Hyper-threading. Each core is sharing L1 and L2 cache between 2 threads. Some workloads will be faster without using it. If you do see a small performance increase you can turn off hyper-threading in the BIOS.
Don’t expect too much performance as those CPUs are 10 years old and are using DDR3 memory probably running at 1333 or even 1066 MHz.
I’ve no idea how quick that system should be, but it’s only 10 cores per CPU and ddr3.
I have a Dell T7910 with dual E5-2699v3 18 core CPUs and 512GB of DDR4 running at 2133Mhz, it can create a plot in 12 Minutes using all 72:threads, but including disk transfer then a series of plots averages at 18 minutes each. This is using Linux Mint.
I was like… there is the problem right there. Sorry to say man but Westmere processors are for sure not what you want to plot with. For reference I hit 1600 seconds on Dual 2667v2’s with just 256GB and madmax for plotting.
I suppose it’s more trouble trying to recover the $200 each they could fetch on ebay - easier to write them off and free up the space. It looks like it’d be a total playground for computer nerds though 8).
This is not accurate. In fact it’s plain wrong. Dual CPU systems are ideal for Bladebit! The more threads you can throw at it, the faster it’ll go. A dual CPU Xeon setup is what you want to buy to run Bladebit. I have several of them and they’re amazing. Even the older cheaper ones!