Posts: 4
Threads: 1
Joined: May 2022
05-03-2022, 04:17 AM
(This post was last modified: 05-03-2022, 04:35 AM by Helfrigg.)
Code: hashcat (v6.2.2) starting in benchmark mode...
Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.
OpenCL API (OpenCL 1.2 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #1: Intel(R) HD Graphics 4400, 1565/1629 MB (407 MB allocatable), 12MCU
* Device #2: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz, skipped
Benchmark relevant options:
===========================
* --optimized-kernel-enable
Hashmode: 22000 - WPA-PBKDF2-PMKID+EAPOL (Iterations: 4095)
Speed.#1.........: 2498 H/s (75.96ms) @ Accel:128 Loops:64 Thr:8 Vec:1
Started: Tue May 03 03:08:09 2022
Stopped: Tue May 03 03:08:15 2022
...
Code: hashcat (v6.2.5) starting in benchmark mode
Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.
OpenCL API (OpenCL 1.2 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #1: Intel(R) HD Graphics 4400, 768/1629 MB (203 MB allocatable), 12MCU
* Device #2: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz, skipped
Benchmark relevant options:
===========================
* --optimized-kernel-enable
-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------
* Device #1: Not enough allocatable device memory for this attack.
...
The first version I downloaded hashcat was 6.2.1. Then I updated to 6.2.2. I have been using this version for a while. Now, when I want to use the 6.2.5 version, it says "Not enough allocatable device memory for this attack."
I have 8gb DDR3 ram. Intel Graphics Drivers latest update installed. I think only one Windows 10 version has changed since I was using it. I did not install OpenCl Runtime Driver separately. This is the version that comes with the graphic drive.
I don't want to use CPU unless necessary. I don't know if this is a problem. I would like to use the recent versions of Hashcat if possible.
Thanks in Advance.
Posts: 374
Threads: 0
Joined: Nov 2017
You're usable graphics memory has changed since trying each version. Something was using the integrated gpu's memory when you ran the newer version of hashcat.
See:
6.2.2 -> * Device #1: Intel(R) HD Graphics 4400, 1565/1629 MB (407 MB allocatable), 12MCU
6.2.5 -> * Device #1: Intel(R) HD Graphics 4400, 768/1629 MB (203 MB allocatable), 12MCU
So whatever is happening on the otherside I cannot relate to but without having enough GPU memory typical task such as benchmarks or mask attacks won't work.
Posts: 4
Threads: 1
Joined: May 2022
05-05-2022, 01:43 AM
(This post was last modified: 05-05-2022, 01:45 AM by Helfrigg.)
(05-04-2022, 04:42 AM)slyexe Wrote: You're usable graphics memory has changed since trying each version. Something was using the integrated gpu's memory when you ran the newer version of hashcat.
See:
6.2.2 -> * Device #1: Intel(R) HD Graphics 4400, 1565/1629 MB (407 MB allocatable), 12MCU
6.2.5 -> * Device #1: Intel(R) HD Graphics 4400, 768/1629 MB (203 MB allocatable), 12MCU
So whatever is happening on the otherside I cannot relate to but without having enough GPU memory typical task such as benchmarks or mask attacks won't work.
Thanks for the answer. I still opened an issue in Github. I also started thinking that there was a problem with the driver. I downloaded Driver with Intel's own tool.
Because there is another strange situation. I usually used hashcat with a dictionary attack. I didn't use it because Mask Attack could ask for a lot of time. I used it a few times with CPU and succeeded. Now I realize that when I use it with GPU, the mask cannot find the password. I specified the letters with a password that I know of 4 letters and 4 numbers, I searched the numbers, the results cannot be found. For now, I just noticed this in the .hc22000 format. I will try other hashs.
Posts: 889
Threads: 15
Joined: Sep 2017
05-05-2022, 04:37 PM
(This post was last modified: 05-05-2022, 04:38 PM by Snoopy.)
Code: OpenCL API (OpenCL 1.2 )
this is really old and you will run into some troubles using these old opencl api with hahscat
next thing opencl on intel iGPU is broken by default
you have to! (plz just believe me)
deinstall intel hd graficsdriver with windows programs (it is there) and install
*ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/13794/opencl_runtime_18.1_x64_setup.msi
this will install opencl runtime 2.1 for cpu only (the speed will be the same) due to the combined cpu/gpu architecture
see whole thread here for more informations
https://hashcat.net/forum/thread-10548.html
the only thing to be aware of, you have to add at least
to your command line to tell hashcat to stick to your cpu
after de- and installing see output of telling your cpu with opencl 2.1
Posts: 4
Threads: 1
Joined: May 2022
(05-05-2022, 04:37 PM)Snoopy Wrote: Code: OpenCL API (OpenCL 1.2 )
this is really old and you will run into some troubles using these old opencl api with hahscat
next thing opencl on intel iGPU is broken by default
you have to! (plz just believe me)
deinstall intel hd graficsdriver with windows programs (it is there) and install
*ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/13794/opencl_runtime_18.1_x64_setup.msi
this will install opencl runtime 2.1 for cpu only (the speed will be the same) due to the combined cpu/gpu architecture
see whole thread here for more informations
https://hashcat.net/forum/thread-10548.html
the only thing to be aware of, you have to add at least
to your command line to tell hashcat to stick to your cpu
after de- and installing see output of telling your cpu with opencl 2.1
I downloaded and installed the Runtime from the link you gave. I uninstalled the graphics driver before installing. I restarted. Version 6.2.5 still allocates less memory than 6.2.2. I wrote the hashcat -I command without installing the graphics driver.
