Yesterday, 03:28 PM
Hey everyone, I’ve been a long-time reader of the hardware benchmarks here, but I finally have a specific project that prompted me to start a thread. I recently came into possession of some retired enterprise gear: a couple of HP ProLiant BL460C blades. These are older units, each running dual Hexa Core 2.4GHz Xeon processors. Now, I know the general consensus in the community is that GPUs are the undisputed kings for Hashcat, but I’m interested in seeing what these can do for specific algorithms where CPU-based cracking isn't completely irrelevant—perhaps for some of the more memory-hard or CPU-intensive KDFs.
The specific point I’m stuck on is the OpenCL implementation for these older Xeons. I’ve seen some conflicting info on whether the Intel CPU Runtime provides meaningful scaling when you're dealing with older architectures like this. In my limited testing, I’ve found that even though I have 12 physical cores (24 threads) available across the blade, the relatively low 2.4GHz clock speed seems to be a real bottleneck for anything beyond very basic wordlist attacks.
Personally, I just love the industrial feel of these blade servers. There’s something incredibly nostalgic about repurposing 10-year-old "top-tier" hardware, even if the fans sound like they’re trying to achieve liftoff when the workload kicks in. My little homelab setup is great for learning, but I’m wondering if I’m just fighting a losing battle against modern hardware efficiency.
Does anyone here have experience running Hashcat on older blade infrastructure, or perhaps tips on optimizing the Intel OpenCL runtime for these specific Xeon families to get the most out of the available threads? Or is the performance-to-power ratio so skewed that I’m better off just using them as a space heater for the winter?
The specific point I’m stuck on is the OpenCL implementation for these older Xeons. I’ve seen some conflicting info on whether the Intel CPU Runtime provides meaningful scaling when you're dealing with older architectures like this. In my limited testing, I’ve found that even though I have 12 physical cores (24 threads) available across the blade, the relatively low 2.4GHz clock speed seems to be a real bottleneck for anything beyond very basic wordlist attacks.
Personally, I just love the industrial feel of these blade servers. There’s something incredibly nostalgic about repurposing 10-year-old "top-tier" hardware, even if the fans sound like they’re trying to achieve liftoff when the workload kicks in. My little homelab setup is great for learning, but I’m wondering if I’m just fighting a losing battle against modern hardware efficiency.
Does anyone here have experience running Hashcat on older blade infrastructure, or perhaps tips on optimizing the Intel OpenCL runtime for these specific Xeon families to get the most out of the available threads? Or is the performance-to-power ratio so skewed that I’m better off just using them as a space heater for the winter?