Bear in mind...
... that "a Linux slice with an effective 105MHz of processing capacity and just under 1GB of memory" is not the whole story. The VMs aren't limited to these figures, these are the averages if each VM was running flat out at the same time. The machine is very good at directing the CPU and memory towards the VMs that need it at each split-second, so whilst one's CPU might be idle whilst it does some I/O, another might have multiple cores running at serveral GHz, crunching some numbers.
"...With all ten engines fired up, that came to 3,139 MIPS of aggregate performance after SMP overhead was taken off the top..."
You get 3.200 MIPS under software emulation using an old 8-socket x86 server for a fraction of the price. It is much cheaper and more flexible.
MIPS = Meaningless indicator of processor speed.
IBM won't license its operating systems zOS, zVM, and zVSE on Hercules. Customers that run apps on those OSs are locked into IBM mainframe hardware.
My point is that using software emulation (which is 5-10x slower than running native code) you get the same performance as this IBM mainframe, using an old 8-socket x86 server. Running native code would yield 5-10x higher performance for the x86 server. So, if you recompile your Mainframe application on the x86 server, you get 5-10x higher performance than this Mainframe. Which is in par with the largest IBM Mainframe.
What does this tell you about the Mainframe cpus? They are quite slow?
Thanks for this excellent article! I started in IT on an old (new at the time) IBM 370/145 with 64K (yes,K) of main memory (32K of which was Memorex bolt-on). I have worked on most models since then, up to the 9672, after which I moved on to other things. I have a soft spot for IBM Mainframes and their operating systems, and it's great to read how far they have progressed. More mainframe articles, please!