Can someone tell me in what condition is pfgw faster then latest llr? Whatever project I try always llr was faster then pfgw. So I am little confused...
____________
92*10^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME
2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie!

They should be very close in speed, when comparing 32-bit to 32-bit or 64-bit to 64-bit. I've had reports of llr being much faster for some tests, but I do not know the reasons for that yet. My assumption is that pfgw's additional error checking is the reason, but I really don't know if that is true or not. I have intended to do some profiling to see if there is code in pfgw that is to causing the slow down. I haven't had time to do that.

Rogue, thanks for answer to my question. For me pfgw was always slower 10% then LLR, so I never use it, but I assume that is reason why that program exist...
____________
92*10^^1585996-1 NEAR-REPDIGIT PRIME :) :) :)
4 * 650^⁴⁹⁸¹⁰¹-1 CRUS PRIME
2022202116^¹³¹⁰⁷²+1 GENERALIZED FERMAT
Proud member of team Aggie The Pew. Go Aggie!

64-bit pfgw was normally faster than 32-bit llr for any tests that weren't base 2. pfgw can do PRP tests on almost any number and it can do primality tests on many forms that llr cannot, such as primorials and factorials.

64-bit pfgw was normally faster than 32-bit llr for any tests that weren't base 2. pfgw can do PRP tests on almost any number and it can do primality tests on many forms that llr cannot, such as primorials and factorials.