volas

github.com/kaelzhang/volas

A Rust-backed, pandas-shaped DataFrame for live OHLCV pipelines: a broad trading-indicator set, incremental O(lookback) refresh on each new bar, and NumPy/Torch-ready output. This page is the project's reproducible benchmark report — regenerated from make benchmark on each release. See the repository for docs and installation.On this run, volas beats TA-Lib on 150 / 158 covered indicators by the default ratio, using the full coverage table below.

Benchmark report

OHLCV technical-indicator computation across libraries · 2026-09-07T10:56:47.508964+00:00 · Apple M1 (Virtual) · Python 3.12.10
Lower time is better; Perf shows the slowest candidate as 1.00× and each faster candidate as {slowest ÷ this}×.

pandaspandas_tastock_pandaspolarstalibvolas

Windowed live stream — 1m bars → 15m, bounded window (per-bar)

The live loop: fold a new 1-minute bar into a 15-minute series and refresh atr:14 over a bounded 30-bar window, every bar, for 20 000 bars. volas does this natively (a windowed, tf-aware frame that auto-compacts, so memory stays O(window + lookback)); the other libraries have no such primitive, so each is given a hand-written, non-lazy equivalent — a trimmed ring of the last window + max_lookback 15-minute bars, folded the same way and re-run through that library's ATR (never an O(n) recompute over all history). Times are per bar, amortized over the 20 000-bar stream (so volas's periodic compaction is included).

atr:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	633.9 ns	631.8 ns	1,577,608	3	67.18×
talib	6.15 µs	6.15 µs	162,620	3	6.90×
polars	33.92 µs	33.92 µs	29,482	3	1.25×
pandas	43.22 µs	42.45 µs	23,135	3	1.00×

Append one new bar → updated indicator

A new bar arrives. volas / stock_pandas refresh their cached column incrementally (O(lookback)); the libraries with no indicator cache (pandas / pandas_ta / polars / talib) must recompute the series (O(n)). Every candidate is measured with the same round count so the rounds column is comparable.

atr:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	588.6 ns	583.0 ns	1,698,987	500	896.80×
talib	15.27 µs	14.88 µs	65,501	500	35.15×
polars	249.08 µs	100.90 µs	4,015	500	5.18×
pandas_ta	126.87 µs	119.56 µs	7,882	500	4.37×
stock_pandas	541.17 µs	448.33 µs	1,848	500	1.17×
pandas	672.67 µs	522.83 µs	1,487	500	1.00×

boll.upper

Candidate	Mean	Median	OPS	rounds	Perf
volas	1.55 µs	1.54 µs	643,434	500	299.37×
talib	12.23 µs	12.00 µs	81,782	500	38.44×
polars	96.81 µs	91.38 µs	10,330	500	5.05×
pandas	165.90 µs	158.35 µs	6,028	500	2.91×
stock_pandas	402.27 µs	386.90 µs	2,486	500	1.19×
pandas_ta	500.87 µs	461.33 µs	1,997	500	1.00×

ema:12

Candidate	Mean	Median	OPS	rounds	Perf
volas	974.3 ns	583.0 ns	1,026,424	500	732.67×
talib	8.84 µs	7.17 µs	113,094	500	59.60×
polars	46.78 µs	40.17 µs	21,379	500	10.63×
pandas	46.73 µs	42.69 µs	21,397	500	10.01×
pandas_ta	60.30 µs	56.29 µs	16,584	500	7.59×
stock_pandas	523.04 µs	427.15 µs	1,912	500	1.00×

ma:20

Candidate	Mean	Median	OPS	rounds	Perf
volas	1.38 µs	1.33 µs	726,134	500	274.00×
talib	6.38 µs	6.29 µs	156,650	500	58.09×
polars	37.68 µs	36.92 µs	26,540	500	9.90×
pandas	58.84 µs	51.79 µs	16,995	500	7.06×
pandas_ta	60.63 µs	55.50 µs	16,495	500	6.59×
stock_pandas	377.98 µs	365.52 µs	2,646	500	1.00×

