Yo dawg I heard you needed memory so I put a swapfile in yo RAM so you can use RAM to get more RAM.

Gigabyte I-RAM was quite interesting back then, as at that time (around 2005-2006) high speed SSD (and NVMe) was not yet a thing.

Don't know if the RAM drive OP mentioned can use registered memory, as these are relatively cheaper per GB compared to the unbuffered counterparts and are available in higher capacity.

But still, it won't be of too much use unless you're having issues with ImDisk (or other RAM disk utilities). There indeed are cases where ImDisk's RAM disk won't work as intended, but some can be worked around by making your default TEMP path a symlink to the RAM disk rather than pointing your TEMP path to the RAM disk directly.

Plasma wrote on 2022-06-17, 03:40:

Yo dawg I heard you needed memory so I put a swapfile in yo RAM so you can use RAM to get more RAM.

🤣😂🤣

Hey, some of us plebians have motherboard ram limits..

Seems like they have a new version, which uses ddr3 dimms.

https://www.youtube.com/watch?v=qmZx-qaYgl8

The DDR4 version has my interest. That would be easy to fill up with high-capacity/high-speed sticks.

1) Should have been that from the start. But gotta sell overpriced RAM boards to crazy Chia miners, right?
2) Too little too late.

you'll just have to wait for the pcie v4 version.

Jo22 wrote on 2022-06-15, 18:54:

It's hard to express that an Atari ST with TOS in EPROM reacts quicker than a modern Ryzen with 16 cores and an NVME SSD:
If you press a key on an Atari ST, the character is faster displayed on screen than on a modern high-end PC.

Is that literally true? I know nothing about the low level workings of integrated circuits, so maybe the reasons for the situation you describe are obvious, but can you explain it to me, please? It does seem impossible that a mid 1980s home computer could do *anything* faster than a modern half-decent computer could. Let alone update the TV screen quicker than a modern PC could update it's monitor image.

Kerr Avon wrote on 2022-12-08, 20:44:

Jo22 wrote on 2022-06-15, 18:54:

It's hard to express that an Atari ST with TOS in EPROM reacts quicker than a modern Ryzen with 16 cores and an NVME SSD:
If you press a key on an Atari ST, the character is faster displayed on screen than on a modern high-end PC.

Is that literally true? I know nothing about the low level workings of integrated circuits, so maybe the reasons for the situation you describe are obvious, but can you explain it to me, please? It does seem impossible that a mid 1980s home computer could do *anything* faster than a modern half-decent computer could. Let alone update the TV screen quicker than a modern PC could update it's monitor image.

It certainly can be. All you need is a simple interrupt service routine on the old computer to process keyboard input and feed it to the probably very tightly coded application (if there is a separate OS at all). On a modern PC, it has to go through the USB stack and various software layers to reach the point where it gets displayed. Dan Lu wrote an article on this topic.

eton975 wrote on 2022-12-09, 03:12:

It certainly can be. All you need is a simple interrupt service routine on the old computer to process keyboard input and feed it to the probably very tightly coded application (if there is a separate OS at all). On a modern PC, it has to go through the USB stack and various software layers to reach the point where it gets displayed. Dan Lu wrote an article on this topic.

PS/2 keyboards use interrupts, so it should also be handled by interrupt service routines. On the other hand, USB keyboards are opposite: The CPU needs to actively poll the USB stack to see if the USB keyboard has any commands. The USB polling rate is usually 125Hz (8ms) but some high end keyboards can do 1000Hz or more.

It's more about the complexity of the system software, that makes things feel less responsive despite having more processing power. It's not just about keyboard input -- this affects every aspect of the modern system. Old systems are much simpler. Simple, well-coded systems that take as few steps as possible to get the work done, can be considered really "real-time", and they will certainly react faster than modern, complex ones in some aspects.

Kerr Avon wrote on 2022-12-08, 20:44:

Jo22 wrote on 2022-06-15, 18:54:

It's hard to express that an Atari ST with TOS in EPROM reacts quicker than a modern Ryzen with 16 cores and an NVME SSD:
If you press a key on an Atari ST, the character is faster displayed on screen than on a modern high-end PC.

Is that literally true? I know nothing about the low level workings of integrated circuits, so maybe the reasons for the situation you describe are obvious, but can you explain it to me, please? It does seem impossible that a mid 1980s home computer could do *anything* faster than a modern half-decent computer could. Let alone update the TV screen quicker than a modern PC could update it's monitor image.

Yes. It has to do with response time (latency).
Modern PCs have a high throughput (MB/s, GB/s etc) but also many layers between applications and the hardware.

While, say, an Atari ST as such is much slower than a modern PC, applications and hardware are much closer to each other.
Many of the "drivers" are part of the firmware (BIOS) already.

For example, the Atari ST's operating system, TOS (CP/M 68000 and GEM), is located directly in memory (in the EPROMs), there's no need to load system files from HDD first (extra drivers for HDD, CD etc can be loaded from floppy/HDD, however).

A keypress generates an interrupt, is handled by the keyboard routine, the scan code is being recognized, then sent to application/drawn on screen.

A modern PC has much more intermediate steps.
A keypress is noticed by the controller chip in the USB keyboard; the USB controller itself is being polled by the PC in a certain interval; the scan code is transmitted via USB protocol, handled by the Windows HID driver, handled by Windows kernal, handled by Win32 API, is then sent into a queue, arrives at the application, is then handled by the composition manager and/or drawn to screen via the Windows graphics drivers.

^This is just a vague description, in reality it might be a bit different.

according to this they now have a pcie 4.0 x8 DDR4 version
https://www.youtube.com/watch?v=Ci2ElmyvOmo

There was a design for an ISA RAM disk with battery backup in one of the 80s/90s electronics magazines. The problem with making it though is it required a custom programmed PIC, which is unavailable and the code was on a BBS that's been offline since the mid 90s. I have it brainmarked as one of those "IF boredom > 99 AND Ambition > 99 THEN grok design, modernise and convert it to Atmel microcontroller.."