[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <CABob6irDKnOu7W9jVwvVi52di07+VrYfRaumWJ0OQy4c8T+bCQ@mail.gmail.com>
Date: Sat, 27 Jul 2013 10:13:43 +0200
From: Lukas Odzioba <lukas.odzioba@...il.com>
To: john-dev@...ts.openwall.com
Subject: phpass-opencl experiments
Just to let you know.
I played for a while with phpass-opencl code here are my remarks.
nv - Titan
amd - core 0 of 7990@...MHz
Current KPC is equal 2048*96:
amd: 1629K
nv: 876K
After I changed KPC to 2048*1024
amd: 1638K
nv 1015K
After applying attached patch:
amd: 1762K
nv: 1115K
Speed of my old code with hash comparision on gpu side with hash
"embedded" in salt, is 2125K on amd (nv - not tested).
With proper comparision on gpu side via kernel_cmp we probably will
not get so much improvement but it should help a bit.
I'll try to implement those changes after cmiyc, now I want to play
with md5crypt, because we could gain more there.
Of course I don't mind if someone else pick it up.
Lukas
Download attachment "phpass_FFtweak.diff" of type "application/octet-stream" (1373 bytes)
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.