macd

Candidate	Mean	Median	OPS	rounds	Perf
volas	724.5 ns	708.0 ns	1,380,266	500	538.02×
talib	18.78 µs	18.33 µs	53,262	500	20.78×
polars	70.49 µs	77.08 µs	14,187	500	4.94×
pandas	145.99 µs	104.88 µs	6,850	500	3.63×
pandas_ta	349.16 µs	297.96 µs	2,864	500	1.28×
stock_pandas	399.12 µs	380.92 µs	2,505	500	1.00×

macd.signal

Candidate	Mean	Median	OPS	rounds	Perf
volas	720.9 ns	667.0 ns	1,387,105	500	570.06×
talib	18.53 µs	18.21 µs	53,978	500	20.88×
polars	107.00 µs	107.12 µs	9,346	500	3.55×
pandas	132.17 µs	128.77 µs	7,566	500	2.95×
pandas_ta	265.42 µs	248.44 µs	3,768	500	1.53×
stock_pandas	395.30 µs	380.23 µs	2,530	500	1.00×

rsi:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	650.9 ns	542.0 ns	1,536,282	500	859.51×
talib	14.70 µs	14.08 µs	68,048	500	33.08×
pandas_ta	71.26 µs	66.08 µs	14,034	500	7.05×
polars	270.92 µs	190.73 µs	3,691	500	2.44×
stock_pandas	419.30 µs	395.13 µs	2,385	500	1.18×
pandas	486.77 µs	465.85 µs	2,054	500	1.00×

willr:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	1.13 µs	1.08 µs	888,342	500	211.26×
talib	12.73 µs	12.33 µs	78,525	500	18.56×
pandas_ta	92.98 µs	88.38 µs	10,755	500	2.59×
polars	99.82 µs	95.56 µs	10,018	500	2.40×
pandas	244.89 µs	228.90 µs	4,083	500	1.00×

Core DataFrame API (construct / slice / mask / assign / copy)

The data-handling plumbing a live system runs around every indicator call — frame construction, column access, row slicing, boolean masking, column assignment, copy — timed against pandas / polars. Not indicator math; the surrounding core APIs.

construct

Candidate	Mean	Median	OPS	rounds	Perf
volas	5.16 µs	4.42 µs	193,965	30038	12.57×
polars	29.04 µs	23.33 µs	34,430	832	2.38×
pandas	65.13 µs	55.50 µs	15,353	640	1.00×

copy

Candidate	Mean	Median	OPS	rounds	Perf
volas	302.1 ns	292.0 ns	3,310,473	42856	102.74×
polars	408.7 ns	375.0 ns	2,446,593	2743	80.00×
pandas	31.73 µs	30.00 µs	31,517	4029	1.00×

getcol

Candidate	Mean	Median	OPS	rounds	Perf
volas	401.9 ns	333.0 ns	2,488,015	27778	21.52×
polars	728.1 ns	542.0 ns	1,373,441	8612	13.22×
pandas	8.34 µs	7.17 µs	119,908	933	1.00×

mask

Candidate	Mean	Median	OPS	rounds	Perf
volas	11.17 µs	10.29 µs	89,525	17455	6.68×
polars	60.96 µs	52.35 µs	16,404	82	1.31×
pandas	70.95 µs	68.71 µs	14,094	532	1.00×

setitem

Candidate	Mean	Median	OPS	rounds	Perf
volas	1.22 µs	1.21 µs	822,232	2769	12.11×
pandas	10.39 µs	8.04 µs	96,287	1772	1.82×
polars	16.93 µs	14.62 µs	59,079	404	1.00×

slice

Candidate	Mean	Median	OPS	rounds	Perf
polars	3.91 µs	2.21 µs	255,779	414	2.32×
volas	2.86 µs	2.46 µs	350,013	1142	2.08×
pandas	6.22 µs	5.13 µs	160,771	8357	1.00×

