	592.56µs	604.74µs	🔴 2.33%	1.05ms	1.06ms	⚪ 0.73%	🟡 76%
execute/ rev_complement	440.62µs	442.82µs	⚪ 0.55%	635.12µs	640.62µs	⚪ 0.77%	🟢 45%
execute/ tiny_keccak	368.83µs	349.83µs	🟢 -4.78%	418.06µs	381.67µs	🟢 -8.66%	🟢 9%
execute/ trunc_f2i	620.25µs	616.44µs	⚪ -0.60%	943.61µs	953.94µs	⚪ 1.00%	🟡 55%
instantiate/ wasm_kernel	54.59µs	56.61µs	🔴 3.62%	54.07µs	55.94µs	🔴 4.46%	🟢 -1%
overhead/ call/typed/0	1.21ms	1.23ms	🔴 1.91%	847.77µs	777.94µs	🟢 -8.23%	🟢 -37%
overhead/ call/typed/16	1.73ms	1.63ms	🟢 -5.83%	2.07ms	1.94ms	🟢 -6.61%	🟢 19%
overhead/ call/untyped/0	1.61ms	1.62ms	⚪ 0.80%	1.38ms	1.23ms	🟢 -11.06%	🟢 -24%
overhead/ call/untyped/16	2.53ms	2.46ms	🟢 -2.63%	3.95ms	3.97ms	⚪ 0.44%	🟡 61%
translate/ bz2/default	1.38ms	1.38ms	⚪ -1.08%	2.42ms	2.46ms	⚪ 1.40%	🟡 79%
translate/ bz2/fuel	1.43ms	1.42ms	⚪ -0.90%	2.58ms	2.61ms	⚪ 0.61%	🟡 84%
translate/ erc1155/default	282.58µs	279.91µs	⚪ -0.59%	476.69µs	470.62µs	⚪ -1.24%	🟡 68%
translate/ erc1155/fuel	301.19µs	299.36µs	⚪ -0.30%	503.03µs	501.65µs	⚪ -0.36%	🟡 68%
translate/ erc20/default	136.00µs	134.27µs	⚪ -1.15%	222.73µs	224.29µs	⚪ 1.01%	🟡 67%
translate/ erc20/fuel	143.84µs	142.60µs	⚪ -1.15%	234.81µs	235.66µs	⚪ 0.60%	🟡 65%
translate/ erc721/default	192.09µs	191.18µs	⚪ -0.51%	321.57µs	324.51µs	⚪ 0.91%	🟡 70%
translate/ erc721/fuel	202.45µs	201.59µs	⚪ -0.48%	339.60µs	339.35µs	⚪ -0.63%	🟡 68%
translate/ pulldown_cmark/default	3.84ms	3.81ms	⚪ -0.66%	6.34ms	6.40ms	⚪ 0.93%	🟡 68%
translate/ pulldown_cmark/fuel	3.94ms	3.93ms	⚪ -0.20%	6.64ms	6.70ms	⚪ 0.96%	🟡 70%
translate/ spidermonkey/default	0.00ns	0.00ns	⚪ -0.27%	0.00ns	0.00ns	⚪ 1.26%	🟢 0%
translate/ spidermonkey/fuel	0.00ns	0.00ns	⚪ -0.58%	0.00ns	0.00ns	⚪ 0.73%	🟢 0%
translate/ wasm_kernel/default	5.07ms	5.04ms	⚪ -0.57%	8.47ms	8.58ms	🔴 1.19%	🟡 70%
translate/ wasm_kernel/fuel	5.21ms	5.21ms	⚪ 0.19%	8.94ms	8.91ms	⚪ -0.28%	🟡 71%

Link to pipeline

[codecov-commenter]

Codecov Report

Attention: 19 lines in your changes are missing coverage. Please review.

Comparison is base (42371ac) 80.88% compared to head (4633081) 81.00%.

Files	Patch %	Lines
...rates/wasmi/src/engine/translator/relink_result.rs	0.00%	14 Missing ⚠️
crates/wasmi/src/engine/executor/instrs/binary.rs	86.66%	4 Missing ⚠️
crates/wasmi/src/engine/executor/instrs.rs	90.90%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #825      +/-   ##
==========================================
+ Coverage   80.88%   81.00%   +0.11%     
==========================================
  Files         255      255              
  Lines       22344    22390      +46     
==========================================
+ Hits        18074    18138      +64     
+ Misses       4270     4252      -18     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Robbepop]

Locally and on CI we can see roughly 3-4% performance improvement for the new divrem benchmark that exclusively tests the new optimization. This is way less than anticipated by still significant. The changes introduced in this PR are not complex and we may get more performance out of this if we succeed to optimize instruction dispatch since instruction dispatch clearly is the bottleneck in the divrem benchmark test.