OpenCL systems and their benchmark results #530

ZerBea · 2025-11-22T09:37:13Z

ZerBea
Nov 22, 2025
Maintainer

Starting with CUDA 13, NVIDIA decided to drop support for my GTX 1080TI. For this reason, I had to look for a cheap alternative to continue my analyses.
I finally ended up with an

AMD Ryzen 5 4600G" and a "SPARKLE Intel Arc A770 TITAN OC Edition 16GB

AMD Ryzen 5 4600G with Radeon Graphics
VGA compatible controller: Intel Corporation DG2 [Arc A770] (rev 08)

If you look at the price-performance ratio, the benchmark results are impressive:
Arch Linux

$ uname -r
6.17.8-arch1-1

hashcat WPA-PBKDF2-PMKID+EAPOL

$ hashcat -m 22000 --benchmark
hashcat (v7.1.2-377-g6c5dc354c) 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.

Kernel /usr/share/hashcat/OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

HIP API (HIP 6.4.43484)
=======================
* Device #01: AMD Radeon Graphics, 6928/6928 MB, 7MCU

OpenCL API (OpenCL 3.0 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #02: Intel(R) Arc(TM) A770 Graphics, 15473/15473 MB (4095 MB allocatable), 16MCU

OpenCL API (OpenCL 2.1 AMD-APP.dbg (3649.0)) - Platform #2 [Advanced Micro Devices, Inc.]
=========================================================================================
* Device #03: AMD Radeon Graphics, skipped

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------

Speed.#01........:    75179 H/s (82.47ms) @ Accel:14 Loops:512 Thr:512 Vec:1
Speed.#02........:   910.2 kH/s (83.47ms) @ Accel:38 Loops:1024 Thr:512 Vec:1
Speed.#*.........:   985.4 kH/s

Started: Sat Nov 22 09:58:55 2025
Stopped: Sat Nov 22 09:59:46 2025

hashcat WPA-PMK-PMKID+EAPOL

$ hashcat -m 22001 --benchmark
hashcat (v7.1.2-377-g6c5dc354c) 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.

Kernel /usr/share/hashcat/OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

HIP API (HIP 6.4.43484)
=======================
* Device #01: AMD Radeon Graphics, 6928/6928 MB, 7MCU

OpenCL API (OpenCL 3.0 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #02: Intel(R) Arc(TM) A770 Graphics, 15473/15473 MB (4095 MB allocatable), 16MCU

OpenCL API (OpenCL 2.1 AMD-APP.dbg (3649.0)) - Platform #2 [Advanced Micro Devices, Inc.]
=========================================================================================
* Device #03: AMD Radeon Graphics, skipped

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------
* Hash-Mode 22001 (WPA-PMK-PMKID+EAPOL) [Iterations: 0]
-------------------------------------------------------

Speed.#01........: 32776.5 kH/s (0.00ms) @ Accel:704 Loops:1024 Thr:1024 Vec:1
Speed.#02........:   410.8 MH/s (0.00ms) @ Accel:160 Loops:1024 Thr:1024 Vec:1
Speed.#*.........:   443.6 MH/s

Started: Sat Nov 22 10:00:34 2025
Stopped: Sat Nov 22 10:01:05 2025

john wpapsk-opencl
Unfortunately this format failed. More information is here.

$ john --format=wpapsk-opencl --test --devices=1
Device 1: Intel(R) Arc(TM) A770 Graphics
Benchmarking: wpapsk-opencl, WPA/WPA2/PMF/PMKID PSK [PBKDF2-SHA1 HMAC-SHA256/AES-CMAC OpenCL]... FAILED (cmp_all(1))

john wpapsk-pmk-opencl

$ john --format=wpapsk-pmk-opencl --test --devices=1
Device 1: Intel(R) Arc(TM) A770 Graphics
Benchmarking: wpapsk-pmk-opencl, WPA/WPA2/PMF/PMKID master key [MD5/SHA-1/HMAC-SHA256/AES-CMAC OpenCL]... LWS=64 GWS=8388608 (131072 blocks) DONE
Raw:	4978K c/s real, 5084K c/s virtual

john wpapsk-opencl

$ john --format=wpapsk-opencl --test --devices=2
Device 2: gfx90c:xnack+ [AMD Radeon Graphics]
Benchmarking: wpapsk-opencl, WPA/WPA2/PMF/PMKID PSK [PBKDF2-SHA1 HMAC-SHA256/AES-CMAC OpenCL]... LWS=64 GWS=57344 (896 blocks) DONE
Raw:	70795 c/s real, 1720K c/s virtual

john wpapsk-pmk-opencl

$ john --format=wpapsk-pmk-opencl --test --devices=2
Device 2: gfx90c:xnack+ [AMD Radeon Graphics]
Benchmarking: wpapsk-pmk-opencl, WPA/WPA2/PMF/PMKID master key [MD5/SHA-1/HMAC-SHA256/AES-CMAC OpenCL]... LWS=256 GWS=7340032 (28672 blocks) DONE
Raw:	4559K c/s real, 5115K c/s virtual