Batch indicator computation

Compute the indicator over the whole series, across every library.

atr:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	5.87 µs	4.33 µs	170,465	14617	112.78×
talib	17.49 µs	14.58 µs	57,184	29740	33.51×
polars	126.99 µs	91.17 µs	7,875	2588	5.36×
stock_pandas	102.24 µs	97.46 µs	9,781	4826	5.02×
pandas_ta	114.96 µs	111.58 µs	8,699	3206	4.38×
pandas	505.52 µs	488.77 µs	1,978	908	1.00×

boll.upper

Candidate	Mean	Median	OPS	rounds	Perf
volas	6.17 µs	6.08 µs	161,968	35191	71.85×
talib	12.09 µs	11.92 µs	82,705	33013	36.68×
polars	92.84 µs	90.75 µs	10,771	3236	4.82×
pandas	156.01 µs	152.54 µs	6,410	1931	2.87×
stock_pandas	164.92 µs	162.38 µs	6,064	2925	2.69×
pandas_ta	454.00 µs	437.08 µs	2,203	919	1.00×

ema:12

Candidate	Mean	Median	OPS	rounds	Perf
volas	5.27 µs	5.21 µs	189,863	56603	10.29×
talib	7.07 µs	6.96 µs	141,423	42180	7.70×
stock_pandas	42.61 µs	38.25 µs	23,467	8572	1.40×
polars	42.66 µs	39.21 µs	23,442	5217	1.37×
pandas	43.50 µs	41.54 µs	22,989	3433	1.29×
pandas_ta	55.87 µs	53.58 µs	17,899	4472	1.00×

hhv:10

Candidate	Mean	Median	OPS	rounds	Perf
volas	5.05 µs	3.21 µs	198,074	20762	22.33×
talib	8.06 µs	5.75 µs	124,057	19418	12.46×
polars	40.96 µs	31.08 µs	24,417	3641	2.30×
stock_pandas	72.14 µs	51.75 µs	13,862	5953	1.38×
pandas	110.00 µs	71.62 µs	9,091	2123	1.00×

llv:10

Candidate	Mean	Median	OPS	rounds	Perf
volas	5.06 µs	3.87 µs	197,738	24845	21.70×
talib	8.54 µs	6.04 µs	117,075	16878	13.92×
polars	49.54 µs	37.94 µs	20,187	1126	2.22×
stock_pandas	89.75 µs	62.38 µs	11,143	1471	1.35×
pandas	130.39 µs	84.08 µs	7,669	1082	1.00×

ma:20

Candidate	Mean	Median	OPS	rounds	Perf
volas	4.51 µs	4.42 µs	221,843	42106	15.23×
talib	6.94 µs	6.21 µs	144,192	42404	10.83×
polars	35.16 µs	33.25 µs	28,443	4951	2.02×
pandas	48.54 µs	45.33 µs	20,602	3173	1.48×
pandas_ta	57.01 µs	53.08 µs	17,541	4394	1.27×
stock_pandas	71.80 µs	67.25 µs	13,927	6757	1.00×

macd

Candidate	Mean	Median	OPS	rounds	Perf
volas	7.06 µs	6.96 µs	141,594	40817	39.71×
talib	18.19 µs	17.92 µs	54,972	22902	15.43×
stock_pandas	51.93 µs	50.67 µs	19,258	8103	5.45×
polars	433.45 µs	76.21 µs	2,307	4210	3.63×
pandas	95.17 µs	92.04 µs	10,507	2737	3.00×
pandas_ta	290.46 µs	276.37 µs	3,443	1459	1.00×

