GRTGen Large NTLM Table Settings

Hello All,
I have plenty of CPU & GPU power at my disposal and would like to generate some larger NTLM tables. I have found some examples for settings but would like to get community input before I dedicate my systems for the long term.
Examples:
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 6 -i 0 --chainlength 50000 --numchains 10000000 -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 7 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 8 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 9 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
Some Rainbow Tables settings I found, not sure if they are applicable to GRTGen:
charsetall -l 7 -i 0 --chainlength 40000 --numchains 67108864
charsetall -l 8 -i 0 --chainlength 200000 --numchains 67108864
Also I found my video cards "GeForce GTX 560 Ti" have a memory limit of --numchains 27199985. This suggests that I can only change the --chainlength and table number to obtain 99% effectivness. Are there any limitations / best practices around --chainlength? Playing with the online table calculator yields various combinations that appear to get me to the 99% effectiviness, but would like community feedback on the best values to use. Thank you so much for your input.
Regards,
Damaged
I have plenty of CPU & GPU power at my disposal and would like to generate some larger NTLM tables. I have found some examples for settings but would like to get community input before I dedicate my systems for the long term.
Examples:
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 6 -i 0 --chainlength 50000 --numchains 10000000 -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 7 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 8 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
GRTGen-CUDA -h MD5 -c ./charsets/charsetall -l 9 -i 0 --chainlength ? --numchains ? -t 128 -b 128 -m 500
Some Rainbow Tables settings I found, not sure if they are applicable to GRTGen:
charsetall -l 7 -i 0 --chainlength 40000 --numchains 67108864
charsetall -l 8 -i 0 --chainlength 200000 --numchains 67108864
Also I found my video cards "GeForce GTX 560 Ti" have a memory limit of --numchains 27199985. This suggests that I can only change the --chainlength and table number to obtain 99% effectivness. Are there any limitations / best practices around --chainlength? Playing with the online table calculator yields various combinations that appear to get me to the 99% effectiviness, but would like community feedback on the best values to use. Thank you so much for your input.
Regards,
Damaged