ZerBea · 2025-11-25T22:01:14Z

ZerBea
Nov 25, 2025
Maintainer Author

Running latest release of intel-compute-runtime and intel-graphics-compiler,

intel-compute-runtime 25.44.36015.5-1
intel-graphics-compiler 1:2.22.2-1

JtR is working again

$ john --format=wpapsk-opencl --test
Device 1: Intel(R) Arc(TM) A770 Graphics
Benchmarking: wpapsk-opencl, WPA/WPA2/PMF/PMKID PSK [PBKDF2-SHA1 HMAC-SHA256/AES-CMAC OpenCL]... LWS=16 GWS=8388608 (524288 blocks) DONE
Raw:	897177 c/s real, 104857K c/s virtual

$ john --format=wpapsk-pmk-opencl --test
Device 1: Intel(R) Arc(TM) A770 Graphics
Benchmarking: wpapsk-pmk-opencl, WPA/WPA2/PMF/PMKID master key [MD5/SHA-1/HMAC-SHA256/AES-CMAC OpenCL]... LWS=64 GWS=8388608 (131072 blocks) DONE
Raw:	5053K c/s real, 5162K c/s virtual

0 replies

kimocoder · 2025-11-26T16:42:03Z

kimocoder
Nov 26, 2025

$ uname -r 6.18.0-061800rc7-generic

hashcat WPA-PBKDF2-PMKID+EAPOL

$ hashcat -m 22000 --benchmark
hashcat (v7.1.2-350-g8fe0ae692) starting in benchmark mode

CUDA API (CUDA 13.0)
====================
* Device #01: NVIDIA GeForce RTX 3090, 23029/24121 MB, 82MCU

OpenCL API (OpenCL 3.0 CUDA 13.0.97) - Platform #1 [NVIDIA Corporation]
=======================================================================
* Device #02: NVIDIA GeForce RTX 3090, skipped

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------

Speed.#01........:  1141.2 kH/s (95.81ms) @ Accel:7 Loops:1024 Thr:768 Vec:1

Started: Wed Nov 26 17:38:54 2025
Stopped: Wed Nov 26 17:39:10 2025

hashcat WPA-PMK-PMKID+EAPOL

$ hashcat -m 22001 --benchmark
hashcat (v7.1.2-350-g8fe0ae692) starting in benchmark mode

CUDA API (CUDA 13.0)
====================
* Device #01: NVIDIA GeForce RTX 3090, 23025/24121 MB, 82MCU

OpenCL API (OpenCL 3.0 CUDA 13.0.97) - Platform #1 [NVIDIA Corporation]
=======================================================================
* Device #02: NVIDIA GeForce RTX 3090, skipped

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------
* Hash-Mode 22001 (WPA-PMK-PMKID+EAPOL) [Iterations: 0]
-------------------------------------------------------

Speed.#01........:   741.0 MH/s (0.00ms) @ Accel:288 Loops:1024 Thr:1024 Vec:1

Started: Wed Nov 26 17:39:33 2025
Stopped: Wed Nov 26 17:39:45 2025

0 replies

kimocoder · 2025-12-07T21:59:42Z

kimocoder
Dec 7, 2025

New test coming up with new kernel and NVIDIA support

0 replies

ZerBea · 2025-12-09T09:02:28Z

ZerBea
Dec 9, 2025
Maintainer Author

ASRock Intel Arc B580 Steel Legend 12 GB OC

AMD Ryzen 7 8700G w/ Radeon 780M Graphics
03:00.0 VGA compatible controller: Intel Corporation Battlemage G21 [Arc B580]

Linux kernel:

$ uname -r
6.17.9-arch1-1

Intel runtime:

$ pacman -Q | grep intel-
intel-compute-runtime 25.44.36015.5-1
intel-gmmlib 22.8.2-1
intel-graphics-compiler 1:2.22.2-1

hashcat WPA-PBKDF2-PMKID+EAPOL

$ hashcat -m 22000 --benchmark
hashcat (v7.1.2) 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.

Kernel /usr/share/hashcat/OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

OpenCL API (OpenCL 3.0 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #01: Intel(R) Arc(TM) B580 Graphics, 11605/11605 MB (11605 MB allocatable), 8MCU

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------

Speed.#01........:   691.6 kH/s (97.28ms) @ Accel:33 Loops:1024 Thr:1024 Vec:1

Started: Tue Dec  9 09:36:15 2025
Stopped: Tue Dec  9 09:36:33 2025

hashcat WPA-PMK-PMKID+EAPOL

$ hashcat -m 22001 --benchmark
hashcat (v7.1.2) 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.