macd.signal

Candidate	Mean	Median	OPS	rounds	Perf
volas	12.24 µs	11.75 µs	81,709	40202	1705.18×
talib	18.20 µs	17.96 µs	54,950	12800	1115.64×
stock_pandas	63.71 µs	62.00 µs	15,696	5993	323.16×
pandas	128.92 µs	125.15 µs	7,757	2464	160.10×
pandas_ta	285.55 µs	274.25 µs	3,502	1395	73.06×
polars	15.37 ms	20.04 ms	65	96	1.00×

rsi:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	12.27 µs	12.08 µs	81,515	39669	38.20×
talib	14.36 µs	14.13 µs	69,621	34683	32.68×
stock_pandas	61.05 µs	58.21 µs	16,380	5486	7.93×
pandas_ta	67.96 µs	63.67 µs	14,715	3096	7.25×
polars	170.40 µs	164.08 µs	5,868	1709	2.81×
pandas	509.60 µs	461.62 µs	1,962	929	1.00×

willr:14

Candidate	Mean	Median	OPS	rounds	Perf
volas	10.01 µs	9.87 µs	99,865	22621	22.24×
talib	11.93 µs	11.75 µs	83,854	26374	18.69×
pandas_ta	89.54 µs	85.83 µs	11,169	3164	2.56×
polars	702.50 µs	96.96 µs	1,423	2972	2.27×
pandas	228.31 µs	219.62 µs	4,380	1523	1.00×

Full coverage — volas vs TA-Lib

Every indicator both volas and TA-Lib implement (the set the parity suite aligns), one row per indicator. The default volas vs TA-Lib column is the Tencent fixture; optional generated lengths and cached append refresh appear as additional ratio columns. Values > 1.00× mean volas is faster.

volas beats TA-Lib on 150 / 158 covered indicators by the default ratio (0 exactly even, 8 slower).

