#Amd radeon pro 5500m driver pro
If it does not, then try making edits to the first few lines of run/opencl/cryptmd5_kernel.cl enabling and disabling the individual optimizations:ĭevice 3: AMD Radeon Pro 5500M Compute Engineīenchmarking: 7z-opencl, 7-Zip (512K iterations).
#Amd radeon pro 5500m driver code
Anyway, in general we do expect people reporting issues on GitHub to be using (or at least having tried before reporting) our latest code from here, not a release.Īnyway, to try and work around whatever miscompile you're getting, you can first try using our latest code from here and see if it just works. In your case, the failure is already detected OK. The changes made to md5crypt-opencl since the 1.9.0-jumbo-1 release are relevant, but are not expected to help - rather, they increase the likelihood of detection (and refusal to work) of OpenCL kernel miscompiles. In particular, does phpass-opencl work, and on which versions (1.9.0-jumbo-1 or/and latest bleeding-jumbo)? Any other OpenCL formats you've been using? This makes me wonder whether other formats work for you. Memmem(): Like Claudio says, this is almost certainly an AMD bug, although this format in particular worked quite well on AMD GPUs lately, so I am a bit surprised. $JOHN is /usr/local/Cellar/john-jumbo/1.9.0/share/john/Įffective limit: Number of salts vs. System-wide home: /usr/local/Cellar/john-jumbo/1.9.0/share/john System-wide exec: /usr/local/Cellar/john-jumbo/1.9.0/share/john Preferred vector width: char 4, short 2, int 1, long 1īuild: darwin20.1.0 64-bit x86_64 SSE4.2 AC Native vector widths: char 4, short 2, int 1, long 1 Preferred vector width: char 1, short 1, int 1, long 1ĭevice #2 (3) name: AMD Radeon Pro 5500M Compute Engine Native vector widths: char 1, short 1, int 1, long 1 Preferred vector width: char 16, short 8, int 4, long 2ĭevice #1 (2) name: Intel(R) UHD Graphics 630 Native vector widths: char 16, short 8, int 4, long 2