Hi ,
[QUOTE=rroonnaalldd;284769]Do you know what causes the underlined differences in err-rates between the 32bit- and 64bit-app?
[/QUOTE]
cudatoolkit_4.0.17_linux_32_ubuntu10.10.run
gpucomputingsdk_4.0.17_linux.run
devdriver_4.1_linux_32_285.05.23.run
with GTX-550Ti
[code]
genefercuda.1.051$ ./GeneferCUDA.cuda4.0.Linux32 -b
GeneferCUDA 2.2.1 (CUDA) based on Genefer v2.2.1
Copyright (C) 2001-2003, Yves Gallot (v1.3)
Copyright (C) 2009, 2011 Mark Rodenkirch, David Underbakke (v2.2.1)
Copyright (C) 2010, 2011, Shoichiro Yamada (CUDA)
A program for finding large probable generalized Fermat primes.

Generalized Fermat Number Bench
2009574^8192+1 Time: 378 us/mul. Err: 3.82e-01 51636 digits
1632282^16384+1 Time: 384 us/mul. Err: 2.53e-01 101791 digits
1325824^32768+1 Time: 437 us/mul. Err: 2.03e-01 200622 digits
1076904^65536+1 Time: 587 us/mul. Err: 1.88e-01 395325 digits
874718^131072+1 Time: 817 us/mul. Err: 3.47e-01 778813 digits
710492^262144+1 Time: 1.33 ms/mul. Err: 4.21e-01 1533952 digits
577098^524288+1 Time: 2.58 ms/mul. Err: 2.01e-01 3020555 digits
468750^1048576+1 Time: 5.25 ms/mul. Err: 1.64e-01 5946413 digits
380742^2097152+1 Time: 10.6 ms/mul. Err: 3.63e-01 11703432 digits
309258^4194304+1 Time: 22.7 ms/mul. Err: 1.56e-01 23028076 digits
251196^8388608+1 Time: 46.5 ms/mul. Err: 1.56e-01 45298590 digits
[/code]

Hi ,
[QUOTE=msft;284826]cudatoolkit_4.0.17_linux_32_ubuntu10.10.run
gpucomputingsdk_4.0.17_linux.run
devdriver_4.1_linux_32_285.05.23.run
with GTX-550Ti
[/QUOTE]

I have installed the same Cuda41rc2-driver and the older Cuda40-sdk. Only difference is my smaller GTS450...


[add]
All your error-rates seems to be the same except: [QUOTE]309258^4194304+1 Time: 22.7 ms/mul. Err: 1.56e-01 23028076 digits
251196^8388608+1 Time: 46.5 ms/mul. Err: 1.56e-01 45298590 digits[/QUOTE]

I have: [QUOTE]309258^4194304+1 Time: 35.8 ms/mul. Err: 4.07e-01 23028076 digits
251196^8388608+1 Time: 73.2 ms/mul. Err: 4.33e-01 45298590 digits[/QUOTE]

[QUOTE=AG5BPilot;284813]
Along the way, I found that at N=4194304 Genefer takes a very long time to start up. Looking at the code, I discovered that the really simple code at the beginning that computes the actual value of b^N, by repeatedly squaring b, takes a very long time when N=44194304. About 2 hours on my Core2Quad.[/QUOTE]

[QUOTE=msft;284818]
Or something like this(GMP ?).[/QUOTE]

Use the GMP, Luke :smile:

[CODE]? N=222222^4194304;
? ##
*** last result computed in 2,749 ms.[/CODE]
That's 2.7 seconds on a 32-bit build of PARI (w/ GMP) running on Core 2 duo 2Ghz. A 64-build version will be much faster.

It can probably made even faster, by removing the 2's from the (even) base, and just doing the relevant number of squarings in the PRP loop.

[QUOTE=axn;284842]Use the GMP, Luke :smile:
[/QUOTE]I like this joke.:smile:
Do not have licence issue?

Ver 1.06
Implement GMP.

gmp.h missing...

Solved by using apt-cache search gmp...

You have to install the external GMP software yourself :smile:

Got it working and did a comparison between v1.051 and 1.060:
[QUOTE]boinc@vmware2k-3:~/Cuda/test$ time ./GeneferCUDA_cudart_cuda40_64bit -b

2009574^8192+1 Time: 696 us/mul. Err: 3.82e-01 51636 digits
1632282^16384+1 Time: 713 us/mul. Err: 2.53e-01 101791 digits
1325824^32768+1 Time: 777 us/mul. Err: 2.03e-01 200622 digits
1076904^65536+1 Time: 1.01 ms/mul. Err: 1.88e-01 395325 digits
874718^131072+1 Time: 1.37 ms/mul. Err: 3.47e-01 778813 digits
710492^262144+1 Time: 2.11 ms/mul. Err: 4.21e-01 1533952 digits
577098^524288+1 Time: 4.09 ms/mul. Err: 2.01e-01 3020555 digits
468750^1048576+1 Time: 8.21 ms/mul. Err: 1.72e-01 5946413 digits
380742^2097152+1 Time: 16.7 ms/mul. Err: 3.63e-01 11703432 digits
309258^4194304+1 Time: 36.8 ms/mul. Err: 1.56e-01 23028076 digits
251196^8388608+1 Time: 74.7 ms/mul. Err: 1.56e-01 45298590 digits

real 34m31.539s
user 14m8.597s
sys 20m12.152s[/quote]

[quote]boinc@vmware2k-3:~/Cuda/test$ time ./GeneferCUDA_gmp_cuda40_64bit -b

2009574^8192+1 Time: 696 us/mul. Err: 3.82e-01 51636 digits
1632282^16384+1 Time: 714 us/mul. Err: 2.53e-01 101791 digits
1325824^32768+1 Time: 777 us/mul. Err: 2.03e-01 200622 digits
1076904^65536+1 Time: 1.01 ms/mul. Err: 1.88e-01 395325 digits
874718^131072+1 Time: 1.37 ms/mul. Err: 3.47e-01 778813 digits
710492^262144+1 Time: 2.11 ms/mul. Err: 4.21e-01 1533952 digits
577098^524288+1 Time: 4.09 ms/mul. Err: 2.01e-01 3020555 digits
468750^1048576+1 Time: 8.21 ms/mul. Err: 1.72e-01 5946413 digits
380742^2097152+1 Time: 16.7 ms/mul. Err: 3.63e-01 11703432 digits
309258^4194304+1 Time: 36.8 ms/mul. Err: 1.56e-01 23028076 digits
251196^8388608+1 Time: 74.7 ms/mul. Err: 1.56e-01 45298590 digits

real 34m31.597s
user 14m10.249s
sys 20m10.672s
[/QUOTE]

[QUOTE=rroonnaalldd;285090]Got it working and did a comparison between v1.051 and 1.060:[/QUOTE]
"-b" option not use GMP.
Please try normal test.
By the way.
[url]http://www.primegrid.com/forum_thread.php?id=3902&nowrap=true#46251[/url]
AG5BPilot make non GMP version.
We can wait his work.:smile:

[QUOTE=msft;285194]"-b" option not use GMP.
Please try normal test.
By the way.
[url]http://www.primegrid.com/forum_thread.php?id=3902&nowrap=true#46251[/url]
AG5BPilot make non GMP version.
We can wait his work.:smile:[/QUOTE]

Hopefully it will work. :)

BTW, you can add this line into the residual check:

[code] check( 1248,4194304, "8f985a974820a6d3");[/code]

Ver 1.061
Add:
[code]
check( 1248,4194304, "8f985a974820a6d3");
[/code]