Indicator	volas	TA-Lib	volas vs TA-Lib (after append)	volas vs TA-Lib	volas vs TA-Lib (100)	volas vs TA-Lib (250)	volas vs TA-Lib (20000)
midpoint:14	7.67 µs	71.96 µs	65.30×	9.39×	n/a	n/a	n/a
cdl.stalledpattern	3.08 µs	22.12 µs	21.41×	7.18×	n/a	n/a	n/a
cdl.kickingbylength	4.92 µs	28.75 µs	26.87×	5.85×	n/a	n/a	n/a
cdl.3starsinsouth	3.08 µs	16.62 µs	16.79×	5.39×	n/a	n/a	n/a
cdl.upsidegap2crows	3.33 µs	15.67 µs	15.41×	4.70×	n/a	n/a	n/a
cdl.3whitesoldiers	6.29 µs	28.17 µs	26.20×	4.48×	n/a	n/a	n/a
minmax.max:30	3.83 µs	16.83 µs	10.97×	4.39×	n/a	n/a	n/a
linearreg_intercept:14	3.15 µs	13.50 µs	13.01×	4.29×	n/a	n/a	n/a
cdl.morningstar	3.29 µs	13.92 µs	13.61×	4.23×	n/a	n/a	n/a
minmax.min:30	3.87 µs	16.37 µs	11.40×	4.23×	n/a	n/a	n/a
cdl.marubozu	3.67 µs	15.25 µs	15.25×	4.16×	n/a	n/a	n/a
linearreg:14	3.46 µs	14.00 µs	13.48×	4.05×	n/a	n/a	n/a
tsf:14	3.46 µs	14.00 µs	13.00×	4.05×	n/a	n/a	n/a
cdl.thrusting	4.00 µs	16.13 µs	15.79×	4.03×	n/a	n/a	n/a
cdl.kicking	7.37 µs	29.54 µs	28.83×	4.01×	n/a	n/a	n/a
cdl.haramicross	3.92 µs	15.46 µs	14.77×	3.95×	n/a	n/a	n/a
cdl.matchinglow	3.04 µs	11.96 µs	12.13×	3.93×	n/a	n/a	n/a
minus_dm:14	3.12 µs	11.96 µs	19.20×	3.83×	n/a	n/a	n/a
linearreg_slope:14	3.13 µs	11.87 µs	11.48×	3.80×	n/a	n/a	n/a
plus_dm:14	3.17 µs	11.96 µs	19.27×	3.78×	n/a	n/a	n/a
cdl.counterattack	4.79 µs	17.96 µs	18.11×	3.75×	n/a	n/a	n/a
rocr:10	704.2 ns	2.58 µs	5.82×	3.67×	n/a	n/a	n/a
cdl.eveningstar	3.33 µs	12.17 µs	11.58×	3.65×	n/a	n/a	n/a
cdl.abandonedbaby	6.08 µs	21.17 µs	19.66×	3.48×	n/a	n/a	n/a
accbands:20	4.50 µs	14.96 µs	10.06×	3.32×	n/a	n/a	n/a
cdl.separatinglines	6.79 µs	21.96 µs	19.85×	3.23×	n/a	n/a	n/a
atr:14	4.17 µs	13.46 µs	27.08×	3.23×	n/a	n/a	n/a
cdl.ladderbottom	2.71 µs	8.67 µs	8.31×	3.20×	n/a	n/a	n/a
cdl.eveningdojistar	5.46 µs	17.46 µs	16.20×	3.20×	n/a	n/a	n/a
cdl.risefall3methods	7.71 µs	24.37 µs	22.63×	3.16×	n/a	n/a	n/a
minmaxindex.max:30	6.71 µs	20.12 µs	32.87×	3.00×	n/a	n/a	n/a
natr:14	4.54 µs	13.46 µs	25.03×	2.96×	n/a	n/a	n/a
tema:30	5.25 µs	15.37 µs	26.51×	2.93×	n/a	n/a	n/a
mom:10	550.0 ns	1.54 µs	3.81×	2.80×	n/a	n/a	n/a
cdl.identical3crows	6.79 µs	18.88 µs	18.35×	2.78×	n/a	n/a	n/a
cdl.sticksandwich	3.79 µs	10.42 µs	10.69×	2.75×	n/a	n/a	n/a
