gpu utilization

[mtr0ll]

Hi, what is the gpu utilization? When I run oclHashcat on my two 7970s I never get above 85%. I don't have anything else running that could use the gpus except for the x11 server that is required to initialize the cards.

[atom]

Give it enough work to do otherwise they can not reach full performance.

For example:

- if you use -a 0, -a 1 or -a 6, make sure the base dictionary (left) has at least 10 mio words (per gpu).
- if you use -a 3 or -a 7, make sure your left mask keyspace > 500k entries (per gpu). for -a 3, this is the keyspace of the first 4 chars of your mask.
- make sure your modifier (right) iterates at least through a keyspace of 1k. in case if -a 0 its the number of rules, in case of -a 1 and -a 7 its the number of words in the right dictionary, in case of -a 6 its the keyspace of the right mask.

The workload modifiers are set to low.

If you dont care about screen lags, raise -n and --gpu-loops. For example -n 160 --gpu-loops 1024.

[mtr0ll]

Give it enough work to do otherwise they can not reach full performance.

For example:

- if you use -a 0, -a 1 or -a 6, make sure the base dictionary (left) has at least 10 mio words (per gpu).
- if you use -a 3 or -a 7, make sure your left mask keyspace > 500k entries (per gpu). for -a 3, this is the keyspace of the first 4 chars of your mask.
- make sure your modifier (right) iterates at least through a keyspace of 1k. in case if -a 0 its the number of rules, in case of -a 1 and -a 7 its the number of words in the right dictionary, in case of -a 6 its the keyspace of the right mask.

The workload modifiers are set to low.

If you dont care about screen lags, raise -n and --gpu-loops. For example -n 160 --gpu-loops 1024.

thanks!

[mtr0ll]

I'm using hashcat lite with ?a on a single ntlm hash. I've added -n 160 --gpu-loops 1024 but that does not seem to make any difference, gpu utilization stays at 85%.

[M@LIK]

And your command line is?

@atom:: Thanks for the valuable info!

[mtr0ll]

And your command line is?

I'm using hashcat lite with ?a on a single ntlm hash. I've added -n 160 --gpu-loops 1024

the above?

./oclHashcat-lite64.bin -m 1000 -1 ?a -n 160 --gpu-loops 1024

[atom]

Then is using the default mask.

Please read this article to understand how masks work: https://hashcat.net/wiki/doku.php?id=mask_attack

[mtr0ll]

So my mask is ?1?2?2?2?2?2?2?3?3?3 for a password of length 10.

- if you use -a 3 or -a 7, make sure your left mask keyspace > 500k entries (per gpu). for -a 3, this is the keyspace of the first 4 chars of your mask.

Since I'm using the lite version, I'm assuming it's equivalent to -a 3 of the plus version? In that case, why are the only first 4 chars are relevant? What mask do I need to get close to 90% for example?

The other thing I found interesting is that when I swapped the processor & motherboard to amd fx8150, the utilization dropped from 85% (with intel processor) to 81% on both cards. Both motherboards use PCIE2.0 16x and I didn't change the command line for hashcat after I swapped the processor + mobo.

[KT819GM]

--restore-timer=60
Also disable integrated gpu if it exist on mobo, disable-remove crossfire.

[mtr0ll]

--restore-timer=60
Also disable integrated gpu if it exist on mobo, disable-remove crossfire.

--restore-timer=60 didn't make any difference. There is no integrated gpu and I'm not using crossfire