Decimal128 by Miguel

Go Benchmark Results

IEEE 754-2019 compliant decimal128 high-performance software solution created by Miguel.

Living document — as-measured results. Category codes, profiles, modes & method: Benchmark Key.

This is the Go view of decimal128 as-measured, band by band, with explicit ratios. It opens with the realistic financial-mix (P-fin) headline, then the per-operation band characterization (P-gen) and FMA. In Go, d128 is measured against no alternative — Go has neither an in-language decimal peer nor a libbid fallback, so its rows are d128-only (- in the alt/ratio columns). It is data only — the categories, magnitude profiles, units, and methodology are defined in the Benchmark Key (and, authoritatively, BenchmarkMatrix.md). The cross-port d128 band-shape matrices (all ports, no alternatives) live in Port-Comparison Benchmark Results; the full index of per-language pages is on the Benchmarks hub.

Summary — Ratio Range by Operation

Go has no in-language decimal peer and takes no libbid fallback, so every row on this page is d128-only — there is no alternative to compute a ratio against.

FinMix — realistic financial mix (P-fin)

The headline: one realistic 64-bit financial operation mix — a MIX add/sub stream, mul CP/WP, div CD/WD/ET/PT — ratio = alt / ours (> 1 ⇒ d128 faster). This is the profile closest to real financial code.

arm64 (M3 Pro).

port op cat profile arch mode ns alt alt ns ratio run notes
go add MIX P-fin arm64 thru 3.21 - - - Rgosw2  
go sub MIX P-fin arm64 thru 4.13 - - - Rgosw2  
go mul CP P-fin arm64 thru 1.98 - - - Rgosw2  
go mul WP P-fin arm64 thru 28.48 - - - Rgosw2  
go div CD P-fin arm64 thru 37.26 - - - Rgosw2  
go div WD P-fin arm64 thru 64.24 - - - Rgosw2  
go div ET P-fin arm64 thru 10.65 - - - Rgosw2  
go div PT P-fin arm64 thru 6.65 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ns alt alt ns ratio run notes
go add MIX P-fin x86_64 thru 8.87 - - - xRgosw2  
go sub MIX P-fin x86_64 thru 11.99 - - - xRgosw2  
go mul CP P-fin x86_64 thru 4.42 - - - xRgosw2  
go mul WP P-fin x86_64 thru 59.93 - - - xRgosw2  
go div CD P-fin x86_64 thru 94.20 - - - xRgosw2  
go div WD P-fin x86_64 thru 151.90 - - - xRgosw2  
go div ET P-fin x86_64 thru 29.66 - - - xRgosw2  
go div PT P-fin x86_64 thru 13.16 - - - xRgosw2  

Add — SQ · NQ · MQ · OQ · FQ

Swept 4096-input average per band (bare thru; ns/op = Time/4096 over the shared decimal128-resources/swept/ corpus, byte-identical operands every port). Relational peer table with explicit ratios; the cross-port d128 band-shape matrices are in Port-Comparison Benchmark Results.

arm64 (M3 Pro).

port op cat profile arch mode ours alt alt ns ratio run notes
go add SQss P-gen arm64 thru 1.54 - - - Rgosw2  
go add SQos P-gen arm64 thru 5.03 - - - Rgosw2  
go add NQss P-gen arm64 thru 6.44 - - - Rgosw2  
go add NQos P-gen arm64 thru 11.20 - - - Rgosw2  
go add MQss P-gen arm64 thru 11.86 - - - Rgosw2  
go add MQos P-gen arm64 thru 19.53 - - - Rgosw2  
go add OQss P-gen arm64 thru 28.60 - - - Rgosw2  
go add OQos P-gen arm64 thru 37.34 - - - Rgosw2  
go add FQss P-gen arm64 thru 17.58 - - - Rgosw2  
go add FQos P-gen arm64 thru 24.02 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ours alt alt ns ratio run notes
go add SQss P-gen x86_64 thru 3.47 - - - xRgosw2  
go add SQos P-gen x86_64 thru 9.82 - - - xRgosw2  
go add NQss P-gen x86_64 thru 12.51 - - - xRgosw2  
go add NQos P-gen x86_64 thru 17.40 - - - xRgosw2  
go add MQss P-gen x86_64 thru 20.81 - - - xRgosw2  
go add MQos P-gen x86_64 thru 34.95 - - - xRgosw2  
go add OQss P-gen x86_64 thru 51.43 - - - xRgosw2  
go add OQos P-gen x86_64 thru 78.79 - - - xRgosw2  
go add FQss P-gen x86_64 thru 32.77 - - - xRgosw2  
go add FQos P-gen x86_64 thru 40.59 - - - xRgosw2  