cdl.morningdojistar	5.71 µs	15.63 µs	14.54×	2.74×	n/a	n/a	n/a
accbands.upper:20	5.50 µs	15.00 µs	10.37×	2.73×	n/a	n/a	n/a
cdl.inneck	3.87 µs	10.42 µs	12.17×	2.69×	n/a	n/a	n/a
linearreg_angle:14	12.75 µs	34.08 µs	29.34×	2.67×	n/a	n/a	n/a
cdl.onneck	4.13 µs	10.96 µs	11.73×	2.66×	n/a	n/a	n/a
accbands.lower:20	5.54 µs	14.50 µs	10.67×	2.62×	n/a	n/a	n/a
cdl.advanceblock	11.00 µs	28.71 µs	25.74×	2.61×	n/a	n/a	n/a
medprice	600.0 ns	1.54 µs	3.46×	2.57×	n/a	n/a	n/a
trix:30	6.50 µs	16.17 µs	26.00×	2.49×	n/a	n/a	n/a
macd	6.83 µs	16.63 µs	25.21×	2.43×	n/a	n/a	n/a
ht_dcphase	194.75 µs	473.37 µs	436.79×	2.43×	n/a	n/a	n/a
boll.middle	4.38 µs	10.62 µs	6.97×	2.43×	n/a	n/a	n/a
ht_sine.leadsine	213.42 µs	516.25 µs	474.68×	2.42×	n/a	n/a	n/a
ht_sine.sine	212.54 µs	513.83 µs	457.44×	2.42×	n/a	n/a	n/a
macdfix	6.75 µs	16.29 µs	24.74×	2.41×	n/a	n/a	n/a
wclprice	708.3 ns	1.71 µs	4.00×	2.41×	n/a	n/a	n/a
cdl.longline	5.21 µs	12.50 µs	12.70×	2.40×	n/a	n/a	n/a
midprice:14	7.75 µs	18.58 µs	14.90×	2.40×	n/a	n/a	n/a
wma:30	2.46 µs	5.87 µs	4.30×	2.39×	n/a	n/a	n/a
avgprice	850.0 ns	2.00 µs	4.90×	2.35×	n/a	n/a	n/a
cdl.highwave	2.54 µs	5.87 µs	5.44×	2.31×	n/a	n/a	n/a
cdl.invertedhammer	5.75 µs	13.25 µs	18.05×	2.30×	n/a	n/a	n/a
ht_trendmode	241.71 µs	556.33 µs	494.18×	2.30×	n/a	n/a	n/a
cdl.dragonflydoji	4.92 µs	11.25 µs	10.58×	2.29×	n/a	n/a	n/a
typprice	766.7 ns	1.75 µs	4.18×	2.28×	n/a	n/a	n/a
cci:14	10.33 µs	23.46 µs	20.15×	2.27×	n/a	n/a	n/a
cdl.dojistar	5.25 µs	11.71 µs	11.27×	2.23×	n/a	n/a	n/a
minmaxindex.min:30	9.17 µs	20.17 µs	35.02×	2.20×	n/a	n/a	n/a
cdl.homingpigeon	3.17 µs	6.96 µs	7.05×	2.20×	n/a	n/a	n/a
cdl.concealbabyswall	4.92 µs	10.54 µs	10.12×	2.14×	n/a	n/a	n/a
dema:30	5.17 µs	10.83 µs	17.47×	2.10×	n/a	n/a	n/a
ultosc	8.25 µs	17.29 µs	9.95×	2.10×	n/a	n/a	n/a
cdl.gravestonedoji	5.67 µs	11.87 µs	11.52×	2.10×	n/a	n/a	n/a
cdl.tasukigap	4.54 µs	9.33 µs	9.05×	2.06×	n/a	n/a	n/a
cdl.mathold	9.79 µs	19.79 µs	17.80×	2.02×	n/a	n/a	n/a
cdl.unique3river	6.50 µs	13.13 µs	12.72×	2.02×	n/a	n/a	n/a
tr	979.0 ns	1.96 µs	4.46×	2.00×	n/a	n/a	n/a
dx:14	8.58 µs	16.75 µs	26.87×	1.95×	n/a	n/a	n/a
cdl.shortline	5.62 µs	10.96 µs	10.33×	1.95×	n/a	n/a	n/a
cdl.gapsidesidewhite	5.25 µs	10.04 µs	9.81×	1.91×	n/a	n/a	n/a
cdl.3inside	7.29 µs	13.71 µs	14.25×	1.88×	n/a	n/a	n/a
