GPGPU using Opencl

C

cayoenrique

Member
Messages
475
I do not know if you saw, here again. I am posting for you the binaries so you do not need to do full install. I guess if you only want to look.

OCLBiss_077_win_static_binary.zip (65.93 KB)
Code: 
https://workupload.com/file/cttjNqqQYXW
 
M

moonbase

VIP
Donating Member
Messages
552
cayoenrique said:
I do not know if you saw, here again. I am posting for you the binaries so you do not need to do full install. I guess if you only want to look.

OCLBiss_077_win_static_binary.zip (65.93 KB)
Code: 
https://workupload.com/file/cttjNqqQYXW
Click to expand...


Thank you, yes I saw the link that you kindly posted and I have now downloaded those files.

Do I need to read all of this topic from the beginning to find out what I do with these binary files?
Or would you please be able to give a simple explanation here?
 
C

C0der

Registered
Messages
270
Basic idea would be:
1. Unzip the files
2. edit OCLBiss.cfg with texteditor
3. run the exe, check speed, repeat at 2.
 
M

moonbase

VIP
Donating Member
Messages
552
C0der said:
Basic idea would be:
1. Unzip the files
2. edit OCLBiss.cfg with texteditor
3. run the exe, check speed, repeat at 2.
Click to expand...

@C0der

Thank you, that sounds simple enough even for a non coding dinosaur like me.
I hope to be back with some KPS results from a spare 2080Ti before moving on to the 4090.
 
C

C0der

Registered
Messages
270
On a GTX1050 with

MULTITHREADSIZE:1
LOCALWORKDROUPSIZE:32
GLOBALWORKDROUPSIZE:10240

I get 2.3E7 kps.

MULTITHREADSIZE doesnt make a difference.

For a 2080 maybe try
LOCALWORKDROUPSIZE:128
 
C

cayoenrique

Member
Messages
475
In general you start the 1rst time with sugestion from GPU. You get them printed by selecting
Code: 
LOCALWORKDROUPSIZE:0
GLOBALWORKDROUPSIZE:0
And make a note of that.
Then you use multiples of the number given for LOCALWORKDROUPSIZE.
Finally GLOBALWORKDROUPSIZE should always be a multiple of LOCALWORKDROUPSIZE.
There is logic in the program to force GLOBALWORKDROUPSIZE to be a multiple of LOCALWORKDROUPSIZE. Because if not kernel will not load.

MULTITHREADSIZE should make a difference. My best guess close to 2X. But if you already have manage to manually select a config that have you GPU saturated then the improve you see is less, as you already have saturated you GPU. Still you should see some improve due to PCIe load unload operations.
 
C

cayoenrique

Member
Messages
475
Ok I will post new DES Opencl application for your evaluation. Posting here so that we can keep all opencl apps in one place.

OCLDes_003W.zip (30.12 KB)
Code: 
https://workupload.com/file/RbtaQduxssr

Code: 
Today is Tue Oct 24 17:26:00 2023
OS TYPE:                                                Windows
OS TYPE:                                                Windows_NT
Timer test for 1 second:                                1.014002
Connected Platform:                                     Advanced Micro Devices, Inc.
Connected Device:                                       Turks
OCLDESVERSION:                                          003
PROGRAM_FILE selected:                                  des_decrypt_1block_001.cl

Device properties:
    Number of Compute Units (CU):                       6
    Recommended Local Work Group size:                  64
    Max Local Work Group size:                          256
    Global Memory size:                                 1073741824

Multithreading properties:
    Number of Threads to launch:                        4 
    Number of PES1 rounds per PES2                      16 

Single Thread properties:
    Number Loops per kernel:                            2048
    Number of keys per kernel:                          2048
    Local Work Group Size:                              256
    Global Work Group Size:                             1536
    Keys per round:                                     50331648(3000000)

File Log from Last Searched present:                    (007ACC40B0000000)
Start from config.ini value:                            (007ACC40B0000000)

Base Selected:                                          (007ACC40B0000000)
Looking for:                                            (7A67101A5BE52601)
BruteForcing for:                                       PES1:36 43 CD E6 3F 67 9D 20 00 00 00 00 00 00 00 00 
                                                        PES2:7D 1C 9F 9D 68 42 34 10 00 00 00 00 00 00 00 00 
                                                        PES3:36 43 CD E6 3F 67 9D 20 00 00 00 00 00 00 00 00 
Range:                                                  (3000000)

