09-24-2018, 03:27 AM
1. Since hashcat support passwords up to length 256, you now need four times the same amount of memory on host and on gpu as with older versions.
2. Since hashcat support utilizing all existing GPU threads available for your device, the amount of memory on host and on gpu typically is another 16 times higher than with older versions.
3. The autotune engine registers the increased memory and decreases your acceleration count (the one that you can set manually with -n) that many times to a size where you have enough memory left again.
4. If you use -n manually the autotune can not find that value in (3) so don't use it
2. Since hashcat support utilizing all existing GPU threads available for your device, the amount of memory on host and on gpu typically is another 16 times higher than with older versions.
3. The autotune engine registers the increased memory and decreases your acceleration count (the one that you can set manually with -n) that many times to a size where you have enough memory left again.
4. If you use -n manually the autotune can not find that value in (3) so don't use it