cdl.harami	3.46 µs	6.50 µs	6.32×	1.88×	n/a	n/a	n/a
cdl.spinningtop	2.75 µs	5.17 µs	4.89×	1.88×	n/a	n/a	n/a
cdl.3blackcrows	5.58 µs	10.33 µs	9.41×	1.85×	n/a	n/a	n/a
cdl.closingmarubozu	4.17 µs	7.71 µs	7.66×	1.85×	n/a	n/a	n/a
beta:5@high,low	6.92 µs	12.67 µs	7.11×	1.83×	n/a	n/a	n/a
boll.upper	5.87 µs	10.62 µs	6.95×	1.81×	n/a	n/a	n/a
cdl.rickshawman	4.17 µs	7.46 µs	7.35×	1.79×	n/a	n/a	n/a
maxindex:30	6.33 µs	11.33 µs	18.87×	1.79×	n/a	n/a	n/a
cdl.2crows	4.54 µs	8.12 µs	8.39×	1.79×	n/a	n/a	n/a
cdl.takuri	4.71 µs	8.17 µs	8.04×	1.73×	n/a	n/a	n/a
cdl.xsidegap3methods	4.21 µs	7.29 µs	8.00×	1.73×	n/a	n/a	n/a
cdl.belthold	4.12 µs	7.08 µs	6.73×	1.72×	n/a	n/a	n/a
boll.lower	5.87 µs	10.00 µs	6.92×	1.70×	n/a	n/a	n/a
minus_di:14	7.67 µs	13.04 µs	21.07×	1.70×	n/a	n/a	n/a
plus_di:14	7.71 µs	12.87 µs	21.20×	1.67×	n/a	n/a	n/a
aroon.up:14	6.00 µs	9.92 µs	9.08×	1.65×	3.69×	2.71×	1.67×
trima:30	4.38 µs	7.21 µs	5.15×	1.65×	n/a	n/a	n/a
cdl.doji	2.60 µs	4.25 µs	3.96×	1.63×	n/a	n/a	n/a
roc:10	1.60 µs	2.58 µs	6.31×	1.62×	3.70×	3.33×	1.07×
cdl.longleggeddoji	3.04 µs	4.92 µs	5.00×	1.62×	n/a	n/a	n/a
rocr100:10	1.61 µs	2.58 µs	6.29×	1.60×	n/a	n/a	n/a
rocp:10	1.61 µs	2.54 µs	5.73×	1.58×	n/a	n/a	n/a
ad	2.67 µs	4.08 µs	6.25×	1.53×	n/a	n/a	n/a
cdl.tristar	5.21 µs	7.92 µs	7.54×	1.52×	3.04×	2.35×	2.53×
bop	2.06 µs	3.04 µs	7.51×	1.47×	n/a	n/a	n/a
hhv:10	3.00 µs	4.42 µs	4.95×	1.47×	3.13×	2.64×	1.52×
macd.signal	11.54 µs	16.63 µs	26.80×	1.44×	n/a	n/a	n/a
macdext	10.00 µs	14.33 µs	7.91×	1.43×	n/a	n/a	n/a
cdl.shootingstar	5.79 µs	8.29 µs	9.04×	1.43×	n/a	n/a	n/a
macdfix.histogram	11.54 µs	16.42 µs	28.30×	1.42×	n/a	n/a	n/a
cdl.piercing	3.33 µs	4.71 µs	4.80×	1.41×	n/a	n/a	n/a
cdl.3outside	2.96 µs	4.17 µs	4.08×	1.41×	n/a	n/a	n/a
cdl.engulfing	3.17 µs	4.46 µs	4.19×	1.41×	n/a	n/a	n/a
adx:14	11.42 µs	16.04 µs	24.15×	1.41×	n/a	n/a	n/a
macd.histogram	11.92 µs	16.71 µs	26.93×	1.40×	n/a	n/a	n/a
adxr:14	11.87 µs	16.46 µs	23.36×	1.39×	n/a	n/a	n/a
minindex:30	9.58 µs	13.12 µs	21.33×	1.37×	n/a	n/a	n/a
sarext	5.12 µs	7.00 µs	9.61×	1.37×	n/a	n/a	n/a
stoch.d	11.58 µs	15.79 µs	14.62×	1.36×	2.76×	2.14×	1.20×
macdfix.signal	11.58 µs	15.79 µs	26.40×	1.36×	n/a	n/a	n/a
obv	2.35 µs	3.21 µs	4.87×	1.36×	n/a	n/a	n/a
sar	4.96 µs	6.75 µs	9.59×	1.36×	n/a	n/a	n/a
