←back to thread

752 points dceddia | 1 comments | | HN request time: 0s | source
Show context
Aloha ◴[] No.36447476[source]
Well of course it does.

WinNT 3.51 was released in 1995 - the fastest PC in 1995 was either a Pentium or Pentium Pro at ~100 MHz - in 2000 a 600 MHz machine is likely a Coppermine PIII.

A fairly common amount of RAM in 1995 to Run WinNT would have been around 32 megs of ram, 64 megs would be especially generous. 128 megs is a high end workstation amount of memory.

The ATA interface also doubled in performance between 1995 and 2000.

There were significant security and stability improvements between NT 3.51 and Windows 2000 - particularly with changes to the driver model that increased stability. (even more so between 2000 and Windows 10/11)

replies(5): >>36447527 #>>36447543 #>>36447578 #>>36449512 #>>36449522 #
AnIdiotOnTheNet ◴[] No.36449512[source]
> Well of course it does.

Why does it follow that software designed for modern hardware, running on modern hardware, should be slower than software designed for older hardware running on slightly newer hardware?

replies(1): >>36456235 #
simooooo ◴[] No.36456235[source]
The software is running on an entirely different platform with different priorities. Security being the main one.
replies(1): >>36458175 #
1. AnIdiotOnTheNet ◴[] No.36458175[source]
The only significant difference in priorities as regards speed would be the modern prioritization of "developer time", ads, and telemetry.

I can run Windows 95 applications at better than era-appropriate speed, in an x86 emulator written in javascript running on a web browser. That's at least 3 layers of virtual machine abstraction and the applications are still faster.

So if you're saying "the comparison isn't fair because modern software is too shit to hold up", then I agree, but if you're trying to tell me there is something else inherent to modern computing that makes software so many orders of magnitude slower, than I request that you show data to support that claim.