Hacker News Reader

Comment by Someone | Hacker News Reader

SoftRAM was a fraud (https://en.wikipedia.org/wiki/SoftRAM ), but RAM Doubler (https://tidbits.com/1994/01/10/ram-doubler/ , https://www.amazon.com/Connectix-R010836-RAM-Doubler-9-0/dp/...; “Currently Unavailable”) worked, and modern OSes may still use the technique of, instead of swapping out memory pages, compressing them so that a page becomes available (https://en.wikipedia.org/wiki/Virtual_memory_compression)

View on HN

> A number of people have expressed disbelief that such a feat is possible, saying that they’d avoid anything like RAM Doubler because it’s obviously doing strange things to memory, which isn’t safe. The answer to these naysayers is that a program like RAM Doubler either works or it doesn’t – it’s a binary decision.

Cool that the technology was available and weird that the knowledge behind why it would work was not more widespread.

I take issue with the statement here though - it's not binary. It's of course important if it works in normal situations, but how does it work under extreme memory pressure or certain incompressible memory pages? Et.c.

MacOS definitely uses memory compression, it's probably the only thing making an 8gb macbook viable for anything more than pure safari and Zoom.

https://www.lifewire.com/understanding-compressed-memory-os-...

So does Fedora's latest release. zram0 partition. zipped RAM drive. And measuring speed of my /tmp directory, have a feeling it resides on RAM.

It's funny how memory works. Human memory that is.

When I read the title of this article and thread, I thought, "This sounds a bit like that fraudulent DOS memory doubler. What was it called again, RAM Doubler?"

Oops. Thanks for setting the record straight!

Reply on news.ycombinator.com