ERROR: cuMemAllocHost() 2

[Zilent]

Ok. Well in the meantime I would advise observing memory utilization while the benchmark is running, maybe you can pinpoint which algorithm is causing you to run out of RAM. Or if there's just a memory leak in general.

I will do some investigation next week. Is there any way to run with verbose output for the benchmark, because all I see is this, within a few seconds:

c:\cudaHashcat-1.37>cudaHashcat64.exe -b
cudaHashcat v1.37 starting in benchmark-mode...

Device #1: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #2: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #3: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #4: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU



ERROR: cuMemAllocHost() 2

[rico]

I guess you could benchmark one or two at a time instead of them all at once and observe memory usage:

Device #1: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #2: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #3: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #4: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU

So the command below will benchmark Device #1 & #2:

cudaHashcat64.exe -b --gpu-devices 1,2

Might help narrow it down.

EDIT: good idea to monitor power usage too.

[Zilent]

So, it works when running the benchmark with 1, 2 or 3 devices simultaniously, but not with all 4, that's when the error occurs.

c:\cudaHashcat-1.37>cudaHashcat64.exe -b --gpu-devices 2,3,4
cudaHashcat v1.37 starting in benchmark-mode...

Device #2: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #3: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU
Device #4: GeForce GTX TITAN X, 12288MB, 1076Mhz, 24MCU

Hashtype: MD4
Workload: 1024 loops, 256 accel

Speed.GPU.#1.: 31125.0 MH/s
Speed.GPU.#2.: 31173.3 MH/s
Speed.GPU.#3.: 31202.6 MH/s
Speed.GPU.#*.: 93500.9 MH/s

----
Additional info:
There's no where near peak in host memory consumption when running for all 4 devices. It might be max just above 25% of the 32 GB (31,9) available.

[Zilent]

Just want to let you know, that I just tried a benchmark again, just for fun of it, and now it works, after I don't know how many driver updates and cudaHashcat v2.01 on the box... I don't know what fixed it exactly.