Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - Printable Version +- hashcat Forum (https://hashcat.net/forum) +-- Forum: Deprecated; Ancient Versions (https://hashcat.net/forum/forum-46.html) +--- Forum: Very old oclHashcat-plus Support (https://hashcat.net/forum/forum-23.html) +--- Thread: Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! (/thread-1137.html) |
Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - hasheponge - 05-02-2012 @ALL for users who use cudahashcat-plus, i tested 0.07 and 0.08 with same script... same files and same rule, same GPU ! differents performances ! and the worst not the same result !!!! Results : with 0.07 : Speed........: 612.6M c/s Real, 582.8M c/s GPU Recovered....: 311/1441897 Digests, 0/1 Salts Progress.....: 20132793600/80453034400 (25.02%) Rejected.....: 4420537600/20132793600 (21.96%) with 0.08 : Speed........: 472.3M c/s Real, 485.6M c/s GPU Recovered....: 7/1441897 Digests, 0/1 Salts Progress.....: 20444744960/80453034400 (25.41%) Rejected.....: 4420537600/20444744960 (21.62%) 0.08 -23% !! and curious Recovered !!! 0.07 = 311, 0.08 = 7 !!! Drivers 296.10 Hashs : MD5 , Rule : d3ad0ne.rule GPU for this test : GTX 470 , win 7-32bits ... Bug ? I tested with my second GPU (GTS 450) same results... :-( ** on this test same GPU-Loops, GPU-Accel... (05-02-2012, 06:08 PM)hasheponge Wrote: @ALL for users who use cudahashcat-plus, i tested 0.07 and 0.08 with same script... same files and same rule, same GPU ! RE: Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - atom - 05-02-2012 I tried to reproduce your case with my own made files on AMD and on NVidia, using multihash MD5 cracking, rule-based. But I am unable to reproduce! Maybe you can share all the required files to reproduce? RE: Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - hasheponge - 05-02-2012 ok Atom i pm you all informations for this test . Thanks RE: Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - atom - 05-03-2012 OK, thanks for the files. I got them. Here is the result from oclHashcat-plus v0.08: AMD hd7970: Quote:Speed........: 1029.1M c/s Real, 1203.9M c/s GPU See what happens to the speed after I've added -n 160 and --gpu-loops 1024: Quote:Speed........: 2175.4M c/s Real, 2216.0M c/s GPU Now the NVidia gtx560Ti: Quote:Speed........: 269.0M c/s Real, 271.7M c/s GPU Again, after I've added -n 160 and --gpu-loops 1024: Quote:Speed........: 380.3M c/s Real, 394.3M c/s GPU Then I tried with the NVidia gtx560Ti with -n 160 and --gpu-loops 1024 but on oclHashcat-plus v0.07: Quote:Speed........: 554.9M c/s Real, 516.2M c/s GPU So you are right, it is slower. But the important part, the number of cracked hashes is the same. I can not explain why this happens on your card, but I can not reproduce it. The speed drop itself must be because of the CUDA 4.2 and not because of a code change since I did not change anything on there. RE: Comparaison CudaHashcat-plus 0.08 vs 0.07 ... !!! - hasheponge - 05-03-2012 thanks Atom, strange strange for the number of cracked hashes... i try install a last Cuda 4.1.28 32 bits. And test... wait. it's crazy ! same results on GTX 470 or GTS 450 0.08 : Speed........: 444.6M c/s Real, 425.9M c/s GPU Recovered....: 3/1441897 Digests, 0/1 Salts Progress.....: 8216719872/80453034400 (10.21%) 0.07 (+ fix 0.071) Speed........: 661.2M c/s Real, 660.4M c/s GPU Recovered....: 159/1441897 Digests, 0/1 Salts Progress.....: 8399222240/80453034400 (10.44%) i don't know what... :-( (05-03-2012, 11:20 AM)hasheponge Wrote: thanks Atom, strange strange for the number of cracked hashes... i try install a last Cuda 4.1.28 32 bits. And test... wait. oh oh... no same rejects : 0.08 : Progress.....: 8766223840/80453034400 (10.90%) Rejected.....: 2770335200/8766223840 (31.60%) 0.07 : Progress.....: 8801912576/80453034400 (10.94%) Rejected.....: 1857324800/8801912576 (21.10%) ... |