why Hashcat v6.2.5 is slower than v5.1.0 - Printable Version +- hashcat Forum (https://hashcat.net/forum) +-- Forum: Support (https://hashcat.net/forum/forum-3.html) +--- Forum: hashcat (https://hashcat.net/forum/forum-45.html) +--- Thread: why Hashcat v6.2.5 is slower than v5.1.0 (/thread-10548.html) |
[Solved] why Hashcat v6.2.5 is slower than v5.1.0 - v71221 - 02-28-2022 Hooray I found this link *ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/13794/opencl_runtime_18.1_x64_setup.msi Now it takes only 11 seconds instead of 16 minutes. Thank you. Code: C:\hashcat-6.2.5-207>hashcat.exe -I Code: C:\hashcat-6.2.5-207>hashcat.exe -D 1 -a 3 -m 22000 "WPA*01*ca5396d611cf330aebefd48ebbfb0e63*020000000001*020000000020*61703031***" "12345678" I think thread name should be changed to [Solved] why Hashcat v6.2.5 is slower than v5.1.0 RE: why Hashcat v6.2.5 is slower than v5.1.0 - v71221 - 03-28-2022 I managed to run Hashcat v6.2.5 on Linux. For simplicity, I run Ubuntu directly from a USB stick without installing it on a hard disk. I chose Ubuntu because it is recommended by Hashcat team. https://hashcat.net/forum/thread-6620-post-35318.html#pid35318 https://hashcat.net/forum/thread-6661-post-40381.html#pid40381 https://hashcat.net/forum/thread-7639-post-41025.html#pid41025 Download an Ubuntu image (why version 16.04 ? Because prerequisite, see below) https://releases.ubuntu.com/16.04.7/ubuntu-16.04.7-desktop-amd64.iso Create a bootable USB flash drive https://ubuntu.com/tutorials/create-a-usb-stick-on-windows Boot from USB flash drive and choose Try Ubuntu without installing Code: ubuntu@ubuntu:~$ uname -r Then you need to download Intel CPU Runtime for OpenCL Applications 18.1 for Linux OS (64bit only). See here https://www.intel.com/content/www/us/en/developer/articles/tool/opencl-drivers.html Or use direct links *ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/15532/l_opencl_p_18.1.0.015.tgz or *ttps://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/15532/l_opencl_p_18.1.0.015.tgz Supported operating systems for this release are: - CentOS 7 (Intel(R) 64) - Red Hat Enterprise Linux 7 (Intel(R) 64) - SUSE Linux Enterprise Server 12 (Intel(R) 64) - Ubuntu 16.04 (Intel(R) 64) Then sudo add-apt-repository universe sudo apt update sudo apt install lsb-core (prerequisite for install.sh) sudo apt install p7zip (needed for unpacking 7z-archives) Unpack tgz-file tar -xzf l_opencl_p_18.1.0.015.tgz Run ./install.sh Download https://hashcat.net/files/hashcat-6.2.5.7z Unpack 7z-file p7zip -d hashcat-6.2.5.7z I got a working hashcat v6.2.5 Code: ubuntu@ubuntu:~/hashcat-6.2.5$ ./hashcat.bin -I Some info was helpful for me here https://www.alexanderjsingleton.com/infosexy-how-to-use-hashcat-to-crack-passwords-in-ubuntu-18-04/ RE: why Hashcat v6.2.5 is slower than v5.1.0 - GWN - 04-21-2022 (02-28-2022, 11:37 AM)Snoopy Wrote: yeah the GPU driver has opencl support but as i siad earlier, the intel opencl runtime for GPU (your used driver) is BROKEN !!! deinstall (a really clean deinstall) it and use the CPU only runtime 18.1 taken directly from the intel website sorry but I cannot find the download link for OpenCl 18.1 for win OS. Any chance you can point me to it? Also BTW would you be able to tell me if this hardware config can even run Hashcat: Hewlett-Packard HP xw6600 Workstation
PROCESSOR Intel® Xeon® CPU E5410 @ 2.33GHz
GRAPHICS = 2 NVIDIA Quadro FX 570
If it can, please point me to a link to how to get Hashcat to load up properly. I have installed 6.2.5 and have installed CUDA in C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6; but any attempt at running it yields: C:\Users\xyz\AppData\Local\Programs\hashcat-6.2.5>hashcat -b -m 14600 -D 1 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. cuFuncSetAttribute is missing from CUDA shared library. clCompileProgram is missing from OpenCL shared library. ATTENTION! No OpenCL, HIP or CUDA installation found. You are probably missing the CUDA, HIP or OpenCL runtime installation. * AMD GPUs on Windows require this driver: "AMD Radeon Adrenalin 2020 Edition" (21.2.1 or later) * Intel CPUs require this runtime: "OpenCL Runtime for