[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a0bced37852884220ee4a31baf09cffc@smtp.hushmail.com>
Date: Tue, 15 Jan 2013 08:20:36 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: fbe9cf13b6 test results
On 15 Jan, 2013, at 0:43 , Lukas Odzioba <lukas.odzioba@...il.com> wrote:
> I ran some tests on my machine. I am using:
>
> Mint 14
> Catalyst 12.11- beta so we may not want to fix ocl problems, but it is
> worth to mention them
> OpenSSL 1.0.1c 10 May 2012
> build linux-x86-64-opencl
>
> Benchmarking: Sybase ASE salted SHA-256 [32/64 OpenSSL]... FAILED
> (get_hash[0](0)) ??? (failed a few times, when I run it by -fo
> parameter it seems to work)
I recently changed this format to dynamically allocate buffers. Using calloc should not be needed (it's cleared in set_key) but I now changed it to do so. Please try that.
> 5850 -test
> OpenCL error (CL_OUT_OF_RESOURCES) in file (opencl_mysqlsha1_fmt.c) at
> line (303) - (failed in clEnqueueNDRangeKernel)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_cryptmd5_fmt.c) at
> line (166) - (Error Unmapping inbuffer)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_rawmd5_fmt.c) at
> line (90) - (Error Unmapping partial_hashes)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_rawmd4_fmt.c) at
> line (93) - (Error Unmapping partial_hashes)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_nsldaps_fmt.c) at
> line (149) - (Error Unmapping outbuffer)
>
> 6950 -test
> OpenCL error (CL_OUT_OF_RESOURCES) in file (opencl_wpapsk_fmt.c) at
> line (167) - (Run initial kernel)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_cryptmd5_fmt.c) at
> line (166) - (Error Unmapping inbuffer)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_rawmd5_fmt.c) at
> line (90) - (Error Unmapping partial_hashes)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_rawmd4_fmt.c) at
> line (93) - (Error Unmapping partial_hashes)
> OpenCL error (CL_INVALID_CONTEXT) in file (opencl_nsldaps_fmt.c) at
> line (149) - (Error Unmapping outbuffer)
Was this with --test=0 (which now lowers GWS) or manually using a low GWS environment variable? If not, it is the known problem of not calling done() between formats.
Also, if those messages come at exit after all tests finished, it's because atexit(done) works poorly. This is known and will be fixed (we'll remove the atexit(done) calls before releasing).
magnum
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