Code: hashcat (v6.2.5) starting in backend information mode
OpenCL Info:
============
OpenCL Platform ID #1
Vendor..: Intel(R) Corporation
Name....: Intel(R) CPU Runtime for OpenCL(TM) Applications
Version.: OpenCL 2.1 WINDOWS
Backend Device ID #1
Type...........: CPU
Vendor.ID......: 8
Vendor.........: Intel(R) Corporation
Name...........: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz
Version........: OpenCL 2.1 (Build 0)
Processor(s)...: 4
Clock..........: 3600
Memory.Total...: 8064 MB (limited to 1008 MB allocatable in one block)
Memory.Free....: 4000 MB
OpenCL.Version.: OpenCL C 2.0
Driver.Version.: 18.1.0.0920
Code: hashcat (v6.2.2) starting in backend information mode...
OpenCL Info:
============
OpenCL Platform ID #1
Vendor..: Intel(R) Corporation
Name....: Intel(R) CPU Runtime for OpenCL(TM) Applications
Version.: OpenCL 2.1 WINDOWS
Backend Device ID #1
Type...........: CPU
Vendor.ID......: 8
Vendor.........: Intel(R) Corporation
Name...........: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz
Version........: OpenCL 2.1 (Build 0)
Processor(s)...: 4
Clock..........: 3600
Memory.Total...: 8064 MB (limited to 2016 MB allocatable in one block)
Memory.Free....: 8000 MB
OpenCL.Version.: OpenCL C 2.0
Driver.Version.: 18.1.0.0920
I downloaded and reinstalled the graphics driver. I couldn't use the cpu again. I can't use runtime with graphics driver. Maybe that's how it should be.
Code: D:\hashcat>hashcat -m 0 -b -D 1
hashcat (v6.2.3) starting in benchmark mode
Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.
OpenCL API (OpenCL 2.1 WINDOWS) - Platform #1 [Intel(R) Corporation]
====================================================================
* Device #1: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz, 8000/8064 MB (2016 MB allocatable), 4MCU
OpenCL API (OpenCL 1.2 ) - Platform #2 [Intel(R) Corporation]
=============================================================
* Device #2: Intel(R) HD Graphics 4400, skipped
* Device #3: Intel(R) Core(TM) i3-4160 CPU @ 3.60GHz, 8000/8064 MB (2016 MB allocatable), 4MCU
Benchmark relevant options:
===========================
* --opencl-device-types=1
* --optimized-kernel-enable
Hashmode: 0 - MD5
It's always worked for me when it comes to cpu usage. Even if the OpenCL driver is outdated. But I don't want to use the cpu for a long time.
Regarding the graphics driver; It's clear I have a problem. I tested my own created md5s. Dictionary attack works fine. Mask attack if the 'password' consists of 5 digits, mask attack cracks the password. But if it's longer than 5, it mostly can't find the password, sometimes it can find the password if I give a hint. I also tried --markov-disable command for this, same result.
Posts: 889
Threads: 15
Joined: Sep 2017
05-09-2022, 03:40 PM
(This post was last modified: 05-09-2022, 03:42 PM by Snoopy.)
it doesnt matter, your CPU is your GPU and vice versa (you have an integrated chip)
use opencl 2.1 with your cpu (your iGPU is inside your CPU so its basically the same, the opencl instructions send to cpu/gpu will be the same)
dont give anything about the ram as the cpu will be able to use all of your free ram, your gpu ist fixed (if you are lucky you can change this inside the bios) but mostly the "RAM" given to the iGPU ist fixed to 1 or 2 GB (which is also de facto taken from your global ram)
yeah, therefore deinstall the crappy graficsdriver, you dont need this
Posts: 889
Threads: 15
Joined: Sep 2017
(05-09-2022, 02:08 PM)Helfrigg Wrote: Regarding the graphics driver; It's clear I have a problem. I tested my own created md5s. Dictionary attack works fine. Mask attack if the 'password' consists of 5 digits, mask attack cracks the password. But if it's longer than 5, it mostly can't find the password, sometimes it can find the password if I give a hint. I also tried --markov-disable command for this, same result.
because the opencl for igpu is broken, dont use this, as you already said, it will be likely, that even when providing the right pass, hashcat will tell you NO, no pass found, so in this case every cracking run will be useless
Posts: 4
Threads: 1
Joined: May 2022
(05-09-2022, 03:44 PM)Snoopy Wrote: (05-09-2022, 02:08 PM)Helfrigg Wrote: Regarding the graphics driver; It's clear I have a problem. I tested my own created md5s. Dictionary attack works fine. Mask attack if the 'password' consists of 5 digits, mask attack cracks the password. But if it's longer than 5, it mostly can't find the password, sometimes it can find the password if I give a hint. I also tried --markov-disable command for this, same result.
because the opencl for igpu is broken, dont use this, as you already said, it will be likely, that even when providing the right pass, hashcat will tell you NO, no pass found, so in this case every cracking run will be useless
Thanks for your answers. :-) How long would you run it without damaging a CPU like mine? I haven't had any overheating problems so far. But it didn't run 100% for a long time.
Posts: 889
Threads: 15
Joined: Sep 2017
as long as your cooling solution will do their job you can run hahscat as long as you want, hahscat will stop by default if your cpu reaches 90 degrees which is still a safe temperature for most cpus
|