adosc	5.33 µs	6.96 µs	9.95×	1.30×	n/a	n/a	n/a
cdl.hikkakemod	8.54 µs	11.04 µs	10.27×	1.29×	n/a	n/a	n/a
stochf.k	9.21 µs	11.87 µs	11.96×	1.29×	n/a	n/a	n/a
cdl.darkcloudcover	3.33 µs	4.29 µs	4.28×	1.29×	n/a	n/a	n/a
llv:10	3.62 µs	4.58 µs	6.00×	1.26×	n/a	n/a	n/a
sum:30	4.12 µs	5.17 µs	3.60×	1.25×	n/a	n/a	n/a
ma:20	4.25 µs	5.29 µs	3.88×	1.25×	n/a	n/a	n/a
t3:5	5.54 µs	6.71 µs	11.79×	1.21×	n/a	n/a	n/a
cdl.3linestrike	5.17 µs	6.25 µs	6.04×	1.21×	2.97×	2.21×	1.20×
cdl.hikkake	3.67 µs	4.42 µs	4.40×	1.20×	3.33×	2.55×	0.95×
stoch.k	13.29 µs	15.83 µs	14.18×	1.19×	n/a	n/a	n/a
var:5	4.37 µs	5.21 µs	5.52×	1.19×	n/a	n/a	n/a
stddev:5	5.00 µs	5.87 µs	6.00×	1.18×	n/a	n/a	n/a
aroon.down:14	8.37 µs	9.83 µs	9.04×	1.17×	3.42×	2.62×	1.55×
ema:12	5.04 µs	5.92 µs	13.01×	1.17×	n/a	n/a	n/a
kama:30	5.08 µs	5.92 µs	9.60×	1.16×	n/a	n/a	n/a
mavp:2,30@close,periods	119.17 µs	138.21 µs	24.97×	1.16×	n/a	n/a	n/a
cdl.breakaway	4.25 µs	4.92 µs	4.52×	1.16×	3.13×	2.40×	2.00×
aroonosc:14	8.21 µs	9.42 µs	8.22×	1.15×	2.91×	2.26×	0.92×
stochf.d	10.33 µs	11.79 µs	11.48×	1.14×	2.85×	2.08×	1.18×
correl:30@high,low	6.54 µs	7.46 µs	3.33×	1.14×	n/a	n/a	n/a
stochrsi.k	20.21 µs	22.96 µs	30.89×	1.14×	n/a	n/a	n/a
ppo	10.17 µs	11.42 µs	6.22×	1.12×	n/a	n/a	n/a
stochrsi.d	20.54 µs	22.92 µs	29.24×	1.12×	1.95×	1.47×	1.11×
cmo:14	11.71 µs	12.88 µs	26.17×	1.10×	n/a	n/a	n/a
ht_trendline	108.71 µs	119.29 µs	136.83×	1.10×	n/a	n/a	n/a
willr:14	9.50 µs	10.42 µs	9.89×	1.10×	n/a	n/a	n/a
rsi:14	11.75 µs	12.79 µs	26.00×	1.09×	n/a	n/a	n/a
cdl.hammer	9.04 µs	9.50 µs	10.03×	1.05×	n/a	n/a	n/a
macdext.signal	13.96 µs	14.33 µs	7.73×	1.03×	n/a	n/a	n/a
mfi:14	4.83 µs	4.96 µs	4.52×	1.03×	3.12×	2.32×	4.10×
mama.fama	108.42 µs	106.83 µs	123.29×	0.99×	n/a	n/a	n/a
apo	10.00 µs	9.83 µs	5.39×	0.98×	n/a	n/a	n/a
mama	104.79 µs	100.79 µs	108.17×	0.96×	1.12×	1.05×	1.15×
ht_phasor.inphase	110.29 µs	104.37 µs	131.79×	0.95×	n/a	n/a	n/a
ht_phasor.quadrature	110.38 µs	104.33 µs	132.06×	0.95×	n/a	n/a	n/a
macdext.histogram	15.54 µs	14.42 µs	7.65×	0.93×	n/a	n/a	n/a
ht_dcperiod	117.17 µs	106.08 µs	127.45×	0.91×	1.02×	0.99×	0.90×
cdl.hangingman	13.37 µs	9.58 µs	10.72×	0.72×	n/a	n/a	n/a

Generated by scripts/benchmark_report.py from pytest-benchmark output.