Odd behavior getting GTS 450 to run alongside GTX 960, and 750TI.

[saltedsalt]

Hello,

I have hashcat installed on an Ubuntu 18.04 box that has a GTX 960, GTX 750TI, and a GTS 450 in it.

I'm able to run the benchmark with each graphics card separately, as well as with the GTX 960 and 750TI together. However, when I attempt to run the benchmark with all three graphics cards simultaneously I'm getting these errors:

Code:

hashcat (v5.1.0) starting in benchmark mode...

Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.

* Device #1: WARNING! Kernel exec timeout is not disabled.
            This may cause "CL_OUT_OF_RESOURCES" or related errors.
            To disable the timeout, see: https://hashcat.net/q/timeoutpatch
* Device #2: This hardware has outdated CUDA compute capability (2.1).
            For modern OpenCL performance, upgrade to hardware that supports
            CUDA compute capability version 5.0 (Maxwell) or higher.
nvmlDeviceGetCurrPcieLinkWidth(): Not Supported

nvmlDeviceGetClockInfo(): Not Supported

nvmlDeviceGetClockInfo(): Not Supported

nvmlDeviceGetTemperatureThreshold(): Not Supported

nvmlDeviceGetTemperatureThreshold(): Not Supported

nvmlDeviceGetUtilizationRates(): Not Supported

OpenCL Platform #1: NVIDIA Corporation
======================================
* Device #1: GeForce GTX 960, 499/1999 MB allocatable, 8MCU
* Device #2: GeForce GTS 450, 241/964 MB allocatable, 4MCU
* Device #3: GeForce GTX 750 Ti, 500/2002 MB allocatable, 5MCU

Benchmark relevant options:
===========================
* --optimized-kernel-enable

Hashmode: 0 - MD5

clCreateContext(): CL_INVALID_DEVICE

Started: Mon Jun 10 18:17:06 2019
Stopped: Mon Jun 10 18:17:09 2019

I understand that the initial errors are saying that the GTS 450 has outdated CUDA compatibility, but given that it is still able to run solo, I'm not sure why it wouldn't be able to run with the other GPUs.

Here's what it looks like when the GTS 450 runs by itself:

Code:

hashcat (v5.1.0) starting in benchmark mode...

Benchmarking uses hand-optimized kernel code by default.
You can use it in your cracking session by setting the -O option.
Note: Using optimized kernel code limits the maximum supported password length.
To disable the optimized kernel code in benchmark mode, use the -w option.

* Device #2: This hardware has outdated CUDA compute capability (2.1).
             For modern OpenCL performance, upgrade to hardware that supports
             CUDA compute capability version 5.0 (Maxwell) or higher.
nvmlDeviceGetCurrPcieLinkWidth(): Not Supported

nvmlDeviceGetClockInfo(): Not Supported

nvmlDeviceGetClockInfo(): Not Supported

nvmlDeviceGetTemperatureThreshold(): Not Supported

nvmlDeviceGetTemperatureThreshold(): Not Supported

nvmlDeviceGetUtilizationRates(): Not Supported

OpenCL Platform #1: NVIDIA Corporation
======================================
* Device #1: GeForce GTX 960, skipped.
* Device #2: GeForce GTS 450, 241/964 MB allocatable, 4MCU
* Device #3: GeForce GTX 750 Ti, skipped.

Benchmark relevant options:
===========================
* --opencl-devices=2
* --optimized-kernel-enable

Hashmode: 0 - MD5

Speed.#2.........:   840.2 MH/s (59.15ms) @ Accel:128 Loops:128 Thr:768 Vec:2

Hashmode: 100 - SHA1

Speed.#2.........:   223.5 MH/s (74.40ms) @ Accel:128 Loops:64 Thr:512 Vec:4

<.. and so on>

Any insight into this issue would be very much appreciated, thank you.

[Chick3nman]

> Any insight into this issue would be very much appreciated, thank you.

* Device #2: This hardware has outdated CUDA compute capability (2.1).


Need any further insight? The card is wildly outdated and not supported by current drivers. The fact that you got it into hashcat at all along side newer cards is surprising.

[saltedsalt]

> Any insight into this issue would be very much appreciated, thank you.

* Device #2: This hardware has outdated CUDA compute capability (2.1).


Need any further insight? The card is wildly outdated and not supported by current drivers. The fact that you got it into hashcat at all along side newer cards is surprising.

I'm not sure that I agree. Just because it's an older card doesn't mean that it's worthless.

Furthermore, it actually still is supported by the Nvidia Linux drivers. So, considering that it is:
a) recognized by the operating system
b) recognized by hashcat whilst running independently
c) still supported by Nvidia

getting it working with the other cards doesn't seem infeasible.

Thanks for your input though.

[Flomac]

I'm not sure that I agree. Just because it's an older card doesn't mean that it's worthless.

It's not supported by hashcat anymore. That's pretty much what the error message is telling you. Only hardware from Maxwell/CUDA 5.0 up is supported. The card is 10 years old and compared to the performance capabilities of the other two cards it doesn't matter if this card is "helping" cracking passwords or if you just spare it out and save the ~100W it consumes. You could consider that worthless.