Intel Core and Intel Xeon Processors" (16.1.1 or later) * NVIDIA GPUs require this runtime and/or driver (both): "NVIDIA Driver" (440.64 or later) "CUDA Toolkit" (9.0 or later) Thank you in advance for helping a complete newbie. RE: why Hashcat v6.2.5 is slower than v5.1.0 - slyexe - 04-22-2022 Install the Nvidia Drivers from official website, that would install the OpenCL runtimes. However, you're GPU does not support CUDA SDK so it would not have the appropriate drivers for using CUDA runtimes. Therefore, don't post issues regarding errors with hashcat as it would be hardware/driver related. RE: why Hashcat v6.2.5 is slower than v5.1.0 - Snoopy - 04-22-2022 *ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/13794/opencl_runtime_18.1_x64_setup.msi this is the direct link for cpu opencl install this and see the output for hashcat -I your cpu should show up with opencl 2.1 RE: why Hashcat v6.2.5 is slower than v5.1.0 - GWN - 04-24-2022 (04-22-2022, 01:39 AM)slyexe Wrote: Install the Nvidia Drivers from official website, that would install the OpenCL runtimes. However, you're GPU does not support CUDA SDK so it would not have the appropriate drivers for using CUDA runtimes. Therefore, don't post issues regarding errors with hashcat as it would be hardware/driver related. Thank you slyexe for your prompt response. Sorry for my delay as I was distracted by other projects. To be clear I was not posting the errors to imply a a problem with hashcat but to document my inability to install it and accompanying tools correctly - so totally on me. As per my comment I did download the NVIDIA drivers and they are installed in C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6 which are part of the path. My confusion stemmed from the reference to the CUDA not being present in "ATTENTION! No OpenCL, HIP or CUDA installation found." when I could see it on my computer. So as a (new) user seeing that error is very confusing. As you will see in my next response to Snoopy, I have followed his instructions and I'm still unsure of my installation. Thank you for your support. RE: why Hashcat v6.2.5 is slower than v5.1.0 - GWN - 04-24-2022 (04-22-2022, 04:05 PM)Snoopy Wrote: *ttp://registrationcenter-download.intel.com/akdlm/irc_nas/vcp/13794/opencl_runtime_18.1_x64_setup.msiThank you Snoopy for your prompt response. Sorry for my delay as I was distracted by other projects. Thank you so much for your link to OpenCL, it was brilliant. My initial attempts at finding this sent me to all sorts of weird Intel sites that wanted me to sign up to request authorization to download, so very strange. Yours was a godsend, easiest install I have had in a long time. After rebooting (yes I know that should be obvious) I now have my computer showing up with 2.1 as you said: C:\Users\xyz\AppData\Local\Programs\hashcat-6.2.5>hashcat -I hashcat (v6.2.5) starting in backend information mode cuFuncSetAttribute is missing from CUDA shared library. clCreateContext(): CL_DEVICE_NOT_AVAILABLE 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) Xeon(R) CPU E5410 @ 2.33GHz Version........: OpenCL 2.1 (Build 0) Processor(s)...: 4 Clock..........: 2330 Memory.Total...: 32767 MB (limited to 4095 MB allocatable in one block) Memory.Free....: 0 MB OpenCL.Version.: OpenCL C 2.0 Driver.Version.: 18.1.0.0920 Then running: C:\Users\xyz\AppData\Local\Programs\hashcat-6.2.5>hashcat -b -m 14600 -D 1 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. cuFuncSetAttribute is missing from CUDA shared library. clCreateContext(): CL_DEVICE_NOT_AVAILABLE OpenCL API (OpenCL 2.1 WINDOWS) - Platform #1 [Intel(R) Corporation] ==================================================================== * Device #1: Intel(R) Xeon(R) CPU E5410 @ 2.33GHz, skipped Benchmark relevant options: =========================== * --opencl-device-types=1 * --optimized-kernel-enable --------------------------------------------- * Hash-Mode 14600 (LUKS) [Iterations: 163044] --------------------------------------------- Then for reference my windows paths are:
I searched for cuFuncSetAttribute on my C drive and windows found this: Name Folder; Date modified; Type; Size cuda.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1.6); 2021-12-17 8:06 PM; H File; 791 KB cudaTypedefs.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1.6); 2021-12-17 8:06 PM; H File 89 KB cuda..occupancy.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1.6); 2021-12-17 8:06 PM; H File 65KB generated_cuda_meta.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1 .6\compute-sanitizer); 2022-03-08 1OE22 PM; H File; 77 KB sanitizer_driver_cbid.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vll.6\compute-sanitizer); 2022-03-08 1OE22 PM; H File; 84KB cuptLdriver_cbid.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1 .6\extras\CUP11); 2022-03-08 7:52 PM; H File; 64 KB generated_cuda_meta.h include (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1 .6\extras\CUP11); 2022-03-08 7:52 PM; H File; 77 KB I also searched for clCreateContext on my C drive and windows found: Name Folder; Date modified; Type; Size cl.h CL (C\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vll.6\include); 2021-12-17 8:06 PM; H File; 78KB cl.hpp CL (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1 .6\include); 2021-12-17 8:06 PM; HPP File; 287 KB cLext.h CL (C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vl 1.6\include); 2021-12-17 8:06 PM; H File; 47KB These results only add to my confusion... My questions now are:
RE: why Hashcat v6.2.5 is slower than v5.1.0 - slyexe - 04-25-2022 1. Is in regards to CUDA which you cannot use, so any error relating to it would be irrelevant to your hardware. 2. You're requesting for a GPU (-D 1) in which your system does not have installed based on your details posted above from hashcat -I 3. No 4. Either use -D 2 or -d 1 to declare what device to use. See wiki for details regarding -D and -d RE: why Hashcat v6.2.5 is slower than v5.1.0 - GWN - 04-26-2022 (04-25-2022, 12:10 AM)slyexe Wrote: 1. Is in regards to CUDA which you cannot use, so any error relating to it would be irrelevant to your hardware. Thank you slyexe #1 Okay, easy enough to disregard the cuFuncSetAttribute message. #3 Thank you - great to hear. #2 I'm confused does the command hashcat -I have an implied -D? Reason I'm asking is that command also gave the CL_DEVICE_NOT_AVAILABLE error. #4 sorry I've scoured the wiki and I cannot find any info on -D or -d. Do you have a link I can use to learn about declaring devices? BTW I searched the forum for declare device and got nothing that was helpful. Then I searched for -D but that's way too general and I got 468 pages of hits. Thank you RE: why Hashcat v6.2.5 is slower than v5.1.0 - GWN - 04-27-2022 (04-26-2022, 07:56 PM)GWN Wrote:(04-25-2022, 12:10 AM)slyexe Wrote: 1. Is in regards to CUDA which you cannot use, so any error relating to it would be irrelevant to your hardware.Thank you slyexe Hi in my travels through the wiki/forum I found this page https://hashcat.net/wiki/doku.php?id=hashcat that helps with #4. -d is --backend-devices which means Backend devices to use and -D is --opencl-device-types which in turn means OpenCL device-types to use Then lower down it states the OpenCL Device Types. It says that Device Type = 1 is CPU So why would declaring -D 1 cause a GPU not found error? Also I have been trying out different options and every time it says that my CPU is "skipped" as per: C:\Users\xyz\AppData\Local\Programs\hashcat>hashcat -b -m 16500 -D 1 -d 1 -w 4 hashcat (v6.2.5) starting in benchmark mode cuFuncSetAttribute is missing from CUDA shared library. clCreateContext(): CL_DEVICE_NOT_AVAILABLE OpenCL API (OpenCL 2.1 WINDOWS) - Platform #1 [Intel(R) Corporation] ==================================================================== * Device #1: Intel(R) Xeon(R) CPU E5410 @ 2.33GHz, skipped Benchmark relevant options: =========================== * --backend-devices=1 * --opencl-device-types=1 * --workload-profile=4 ---------------------------------------- * Hash-Mode 16500 (JWT (JSON Web Token)) ---------------------------------------- Started: Wed Apr 27 09:53:05 2022 Stopped: Wed Apr 27 09:53:08 2022 Why does it "skip" my device even though I've used -D and -d both set to 1? What am I missing? Thank you |