Launching threadnum# 0 From:(7ACC40B0000000) To:(7ACC40B3000000)
Launching threadnum# 1 From:(7ACC40B3000000) To:(7ACC40B6000000)
Launching threadnum# 2 From:(7ACC40B6000000) To:(7ACC40B9000000)
Launching threadnum# 3 From:(7ACC40B9000000) To:(7ACC40BC000000)
T# Round#       From           To           Time    kps:Round  Last4     Keys Found 
0 00000001 7ACC40B0000000 7ACC40B2FFFFFF  17:26:14  7.68e+006 2.24e+007 00 Total 00: 
1 00000002 7ACC40B3000000 7ACC40B5FFFFFF  17:26:14  7.68e+007 3.48e+007 00 Total 00: 
2 00000003 7ACC40B6000000 7ACC40B8FFFFFF  17:26:14  4.61e+008 1.43e+008 00 Total 00: 
3 00000004 7ACC40B9000000 7ACC40BBFFFFFF  17:26:15  4.61e+008 2.52e+008 00 Total 00: 
0 00000005 7ACC40BC000000 7ACC40BEFFFFFF  17:26:21  8.38e+006 2.52e+008 00 Total 00: 
1 00000006 7ACC40BF000000 7ACC40C1FFFFFF  17:26:22  4.61e+007 2.44e+008 00 Total 00: 
2 00000007 7ACC40C2000000 7ACC40C4FFFFFF  17:26:22  4.61e+008 2.44e+008 00 Total 00: 
3 00000008 7ACC40C5000000 7ACC40C7FFFFFF  17:26:22  4.61e+008 2.44e+008 00 Total 00: 
0 00000009 7ACC40C8000000 7ACC40CAFFFFFF  17:26:28  8.38e+006 2.44e+008 00 Total 00: 
1 0000000A 7ACC40CB000000 7ACC40CDFFFFFF  17:26:29  5.12e+007 2.45e+008 00 Total 00: 
0 0000000D 7ACC40D4000000 7ACC40D6FFFFFF  17:26:35  8.38e+006 2.45e+008 01 Total 01: Key 001:7A67101A5BE52601 

Finish

or you can do a vectortest and post all fake keys found
Code: 
Today is Tue Oct 24 17:40:17 2023
OS TYPE:                                                Windows
OS TYPE:                                                Windows_NT
Timer test for 1 second:                                1.014002
Connected Platform:                                     Advanced Micro Devices, Inc.
Connected Device:                                       Turks
OCLDESVERSION:                                          003
PROGRAM_FILE selected:                                  des_decrypt_1block_001.cl

Device properties:
    Number of Compute Units (CU):                       6
    Recommended Local Work Group size:                  64
    Max Local Work Group size:                          256
    Global Memory size:                                 1073741824

Multithreading properties:
    Number of Threads to launch:                        4 
    Number of PES1 rounds per PES2                      16 

Single Thread properties:
    Number Loops per kernel:                            2048
    Number of keys per kernel:                          2048
    Local Work Group Size:                              256
    Global Work Group Size:                             1536
    Keys per round:                                     50331648(3000000)

File Log from Last Searched present:                    (007ACC40B0000000)
Start from config.ini value:                            (007ACC40B0000000)
Range:                                                  (3000000)

