hashcat
advanced password recovery

I am running oclHashCat 1.30 under Windows 7 on AMD R9 290 card, factory-installed with Alienware Aurora machine.
All works well, thanks to the new update to HashCat that supports AMD beta drivers
(I have 14.7, Aug 12, 2014, v.14.20.1004-140811a-174673E).

Problem is, machine reboots sometimes. Quietly, no logs, no crash dumps, just powers down and up.
Some cracking modes (Combination, for example) reboot machine almost immediately after start, some last longer (BruteForce without Markov lasts a night, BruteForce with Makrov reboots rather quickly).
Has anyone experienced this? I am wondering if this has to do with some Catalyst setting - I am running all defaults.
The file I an trying to crack in 1.5 million hashes, salted, vBulletin < v3.8.5.
Screen looks like this:
---
oclHashcat v1.30 starting in restore-mode...

Device #1: Hawaii, 3072MB, 1000Mhz, 44MCU

Hashes: 1523668 hashes; 1523668 unique digests, 1453164 unique salts
Bitmaps: 24 bits, 16777216 entries, 0x00ffffff mask, 67108864 bytes
Applicable Optimizers:
* Zero-Byte
* Precompute-Init
* Precompute-Merkle-Demgard
* Early-Skip
* Not-Iterated
* Brute-Force
Watchdog: Temperature abort trigger set to 90c
Watchdog: Temperature retain trigger set to 75c
Device #1: Kernel ./kernels/4098/m2610_a3.Hawaii_1526.3_1526.3 (VM).kernel (255900 bytes)
Device #1: Kernel ./kernels/4098/markov_le_v2.Hawaii_1526.3_1526.3 (VM).kernel (133692 bytes)
Device #1: Kernel ./kernels/4098/bzero.Hawaii_1526.3_1526.3 (VM).kernel (30484 bytes)


INFO: approaching final keyspace, workload adjusted

[s]tatus [p]ause [r]esume [b]ypass [q]uit =>

Session.Name...: NewTest
Status.........: Running
Input.Mode.....: Mask (?1?1?1?1?1) [5]
Hash.Target....: File (C:\Temp\BBS\uc_members_HashesOnly.raw)
Hash.Type......: vBulletin < v3.8.5
Time.Started...: Fri Sep 12 11:12:52 2014 (2 secs)
Time.Estimated.: Fri Sep 12 11:32:17 2014 (19 mins, 16 secs)
Speed.GPU.#1...: 9 H/s
Recovered......: 2920/1523668 (0.19%) Digests, 2666/1453164 (0.18%) Salts
Progress.......: 87854113394928/87867270180864 (99.99%)
Skipped........: 45360/87854113394928 (0.00%)
Rejected.......: 0/87854113394928 (0.00%)
HWMon.GPU.#1...: 0% Util, 65c Temp, 44% Fan

Do you have a big enough power supply?

yep, definitely sounds like a power issue.

The machine has 875W Power Supply, I am not sure how much the card can take at peak loads - I saw numbers from 250W to 500W.
Connected to UPS SmartUPS 1000, not showing more then 15% load on the UPS.

does your motherboard have supplemental power supplied to the pci-e bus?

the only other time i've encountered what you are describing is when the 24-pin ATX connector melted due to too much power being pulled through it because of a lack of supplemental power.

transout, I might be wrong but the memory shown by Hashcat is only 3072MB, although all 290X have 4GB+. Maybe that's your source of probs.

It's a general AMD OpenCL runtime limitation

Thank you for pointing out power issues.
Looks like it's a common issue with Aurora machine:
http://en.community.dell.com/owners-club...t/19590337
Out of the box, the card is wired to one power supply cable, split into 2 connectors (6-pin and 6+2pin).
Power supply has a 2 GPU cables, so rewiring one of the connectors to feed from a separate cable seem to fix the problem. It's running now in the combination cracking mode, which I couldn't use before.
Hope it doesn't fry the card due to improper grounding.