[phpBB Debug] PHP Warning: in file [ROOT]/phpbb/session.php on line 580: sizeof(): Parameter must be an array or an object that implements Countable
[phpBB Debug] PHP Warning: in file [ROOT]/phpbb/session.php on line 636: sizeof(): Parameter must be an array or an object that implements Countable
3.14.by forum • New codebase testing 0.B version
Page 1 of 4

New codebase testing 0.B version

Posted: Sun Oct 25, 2009 7:24 am
by BarsMonster
Due to massive changes, command line parameters are different now. Also, you may (finally) use config file.
Default config file have all necessary parameters, so you may just run the executable, and it will run.

Also, should show a little higher speed (due to infrastructure, not algorithm) :crazy: :crazy: :crazy:

Update:
in 0.B - some more changes.
Probably significant speed increase on tiny gpu_time.
Invalid device pointer thing should be fixed now.

http://3.14.by/files/BarsWF_CUDA_x32_0b.zip
http://3.14.by/files/BarsWF_CUDA_x64_0b.zip

Let me know your speed observations, problems, suggestions. That's codebase that would be used in distributed version.
Allowed options:
-? [ --help ] Display this help screen
-r [ --restore ] Continue interrupted work
-h [ --hash ] arg Hash to attack
--cpu_n arg (=-1) Number of CPU cores
--gpu_mask arg (=-1) 1 enables first GPU, 2 - second, 15 - first 4,
e.t.c
--grid_n arg (=-1) Grid count for nVidia
--thread_n arg (=-1) Thread count for nVidia
--gpu_time arg (=100) How long GPU core is allowed to run
--min_len arg (=1) Minimal password length
-c [ --charset ] arg
-C [ --custom_characters ] arg
-X [ --hex_characters ] arg

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 8:13 am
by D3ad0ne
I know I posted in the other post I had but do you plan to release a 32 bit version? I'm currently stuck with XP 32 because vista 64 only supports 5 cards.

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 8:34 am
by BarsMonster
D3ad0ne wrote:I know I posted in the other post I had but do you plan to release a 32 bit version? I'm currently stuck with XP 32 because vista 64 only supports 5 cards.
Doh... Currently no. May be in a day or two, sorry :sad:

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 11:15 am
by IvanG
Still problems with hybrid systems (i.e. GT200 + G92 GPUs at the same time).

Image

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 11:49 am
by LordMike
Dammit... 64 bit... :P

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 12:52 pm
by kiando
Nice to see you're working on your programm!

32-Bit would be awesome.

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 6:57 pm
by BarsMonster
You are making me sad :-)
It's almost 2010, and we are still on the 1990 tech with these 32-bit OSes :sad:

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 6:58 pm
by BarsMonster
IvanG wrote:Still problems with hybrid systems (i.e. GT200 + G92 GPUs at the same time).
Could you confirm if it works with -gpu_mask 1 and -gpu_mask 2 ?
Could you send me barswf.log when it crashes?

Re: New codebase testing 0.A version

Posted: Sun Oct 25, 2009 7:46 pm
by IvanG
BarsMonster wrote:Could you confirm if it works with -gpu_mask 1 and -gpu_mask 2 ?
Yes, it works separately.
Could you send me barswf.log when it crashes?
It doesn't crash in fact, it's locks up with heart beating. Pressing Esc leading to "Stopping threads..." message but nothing happens. Log file isn't very interesting too:

Code: Select all

2009-10-25 19:41:32.5 PHENOM9550[1960]:BarsWF:I:------------------------------
2009-10-25 19:41:32.6 PHENOM9550[1960]:BarsWF:I:BarsWF started
2009-10-25 19:41:32.6 PHENOM9550[1960]:BarsWF:I:Init CUDA
2009-10-25 19:41:32.6 PHENOM9550[1960]:BarsWF:I:Main CUDA init #0
2009-10-25 19:41:32.6 PHENOM9550[1960]:BarsWF:I:Main CUDA init #1
2009-10-25 19:41:32.6 PHENOM9550[1960]:BarsWF:I:Init SSE2
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Init Display
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Starting GPU thread #0
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Starting GPU thread #1
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Init CUDA thread #0
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Starting SSE2 thread #0
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Init CUDA thread #1
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Detecting CUDA performance...
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Starting SSE2 thread #1
2009-10-25 19:41:32.7 PHENOM9550[1960]:BarsWF:I:Detecting CUDA performance...
2009-10-25 19:41:32.8 PHENOM9550[1960]:BarsWF:I:Detecting CUDA performance...
2009-10-25 19:41:32.8 PHENOM9550[1960]:BarsWF:I:Starting SSE2 thread #2
2009-10-25 19:41:32.8 PHENOM9550[1960]:BarsWF:I:Detecting CUDA performance...
2009-10-25 19:41:32.9 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #0
2009-10-25 19:41:32.9 PHENOM9550[1960]:BarsWF:I:Starting SSE2 thread #3
2009-10-25 19:41:32.9 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #1
2009-10-25 19:41:32.9 PHENOM9550[1960]:BarsWF:I:State saved.
2009-10-25 19:41:32.9 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #0
2009-10-25 19:41:33.0 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #1
2009-10-25 19:41:33.0 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #1
2009-10-25 19:41:33.1 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #1
2009-10-25 19:41:33.1 PHENOM9550[1960]:BarsWF:I:Time spent: 78.125
2009-10-25 19:41:33.1 PHENOM9550[1960]:BarsWF:I:New amount of work per kernel: 204 : 384
2009-10-25 19:41:33.2 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #1
2009-10-25 19:41:33.6 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #1
2009-10-25 19:41:33.6 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #1
2009-10-25 19:41:34.0 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #1
2009-10-25 19:41:34.0 PHENOM9550[1960]:BarsWF:I:Time spent: 437.500
2009-10-25 19:41:34.1 PHENOM9550[1960]:BarsWF:I:New amount of work per kernel: 466 : 384
2009-10-25 19:41:34.1 PHENOM9550[1960]:BarsWF:I:Run CUDA thread #1
2009-10-25 19:41:34.1 PHENOM9550[1960]:BarsWF:I:Stop CUDA thread #1

