HELP WITH GPU SPEED
#1
Exclamation 
Hi I am new to all of this, have spend a lot of time on the hashcat wiki.

My question is.. are my GPUs speeds considerable good or bad ?

I tried to put all the GPUs / Hashcat info bellow, to help.

Now I am trying to crack a 90+gb wordlist using a ssd, idk if that matters.

Idk if I should OC the GPUs in another way, or maybe dont OC at all, I am also using power limits around 70%, in msi after burner.

I am using this command:

hashcat -D 2 -d 1,2,3,4,5,6 -m 22000 file.hc22000 namewordslist.txt

Should change the command to improve something?

rn the setup is only used for hashcat, I give the commands and let it run.

I also got considarable RAM.

My CPU idk if is cosiderable good for hashcat ?! Do I need a upgrade ?

I am still a newbie, did I config this right ? should I modify something ?

Any help is appreciated !

Thank you !

Speed.#1.........:  400.3 kH/s (6.20ms) @ Accel:8 Loops:128 Thr:256 Vec:1
Speed.#2.........:  409.5 kH/s (6.21ms) @ Accel:8 Loops:128 Thr:256 Vec:1
Speed.#3.........:  443.7 kH/s (9.98ms) @ Accel:16 Loops:128 Thr:256 Vec:1
Speed.#4.........:  280.5 kH/s (9.87ms) @ Accel:16 Loops:128 Thr:256 Vec:1
Speed.#5.........:  217.4 kH/s (11.54ms) @ Accel:16 Loops:128 Thr:256 Vec:1
Speed.#6.........:  259.7 kH/s (10.10ms) @ Accel:64 Loops:128 Thr:64 Vec:1
CUDA API (CUDA 12.0)
====================
* Device #1: NVIDIA GeForce RTX 3070, 7113/8191 MB, 46MCU
* Device #2: NVIDIA GeForce RTX 3070, 7113/8191 MB, 46MCU
* Device #3: NVIDIA GeForce RTX 3060 Ti, 7142/8191 MB, 38MCU
* Device #4: NVIDIA GeForce GTX 1660 Ti, 5143/6143 MB, 24MCU
* Device #5: NVIDIA GeForce GTX 1660, 5148/6143 MB, 22MCU
* Device #6: NVIDIA GeForce GTX 1660 SUPER, 5148/6143 MB, 22MCU

OpenCL API (OpenCL 3.0 CUDA 12.0.139) - Platform #1 [NVIDIA Corporation]
========================================================================
* Device #7: NVIDIA GeForce RTX 3070, skipped
* Device #8: NVIDIA GeForce RTX 3070, skipped
* Device #9: NVIDIA GeForce RTX 3060 Ti, skipped
* Device #10: NVIDIA GeForce GTX 1660 Ti, skipped
* Device #11: NVIDIA GeForce GTX 1660, skipped
* Device #12: NVIDIA GeForce GTX 1660 SUPER, skipped

OpenCL API (OpenCL 1.2 ) - Platform #2 [Intel(R) Corporation]
=============================================================
* Device #13: Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz, skipped
* Device #14: Intel(R) HD Graphics 2500, skipped

Speed.#*.........:  2011.0 kH/s
Recovered........: 24/465 (5.16%) Digests (total), 0/465 (0.00%) Digests (new), 8/39 (20.51%) Salts
Progress.........: 156342866317/329903349321 (47.39%)
Rejected.........: 853839831/156342866317 (0.55%)
Restore.Point....: 4007722609/8459060239 (47.38%)
Restore.Sub.#1...: Salt:7 Amplifier:0-1 Iteration:2432-2560
Restore.Sub.#2...: Salt:19 Amplifier:0-1 Iteration:0-1
Restore.Sub.#3...: Salt:0 Amplifier:0-1 Iteration:0-1
Restore.Sub.#4...: Salt:24 Amplifier:0-1 Iteration:0-1
Restore.Sub.#5...: Salt:8 Amplifier:0-1 Iteration:1792-1920
Restore.Sub.#6...: Salt:33 Amplifier:0-1 Iteration:3712-3840
Candidate.Engine.: Device Generator
Candidates.#1....: adosdg@hasng.com -> aasdaBsads123
Candidates.#2....: adoasph.halley -> adolphins72
Candidates.#3....: adoBs11 -> adaull
Candidates.#4....: adolfo.valliere -> adah.hallett
Candidates.#5....: adolphins8 -> adsdg@chasdyte.com
Candidates.#6....: adosadobjeCt -> adsde76341
Hardware.Mon.#1..: Temp: 56c Fan: 56% Util: 93% Core:1425MHz Mem:7800MHz Bus:1
Hardware.Mon.#2..: Temp: 56c Fan: 56% Util: 94% Core:1425MHz Mem:7800MHz Bus:1
Hardware.Mon.#3..: Temp: 54c Fan: 31% Util:  0% Core:1785MHz Mem:7800MHz Bus:1
Hardware.Mon.#4..: Temp: 54c Fan: 57% Util: 97% Core:1920MHz Mem:5951MHz Bus:1
Hardware.Mon.#5..: Temp: 52c Fan: 49% Util: 97% Core:1530MHz Mem:4600MHz Bus:1
Hardware.Mon.#6..: Temp: 59c Fan: 29% Util: 97% Core:1785MHz Mem:7650MHz Bus:1
Reply
#2
Just another metion.

The Util % of the GPUs keep changing rn it is around 90% on all of the gpus.

Some times it goes to zero for short period, like in the info I posted above.

Is that normal ?
Reply
#3
you could run a benchmark to see the "theoretically" max hash/s per card for this hashmode and compare it with your output (add -D2 when needed)

hashcat -b -m22000

jfyi the speed of plain wordlist is always slower, you can combine your wordlist with rules to take some advantages of the gpus

i think it drops to 0 when hashcat copies a new bulk of candidates to the gpu ram, so dont worry
Reply
#4
(01-27-2023, 02:42 PM)Snoopy Wrote: you could run a benchmark to see the "theoretically" max hash/s per card for this hashmode and compare it with your output (add -D2 when needed)

hashcat -b -m22000

jfyi the speed of plain wordlist is always slower, you can combine your wordlist with rules to take some advantages of the gpus

i think it drops to 0 when hashcat copies a new bulk of candidates to the gpu ram, so dont worry

What kind of rule can I use, also can you please give an example.

So.. wordlists are the less efficient way ?

Should I be trying something else, instead of wordlists ?

in what situation shold I use command -D2 ?

Thanks.
Reply
#5
(01-27-2023, 02:42 PM)Snoopy Wrote: you could run a benchmark to see the "theoretically" max hash/s per card for this hashmode and compare it with your output (add -D2 when needed)

hashcat -b -m22000

jfyi the speed of plain wordlist is always slower, you can combine your wordlist with rules to take some advantages of the gpus

i think it drops to 0 when hashcat copies a new bulk of candidates to the gpu ram, so dont worry  fnf

Many thanks bro. Will try that
Reply
#6
(03-29-2023, 03:45 PM)chainyc Wrote:
(01-27-2023, 02:42 PM)Snoopy Wrote: you could run a benchmark to see the "theoretically" max hash/s per card for this hashmode and compare it with your output (add -D2 when needed)

hashcat -b -m22000

jfyi the speed of plain wordlist is always slower, you can combine your wordlist with rules to take some advantages of the gpus

i think it drops to 0 when hashcat copies a new bulk of candidates to the gpu ram, so dont worry  fnf

Many thanks bro. Will try that

Have you tried it?
Reply