Retroid

So, for the curious, XP to XP file transfer performance over the same physical conditions as the above test of 98SE to XP file transfer performance showed a significant (>50%) boost in performance. That test clocked in at 948 Kbytes/sec (7.4 mbps). That got me closer to the USB 1.1 theoretical limit, but I have a feeling that WiFi latencies are what are holding me back. If anyone actually still cares and responds to this post, I will perform the same two tests (XP to XP and 98SE to XP transfers) strictly over the wired LAN to see what improvements might be had there. So, since it seems that the 98SE IP stack is kind of wussy out of the box, would anybody care to point me to a thread about optimizing TCP/IP performance in 98SE? I've long since thrown out all those wonderful articles I used to reference regarding tweaking the MTU settings in the Windows 9x family. Blessings, all!

So not sure if you made any progress here. Problem is when you inject "WiFi" into the picture, so many variables come into play that it's really hard to give good advice, other than generically to ensure that you don't have lots of competition on your chosen WiFi channel either from neighboring devices or your own devices. Probably best to hook your Windows 7 machine into the LAN and do some tests there, to see if it's Windows 7 to 98SE as the problem or if it's the WiFi that's the problem. Just as an FYI, in a quick home test I was able to pull from a Windows XP VM running over WiFi a 150MB file from my Windows 98SE share of a USB stick folder (USB 1.1) at 606 Kbytes/sec. The Wireless G hotspot should support theoretical limit of 54 Mbps, which is higher than the limits of USB 1.1 (12 Mbps). However, 606 Kbytes/sec is only 4.7 Mbps, which is well under both of those limits, so I am guessing other performance limiters are the 500MHz Pentium III CPU speed, and probably crummier than average TCP/IP implementation in 98SE. Just out of curiosity, I think I will boot this 98SE laptop into XP and run the same test, just to see what kind of performance difference results.

Wow. I do have to admit that I am very impressed with what everyone is able to accomplish. @MrMateczko: I had given up on trying to run 98SE on modern hardware a long time ago. I guess I never could justify purchasing new equipment for that purpose, but given that I had kept around so many old pieces of equipment for a long time, including some PCI graphics cards, I suppose I probably could have done it if I had been targeting it. Unfortunately, so much of that stuff has made its way to the landfill since then. HOWEVER, since I do make regular trips to the landfill and often see old 98/XP era machines laying around there, perhaps I could rummage for some precious items in your spec list. @DougB: Thanks for giving me the faith that this actually was a worthwhile endeavor. I am actually typing this post now on the 128MB Pentium III laptop in Pale Moon 3.6.32, and it is performing quite decently here on MSFN, which had been crushing poor Opera 12.02 for me yesterday evening. I have to imagine that you are an x86 assembly developer, if you are booting to DOS 7.1 on a regular basis! If I were to be using this laptop on a regular basis, it would likely be to be reading the NIV84 on the one remaining site on the Internet that I have found that contains it, as well as using Acrobat Reader 6 to pull up PDFs for my Bible Study. @LoneCrusader: I am very similar in liking to prove that computing tasks can continue to be done on machines that have long since been written off by others. Part of it is my built-in annoyance that software has become so ridiculously bloated these days, because developers frankly just don't need to be optimal anymore. Sloppiness can still unfortunately achieve acceptable results with today's ridiculous amount of horsepower most people have at their disposal. Like you, another part of my passion is just simply the technical challenge of proving it can be done. And with everyone else's help here on MSFN, you have helped me get a LOT done with a meager 128MB of memory. Just imagine how much more virtual computing could be accomplished if we simply made the effort to keep our VMs operating with minimal resources consumed. Before DougB showed me the Pale Moon 3.6.32 light, i was experimenting with Tiny Core Linux and was actually starting to make progress running even more modern browsers on this same hardware. But so far here on MSFN, I have not gotten anything to perform better than 98SE with PM 3! @rloew: Once I get strapped into a particular workstation, I try to stay there as long as possible to prove that I can keep on functioning without pulling the parachute. One of my "bail out" (or "bail sideways") tricks is to pull up Microsoft Remote Desktop and start controlling other modern machines on my network. Of course, I have to admit that I really am a Mac guy, so I probably need to figure out what is the most modern VNC software available for 98SE and cross my fingers that I can get through the security protocol modernization that Apple has been pushing out over the past few years. @all: Thanks a lot for keeping all this rolling. I believe there is still a future for squeezing the most life out of the fewest system resources possible. I have a Xen Hypervisor running on one of the stray machines on my network, and am wondering if any of these tricks we've been coming up with might have some use in that space. Thanks for the quick responses to my posts, gentlemen. I am praying many blessings on you all!

So as many of you probably already know, this Javascript engine in Opera 12.02 has abysmal performance. In XP, I had installed Opera 20 (just had it laying around on my NAS) to compare with version 12.02, and the performance in the Chromium Javscript engine is vastly superior, even on this 500MHz Pentium III machine. I wonder if a Chromium-like optimized Javascript engine was ever built for the Win 9x platform. Will do some additional research after I finally shut off this poor slug :-)

Hi, y'all. Drawn here because I couldn't bear to toss out this circa 2000 Dell laptop. And if you could just hear the Ensoniq General MIDI performance, you would know that it's worth keeping around :-) Looking forward to getting to know y'all more over the coming days!

Just out of curiosity, are you guys using 98SE for functional work, or just toying around to see what's still possible? @dencorso You are making me feel old. It really doesn't feel like it's been that long. I even remembered to run "msinfo32" to get a machine report, although I forgot how to find how many MHz my CPU is actually trying to clock at on this machine. Might need third-party software for that. @MrMateczko Thank you for the kind words. You encouraged me to go check out Opera again, as my first attempt turned out to be a dismal failure. Turns out I had downloaded 12.12, not 12.02. Oopsie. I feel like somebody just turned the lights back on so I can see the msfn site again. However, the Javascript execution here in Opera is really crushing my poor little Pentium III's performance. I do also have XP dual booted on this old fella, so I am going to check and see if this is a win98 specific thing or if it's just a sign that a nearly 20 year old machine is going to have inherent limitations browsing the modern web.

Why do I feel the need to submit my test results from https://www.ssllabs.com/ssltest/viewMyClient.html to demonstrate that roytam1's Retrozilla actually supports TLS 1.2? Why am I struggling to create a JPEG file to give evidence of this fact on this Dell Pentium III Laptop with a scant 128MB? Why do I sort of want to hug the gentleman who previously stated I could make those security errors go away by rolling back to KernelEx 2016.16? And, most importantly...how is it that I am able to submit this post near the end of 2018 on an operating system that hasn't been supported for so long, I'm not even sure when Microsoft stopped supporting it? You guys are crazy. And you're sucking me right in with you. Help!