Todo list (improvement suggestions)

[labbo]

Pause and resume whould be good and tp load more that 4000 hashes, but this is already in the next release or?

[Bitweasil]

Pause and resume whould be good and tp load more that 4000 hashes, but this is already in the next release or?

More than 4000 hashes is supported in 0.7 - pause & resume not yet, though.

[mjc]

Hello,

Firstly thanks for the x86_64 Linux release running the application with the following parameters:

Code: Select all

./CUDA-Multiforcer-Linux-64-0.7.bin --hashtype=MD5 --min=8 --max=8 --charsetfile charsets/charsetfull --hashfile=test_hash_files/Hashes-MD5-Full.txt --threads 512 --blocks 512 -m 500

Gives me around "Step rate: 417.5M/s Search rate: 4175.7M/sec"

Although, I did note that;
* Only one core on my Intel Core i7 920 chip was being used
* Only one GPU out of the 4 in my GTX 295 Quad-SLI rig seemed to be used

I understand you are reconciling the Windows/Mac/Linux ports before you release source code, are you open to the idea of releasing it to a couple of "alpha" testers?

Regards,
mjc

[Bitweasil]

It only uses one GPU core, and the CPU core isn't being "used" as much as "busy waiting for the kernels to complete so it can respond as fast as possible."

I don't have multi-GPU or actual CPU usage support yet (it only uses the CPU for some basic housekeeping tasks and then busy waits).

I intend to fix this in the next version, which will be a total redesign of the entire concept. You'll be able to use every system in your house/office to do the work. Including, if someone else contributes the code, both systems with ATI and nVidia GPUs.

[mjc]

Sounds great, can't wait to see the code ... I have a few IBM PowerXCell 8i (http://en.wikipedia.org/wiki/Cell_(microprocessor)) chips available and I'm keen to contribute some code to a worthy cause.