Kernel /usr/share/hashcat/OpenCL/m22001-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

OpenCL API (OpenCL 3.0 ) - Platform #1 [Intel(R) Corporation]
=============================================================
* Device #01: Intel(R) Arc(TM) B580 Graphics, 11605/11605 MB (11605 MB allocatable), 8MCU

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------
* Hash-Mode 22001 (WPA-PMK-PMKID+EAPOL) [Iterations: 0]
-------------------------------------------------------

Speed.#01........:   698.8 MH/s (0.00ms) @ Accel:128 Loops:1024 Thr:1024 Vec:1

Started: Tue Dec  9 09:36:38 2025
Stopped: Tue Dec  9 09:36:50 2025

john wpapsk-opencl

$ john --format=wpapsk-opencl --test --devices=1
Device 1: Intel(R) Arc(TM) B580 Graphics
Benchmarking: wpapsk-opencl, WPA/WPA2/PMF/PMKID PSK [PBKDF2-SHA1 OpenCL]... DONE
Speed for cost 1 (key version [0:PMKID 1:WPA 2:WPA2 3:802.11w]) of 2
Raw:	501960 c/s real, 501960 c/s virtual

john wpapsk-pmk-opencl

$ john --format=wpapsk-pmk-opencl --test --devices=1
Device 1: Intel(R) Arc(TM) B580 Graphics
Benchmarking: wpapsk-pmk-opencl, WPA/WPA2/PMF/PMKID master key [MD5/SHA-1/SHA-2 OpenCL]... DONE
Speed for cost 1 (key version [0:PMKID 1:WPA 2:WPA2 3:802.11w]) of 1 and 2
Warning: "Many salts" test limited: 18/256
Many salts:	370085K c/s real, 370085K c/s virtual
Only one salt:	22075K c/s real, 22075K c/s virtual

0 replies

ZerBea · 2026-01-27T10:39:40Z

ZerBea
Jan 27, 2026
Maintainer Author

AMD Ryzen 5 5500GT with Radeon Graphics

For this test we use opencl-mesa!

$ clinfo -l
Platform #0: rusticl
 `-- Device #0: AMD Radeon Graphics (radeonsi, renoir, ACO, DRM 3.64, 6.18.6-arch1-1)

hashcat WPA-PBKDF2-PMKID+EAPOL

$ hashcat -m 22000 --benchmark
hashcat (v7.1.2) 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.

Kernel /usr/share/hashcat/OpenCL/m22000-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

OpenCL API (OpenCL 3.0 ) - Platform #1 [Mesa/X.org]
===================================================
* Device #01: AMD Radeon Graphics (radeonsi, renoir, ACO, DRM 3.64, 6.18.6-arch1-1), 512/512 MB (1919 MB allocatable), 7MCU

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------------
* Hash-Mode 22000 (WPA-PBKDF2-PMKID+EAPOL) [Iterations: 4095]
-------------------------------------------------------------

Speed.#01........:    57775 H/s (92.34ms) @ Accel:24 Loops:512 Thr:256 Vec:1

Started: Tue Jan 27 11:35:30 2026
Stopped: Tue Jan 27 11:35:38 2026

hashcat WPA-PMK-PMKID+EAPOL

$ hashcat -m 22001 --benchmark
hashcat (v7.1.2) 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.

Kernel /usr/share/hashcat/OpenCL/m22001-optimized.cl:
Optimized kernel requested, but not available or not required
Falling back to pure kernel

OpenCL API (OpenCL 3.0 ) - Platform #1 [Mesa/X.org]
===================================================
* Device #01: AMD Radeon Graphics (radeonsi, renoir, ACO, DRM 3.64, 6.18.6-arch1-1), 512/512 MB (1919 MB allocatable), 7MCU

Benchmark relevant options:
===========================
* --backend-devices-virtmulti=1
* --backend-devices-virthost=1
* --optimized-kernel-enable

-------------------------------------------------------
* Hash-Mode 22001 (WPA-PMK-PMKID+EAPOL) [Iterations: 0]
-------------------------------------------------------

Speed.#01........: 29750.4 kH/s (0.00ms) @ Accel:48 Loops:1024 Thr:1024 Vec:1

Started: Tue Jan 27 11:36:34 2026
Stopped: Tue Jan 27 11:36:50 2026

0 replies

Navigation Menu

OpenCL systems and their benchmark results #530

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Replies: 5 comments

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

OpenCL systems and their benchmark results #530

Uh oh!

Uh oh!

ZerBea Nov 22, 2025 Maintainer

AMD Ryzen 5 4600G" and a "SPARKLE Intel Arc A770 TITAN OC Edition 16GB

Replies: 5 comments

Uh oh!

ZerBea Nov 25, 2025 Maintainer Author

Uh oh!

Uh oh!

kimocoder Nov 26, 2025

hashcat WPA-PBKDF2-PMKID+EAPOL

hashcat WPA-PMK-PMKID+EAPOL

Uh oh!

kimocoder Dec 7, 2025

Uh oh!

Uh oh!

ZerBea Dec 9, 2025 Maintainer Author

ASRock Intel Arc B580 Steel Legend 12 GB OC

Uh oh!

ZerBea Jan 27, 2026 Maintainer Author

AMD Ryzen 5 5500GT with Radeon Graphics

ZerBea
Nov 22, 2025
Maintainer

ZerBea
Nov 25, 2025
Maintainer Author

kimocoder
Nov 26, 2025

kimocoder
Dec 7, 2025

ZerBea
Dec 9, 2025
Maintainer Author

ZerBea
Jan 27, 2026
Maintainer Author