I think 8th got over 10x more bfloat calculation speed by switching from MAPM to libBF.
Here is a 8th sample code to calculate 1000 PI digits:
Result:
Here is a 8th sample code to calculate 1000 PI digits:
Code:
: pi \ n -- n dup ## n:1+ n# F1 F5 n:/ n:atan F4 n:* F1 F239 n:/ n:atan n:- F4 n:* ;: app:main d:msec 1000 pi d:msec rot n:- swap >s . cr cr "took: %d msecs\n" s:strfmt . ;Code:
3.1415926535897932384626433832795028841971693993751058209749445923078164062862089986280348253421170679821480865132823066470938446095505822317253594081284811174502841027019385211055596446229489549303819644288109756659334461284756482337867831652712019091456485669234603486104543266482133936072602491412737245870066063155881748815209209628292540917153643678925903600113305305488204665213841469519415116094330572703657595919530921861173819326117931051185480744623799627495673518857527248912279381830119491298336733624406566430860213949463952247371907021798609437027705392171762931767523846748184676694051320005681271452635608277857713427577896091736371787214684409012249534301465495853710507922796892589235420199561121290219608640344181598136297747713099605187072113499999983729780499510597317328160963185950244594553469083026425223082533446850352619311881710100031378387528865875332083814206171776691473035982534904287554687311595628638823537875937519577818577805321712268066130019278766111959092164201989took: 5 msecsStatistics: Posted by jalih — Sat Apr 05, 2025 4:57 pm