NM. Not contributing to thread. 
If you’re sufficiently motivated, you can download http://download.sysinternals.com/files/TestLimit.zip and see what it’s like with half the memory gone.
It’s not the same as only having one stick, but it is probably about the same as only having two 4G sticks.
+1 to Everything! Been using it for years and love it!
I’ll add Teracopy to the must have list.
I have 4 GB on a laptop running “Windows 10 Pro.” It’s often quite sluggish and it’s gotten worse recently. (Is this because Microsoft keeps forcing mandatory bloatware additions down my throat?)
I’d love to remove disable the unneeded bloatware but what I read at blackviper.com looks more complicated than my comfort zone.
Although Task Manager shows 78% of my 4 GB used as I type, adding the memory column sums to only about 2 GB — 1.2 GB of that is used by “System.”
An unrelated issue: If I run a compute-bound task (almost zero I/O and zero memory) under cygwin on an otherwise almost-idle machine it gets only about 9% cpu usage (shell’s ‘time’ shows 99+%). What gives with that?
10 core CPU ?
Or the multi-core multi-thread cygwin bug. Try updating cygwin, or set the core-affinity to use just one core.
I just ran an experiment with bizarre results. I wrote a trivial program that spins through hundreds of billions of arithmetic steps, then prints a single integer result and exits.
On a 2-core machine which is otherwise fairly idle, I run 1, 2 or 3 copies of this program, each from its own cygwin xterm.
- When I run one copy of the program it took 130 seconds to complete. Task Manager reported it using about 8.7% CPU.
- With two copies of the program running concurrently, they took 150 seconds. The report was 8.5% each, drifting occasionally and briefly up to 18% or so.
- With three copies, they completed in 58 seconds! Each held steady at about 31% as reported by Task Manager.
I’m so ignorant these days, I hardly believe myself that I was a high-paid consultant once upon a time, sometimes specializing in identifying performance issues. But running a useless compute-bound process in the background to improve the performance of another compute-bound task? That’s a new one for me!
Intel® Core™ i7-5500U Processor – 2 cores.
The application code uses no thread management; it is single-thread.
I know nothing about the multi-core multi-thread cygwin bug, nor how to set core-affinity but will Google the if you think it may help.
If your actual task only takes 150 seconds, then you save more time by not googling for the solution. and as you have demonstrated, you can set core affinity to 1 just by filling up the other core with something
I don’t do compute bound tasks, and I don’t use cygwin if I can help it. I just happened to remember that.
As a matter of interest, I assume that it’s the inverse of the recent spectre/whatever bugs. It’s wasting time reloading the context when it tries to use two cores. I have similar problems with virtualization.