This appeared to build OK, but then it conked out again:
I'd continue, but I'll need to look into this later as I have somewhere to be in a bit.
Code:
nvcc -gencode=arch=compute_52,code=\"sm_52,compute_52\" -gencode=arch=compute_50,code=\"sm_50,compute_50\" -I/usr/local/cuda/include -I. --ptxas-options="-v" -gencode=arch=compute_20,code=\"sm_21,compute_20\" --maxrregcount=80 -o scrypt/salsa_kernel.o -c scrypt/salsa_kernel.cu
nvcc fatal : Unsupported gpu architecture 'compute_20'
nvcc fatal : Unsupported gpu architecture 'compute_20'
I'd continue, but I'll need to look into this later as I have somewhere to be in a bit.
Took another whack at it just now...fixed Makefile.am to build this with compute_30 instead of compute_20. It now builds (the fixes I applied are in https://gitlab.com/salfter/portage/tree/master/net-misc/sp-ccminer/files), but a quick run mining with the quark algo conked out after a few minutes with a floating-point exception. I'm not using excessive optimization...just -O2 -march=native, and the CPU for which it's building is a Celeron G3920.