Release 1.2 of double-double arithmetic posted to registered developer website

Release notes:

  • Release 1.2
  • (1) Deployed new implementation of div_dbldbl() and sqrt_dbldbl() based on
  • Newton-Raphson iteration, providing significant speedup.
  • (2) Added new function rsqrt_dbldbl() which provides reciprocal square root.

Download instructions:

(1) Log in at
(2) Click green link “CUDA/GPU Computing Registered Developer Program”
(3) Click green link “Download” after “CUDA Double-Double Precision Arithmetic (Updated April 2013)”
(4) Clock on green link “dbldbl_v1_2.tar”
(5) Confirm legal terms by clicking button “Agree & Download”
(6) Download should start