You are not logged in.

  • Login

Dear visitor, welcome to ocaholic - Board. If this is your first visit here, please read the Help. It explains in detail how this page works. To use all features of this page, you should consider registering. Please use the registration form, to register here or read more information about the registration process. If you are already registered, please login here.

1

Wednesday, April 4th 2012, 9:00pm

MD5 Brute Force with CUDA

Have you ever benched your system with MD5 brute force tools supporting CUDA?

There are significant improvements for each new graphics card that is released.
The most performant tool seems to be Whitepixel. But since its only for Linux its not very useful to me.
With oclhashcatlite my Geforce 580 @810/1620/2025 can process ~2.3 billions (milliarden) password hashes/sec.
This tool is only for GPU

BarsWF uses also the CPU but is not the fastest with GPU processing. But only to see how much my 580 GTX with CUDA performs in comparison to my 6 Core (3.2GHz) XEON with 12 Threads.


Others have more cards...;
Source

This post has been edited 3 times, last edit by "Triton" (Apr 4th 2012, 9:08pm)


mr.k

Mainboard Mörder

Posts: 81

Occupation: Lehrling Informatiker mit Schwerpunkt Systemtechnik 3.Lj.

  • Send private message

2

Wednesday, April 4th 2012, 9:08pm

I used it for only for hacking :D


Kollege: - Ocen tut man mit Trockenfleisch :pinch:
Ich: - LOL:roflmao:
Kollege: - Oh man ich han so Hunger..

3

Wednesday, April 4th 2012, 9:11pm

That's another use for it^^ Probably the main purpose ;) But its also possible to bench the system with it

rewarder

oc-Alcoholic

Posts: 7,745

Location: Watt

Occupation: ocaholic

  • Send private message

4

Wednesday, April 4th 2012, 9:56pm

I was also thinking about using it for standard benchmarks. I think it could be really interesting to compare CUDA performance of GPUs in a completely different kind of GPU reviews in this case. There could for example be reviews for gaming and reviews for professional use.
Sys 1 Arbeitsviech: Gigabyte X58A-UD7, Core i7 920 @ 3.6GHz, ATI Radeon HD 6970, 6 GB Corsair Dominator DDR3 1600, OCZ Vertex 3 240 GByte SSD, 750 GB Samsung HDD
Sys 2 Linux Server: Gigabyte X48-DQ6, QX6850, Billige Graka, 4 GB DDR3, 1 x 36 GB WD Raptor, 6 x 2TB, 2 x 1.5 TB
MacBook Pro 13 Zoll, 8 GB RAM, Samsung PM 830 256 GB SSD

5

Thursday, April 5th 2012, 12:35pm

I think this type of "benchmark" could also be used to test the stability under full load. Because every CPU and every GPU runs at 100% load. @ My System, the power consumption jumped from ~125W/h to ~465W/h (Don't know for sure have to look again at home).

rewarder

oc-Alcoholic

Posts: 7,745

Location: Watt

Occupation: ocaholic

  • Send private message

6

Thursday, April 5th 2012, 5:09pm

I think this type of "benchmark" could also be used to test the stability under full load. Because every CPU and every GPU runs at 100% load. @ My System, the power consumption jumped from ~125W/h to ~465W/h (Don't know for sure have to look again at home).


For stability testing I use Furmark. The guys who program the software really make sure that even the last feature of the architecture is being used. So I don't thing one can give more load to a GPU than with Furmark. But still I think this kind of benchmark is interesting. Probably it would be helpful to not look at it from the hardware performance point of view. Instead it might be interesting to "monitor" the evolution of this kind of CUDA software.
Sys 1 Arbeitsviech: Gigabyte X58A-UD7, Core i7 920 @ 3.6GHz, ATI Radeon HD 6970, 6 GB Corsair Dominator DDR3 1600, OCZ Vertex 3 240 GByte SSD, 750 GB Samsung HDD
Sys 2 Linux Server: Gigabyte X48-DQ6, QX6850, Billige Graka, 4 GB DDR3, 1 x 36 GB WD Raptor, 6 x 2TB, 2 x 1.5 TB
MacBook Pro 13 Zoll, 8 GB RAM, Samsung PM 830 256 GB SSD