Re: New codebase testing 0.A version

Posted: Mon Oct 26, 2009 2:55 am
by BarsMonster

Re: New codebase testing 0.A version

Posted: Mon Oct 26, 2009 5:58 am
by BarsMonster
Check first post on 0.B version.

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 9:48 am
by IvanG
Yeah, it finally works.
Image

Now there is an other subproblem -- as system config is GTX260+GTS250 it'll nice to have thread_n = 192 for first GPU and thread_n = 128 for second one ;).

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 4:36 pm
by D3ad0ne
I'm getting a procedure Entry Point Not Found - "entry point cudaSetDeviceFlags could not be located in the dynamic link library cudart.dll"

This is using the same cudart.dll that i used for the other version.

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 5:04 pm
by BarsMonster
D3ad0ne wrote:I'm getting a procedure Entry Point Not Found - "entry point cudaSetDeviceFlags could not be located in the dynamic link library cudart.dll"

This is using the same cudart.dll that i used for the other version.
Well, you have placed cudart.dll to the BarsWF directory or what? This shouldn't be needed, cudart is linked statically.
Meanwhile, not sure what changed, this function was used in 0.A version too.
Do you have pretty recent display driver?

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 5:42 pm
by BarsMonster
IvanG wrote:Yeah, it finally works.
Now there is an other subproblem -- as system config is GTX260+GTS250 it'll nice to have thread_n = 192 for first GPU and thread_n = 128 for second one ;).
Could you send me your log so that I would see what did it selected automatically? Or you specified thread_n and grid_n manuallY?

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 5:55 pm
by D3ad0ne
Well, you have placed cudart.dll to the BarsWF directory or what? This shouldn't be needed, cudart is linked statically.
Meanwhile, not sure what changed, this function was used in 0.A version too.
Do you have pretty recent display driver?
Yea cudart was in the same directory, it was there from previous versions. So I'll take that out when I get home from work. And I try and keep up to date with latest drivers.

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 10:19 pm
by DarkPrince
D3ad0ne wrote:I'm getting a procedure Entry Point Not Found - "entry point cudaSetDeviceFlags could not be located in the dynamic link library cudart.dll"

This is using the same cudart.dll that i used for the other version.

I have got the same problem

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 11:11 pm
by BarsMonster
DarkPrince wrote:
D3ad0ne wrote:I'm getting a procedure Entry Point Not Found - "entry point cudaSetDeviceFlags could not be located in the dynamic link library cudart.dll"

This is using the same cudart.dll that i used for the other version.

I have got the same problem
Did you deleted .dll?

Re: New codebase testing 0.B version

Posted: Mon Oct 26, 2009 11:59 pm
by D3ad0ne
Yes I deleted both dlls in the local directory.

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 12:16 am
by BarsMonster
D3ad0ne wrote:Yes I deleted both dlls in the local directory.
And 0.A version was working, right?
I am totally confused :-S

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 12:25 am
by BarsMonster
I've reuploaded archives, it appeared it indeed had cudart.dll dependancy.

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 1:03 am
by protter
There is a problem with a time estimation, when I use:

gpu_time=1000
charset=aA
min_len=16
hash=1b0e9fd3086d90a159a1d6cb86f11b4c

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 1:19 am
by BarsMonster
protter wrote:There is a problem with a time estimation, when I use:

gpu_time=1000
charset=aA
min_len=16
hash=1b0e9fd3086d90a159a1d6cb86f11b4c
What kind of a problem? Screenshot and details please :-)

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 2:58 am
by D3ad0ne
Works perfect! :mrgreen:

And gave me a new high score :joy:
new.JPG
new.JPG (44.91 KiB) Viewed 125808 times

Re: New codebase testing 0.B version

Posted: Tue Oct 27, 2009 4:00 am
by BarsMonster
D3ad0ne wrote:Works perfect! :mrgreen:

And gave me a new high score :joy:
Image
Gratz :-)