You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Current Par implementation could be made more efficient (less allocations) #91
Review update against FinanceModels v7.0.0 (6cb71714), 2026-10-11.
Still applicable; low priority.par still allocates cashflow/time vectors and solves an IRR for regular coupon schedules, where the periodic par yield can be obtained directly from discounted coupon payments.
A warmed 30-year semiannual example allocated 1,152 bytes in the review; the equivalent direct regular-schedule formula allocated zero. A 10-year example allocated 448 bytes. These are local Julia 1.13.1 measurements, not an end-to-end timing benchmark.
See current implementation. Add a regular-schedule fast path if profiling warrants it. Preserve the existing distinction between par yield and par coupon on stub schedules, rather than applying the regular formula to every maturity.
Activity
alecloudenback commentedon Oct 11, 2026
Review update against FinanceModels v7.0.0 (
6cb71714), 2026-10-11.Still applicable; low priority.
parstill allocates cashflow/time vectors and solves an IRR for regular coupon schedules, where the periodic par yield can be obtained directly from discounted coupon payments.A warmed 30-year semiannual example allocated 1,152 bytes in the review; the equivalent direct regular-schedule formula allocated zero. A 10-year example allocated 448 bytes. These are local Julia 1.13.1 measurements, not an end-to-end timing benchmark.
See current implementation. Add a regular-schedule fast path if profiling warrants it. Preserve the existing distinction between par yield and par coupon on stub schedules, rather than applying the regular formula to every maturity.