Subtract — SQ · NQ · MQ · OQ · FQ

Swept 4096-input average per band (bare thru; ns/op = Time/4096 over the shared decimal128-resources/swept/ corpus, byte-identical operands every port). Relational peer table with explicit ratios; the cross-port d128 band-shape matrices are in Port-Comparison Benchmark Results.

arm64 (M3 Pro).

port op cat profile arch mode ours alt alt ns ratio run notes
go sub SQss P-gen arm64 thru 2.92 - - - Rgosw2  
go sub SQos P-gen arm64 thru 1.62 - - - Rgosw2  
go sub NQss P-gen arm64 thru 10.68 - - - Rgosw2  
go sub NQos P-gen arm64 thru 5.59 - - - Rgosw2  
go sub MQss P-gen arm64 thru 18.11 - - - Rgosw2  
go sub MQos P-gen arm64 thru 11.53 - - - Rgosw2  
go sub OQss P-gen arm64 thru 36.44 - - - Rgosw2  
go sub OQos P-gen arm64 thru 27.58 - - - Rgosw2  
go sub FQss P-gen arm64 thru 22.72 - - - Rgosw2  
go sub FQos P-gen arm64 thru 16.72 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ours alt alt ns ratio run notes
go sub SQss P-gen x86_64 thru 8.78 - - - xRgosw2  
go sub SQos P-gen x86_64 thru 3.88 - - - xRgosw2  
go sub NQss P-gen x86_64 thru 17.03 - - - xRgosw2  
go sub NQos P-gen x86_64 thru 12.17 - - - xRgosw2  
go sub MQss P-gen x86_64 thru 34.93 - - - xRgosw2  
go sub MQos P-gen x86_64 thru 20.41 - - - xRgosw2  
go sub OQss P-gen x86_64 thru 69.12 - - - xRgosw2  
go sub OQos P-gen x86_64 thru 50.94 - - - xRgosw2  
go sub FQss P-gen x86_64 thru 45.88 - - - xRgosw2  
go sub FQos P-gen x86_64 thru 37.08 - - - xRgosw2  

Multiply — CP · WP · XP

Swept 4096-input average per band (bare thru; ns/op = Time/4096 over the shared decimal128-resources/swept/ corpus, byte-identical operands every port). Relational peer table with explicit ratios; the cross-port d128 band-shape matrices are in Port-Comparison Benchmark Results.

arm64 (M3 Pro).

port op cat profile arch mode ours alt alt ns ratio run notes
go mul CP P-gen arm64 thru 2.46 - - - Rgosw2  
go mul WP P-gen arm64 thru 28.35 - - - Rgosw2  
go mul XP P-gen arm64 thru 39.72 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ours alt alt ns ratio run notes
go mul CP P-gen x86_64 thru 6.43 - - - xRgosw2  
go mul WP P-gen x86_64 thru 52.28 - - - xRgosw2  
go mul XP P-gen x86_64 thru 78.60 - - - xRgosw2  

Divide — CD · WD · XD (+ ET · PT)

Swept 4096-input average per band (bare thru; ns/op = Time/4096 over the shared decimal128-resources/swept/ corpus, byte-identical operands every port). Relational peer table with explicit ratios; the cross-port d128 band-shape matrices are in Port-Comparison Benchmark Results.

arm64 (M3 Pro).

port op cat profile arch mode ours alt alt ns ratio run notes
go div CD P-gen arm64 thru 35.09 - - - Rgosw2  
go div WD P-gen arm64 thru 61.48 - - - Rgosw2  
go div XD P-gen arm64 thru 58.03 - - - Rgosw2  
go div ET P-gen arm64 thru 12.35 - - - Rgosw2  
go div PT P-gen arm64 thru 6.64 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ours alt alt ns ratio run notes
go div CD P-gen x86_64 thru 92.98 - - - xRgosw2  
go div WD P-gen x86_64 thru 157.10 - - - xRgosw2  
go div XD P-gen x86_64 thru 117.50 - - - xRgosw2  
go div ET P-gen x86_64 thru 35.98 - - - xRgosw2  
go div PT P-gen x86_64 thru 15.15 - - - xRgosw2  

FMA — FN (Barrett) · FF (fits-128)

Swept 4096-input average per band (bare thru; ns/op = Time/4096 over the shared decimal128-resources/swept/ corpus, byte-identical operands every port). Relational peer table with explicit ratios; the cross-port d128 band-shape matrices are in Port-Comparison Benchmark Results.

arm64 (M3 Pro).

port op cat profile arch mode ours alt alt ns ratio run notes
go fma FN FMA arm64 thru 159.20 - - - Rgosw2  
go fma FF FMA arm64 thru 72.84 - - - Rgosw2  

x86_64 (Intel i9-9880H).

port op cat profile arch mode ours alt alt ns ratio run notes
go fma FN FMA x86_64 thru 313.30 - - - xRgosw2  
go fma FF FMA x86_64 thru 165.00 - - - xRgosw2