Launching threadnum# 0 From:(7ACC40B0000000) To:(7ACC40B3000000)
Launching threadnum# 1 From:(7ACC40B3000000) To:(7ACC40B6000000)
Launching threadnum# 2 From:(7ACC40B6000000) To:(7ACC40B9000000)
Launching threadnum# 3 From:(7ACC40B9000000) To:(7ACC40BC000000)
T# Round#       From           To           Time    kps:Round  Last4     Keys Found 
0 00000001 7ACC40B0000000 7ACC40B2FFFFFF  17:40:28  7.32e+006 2.25e+007 04 Total 04: Key 001:7A6710160BA71AF1 002:7A67101616E5B037 
                                                                                         003:7A6710161A047985 004:7A671016264962A7 
1 00000002 7ACC40B3000000 7ACC40B5FFFFFF  17:40:28  1.54e+008 5.40e+007 04 Total 08: Key 001:7A67101637F1C7DF 002:7A67101638B90443 
                                                                                         003:7A6710163DBF73D9 004:7A67101643256145 
2 00000003 7ACC40B6000000 7ACC40B8FFFFFF  17:40:28  4.61e+008 1.62e+008 05 Total 13: Key 001:7A67101662CE25A8 002:7A6710166B4068D5 
                                                                                         003:7A671016735E9154 004:7A6710167A493EE6 
                                                                                         005:7A6710168C3E9D34 
3 00000004 7ACC40B9000000 7ACC40BBFFFFFF  17:40:28  4.61e+008 2.71e+008 03 Total 16: Key 001:7A671016916B6B37 002:7A671016B980FD61 
                                                                                         003:7A671016BF5B15BF 
0 00000005 7ACC40BC000000 7ACC40BEFFFFFF  17:40:34  7.95e+006 2.71e+008 02 Total 18: Key 001:7A671016C7857FC4 002:7A671016DC79915B 
1 00000006 7ACC40BF000000 7ACC40C1FFFFFF  17:40:35  6.58e+007 2.49e+008 01 Total 19: Key 001:7A6710190E49258F 
2 00000007 7ACC40C2000000 7ACC40C4FFFFFF  17:40:35  4.61e+008 2.49e+008 00 Total 19: 
3 00000008 7ACC40C5000000 7ACC40C7FFFFFF  17:40:35  4.61e+008 2.49e+008 03 Total 22: Key 001:7A6710195EA71A4C 002:7A6710196B3D1973 
                                                                                         003:7A6710197F018FA1 
0 00000009 7ACC40C8000000 7ACC40CAFFFFFF  17:40:42  7.95e+006 2.49e+008 06 Total 28: Key 001:7A6710198F4C8CD3 002:7A6710199220C4CB 
                                                                                         003:7A67101998F80B49 004:7A6710199D046D9D 
                                                                                         005:7A671019A1A26457 006:7A671019A87C6415 
1 0000000A 7ACC40CB000000 7ACC40CDFFFFFF  17:40:42  7.68e+007 2.52e+008 03 Total 31: Key 001:7A671019BF4CD383 002:7A671019C2809225 
                                                                                         003:7A671019D957B302 
2 0000000B 7ACC40CE000000 7ACC40D0FFFFFF  17:40:43  4.61e+008 2.52e+008 04 Total 35: Key 001:7A671019E0CB20F2 002:7A67101A08C10DA7 
                                                                                         003:7A67101A0EC2F452 004:7A67101A0D8FF8EC 
3 0000000C 7ACC40D1000000 7ACC40D3FFFFFF  17:40:43  4.61e+008 2.52e+008 04 Total 39: Key 001:7A67101A1525FD6D 002:7A67101A190201CE 
                                                                                         003:7A67101A1F86E6AD 004:7A67101A2F706B1A 
0 0000000D 7ACC40D4000000 7ACC40D6FFFFFF  17:40:49  7.95e+006 2.52e+008 09 Total 48: Key 001:7A67101A4649D546 002:7A67101A462FD940 
                                                                                         003:7A67101A514AB531 004:7A67101A58EFFB10 
                                                                                         005:7A67101A5BE52601 006:7A67101A5EC84F57 
                                                                                         007:7A67101A61EFC786 008:7A67101A64B023D0 
                                                                                         009:7A67101A6883EAA1 
1 0000000E 7ACC40D7000000 7ACC40D9FFFFFF  17:40:50  6.58e+007 2.49e+008 03 Total 51: Key 001:7A67101A80202576 002:7A67101A91378F1F 
                                                                                         003:7A67101A92CB266D 
2 0000000F 7ACC40DA000000 7ACC40DCFFFFFF  17:40:50  4.61e+008 2.49e+008 04 Total 55: Key 001:7A67101AC1203110 002:7A67101AC7CDB3A2 
                                                                                         003:7A67101ACB106BBF 004:7A67101ACD9DEC07 


-Vector Test PASS exiting
 
C

cayoenrique

Member
Messages
475
Here the files with the pdf that contains the Tutorial Installations from post #6 so that it does not get lost from workupload

 
C

cayoenrique

Member
Messages
475
Here the files with the pdf that contains the Tutorial Installations from post #6 so that it does not get lost from workupload
 

Attachments

  • Tutorial_Opencl_installation_ver_0.01.tar.7z.zip
    1.2 MB · Views: 13
You must log in or register to reply here.
Top