Building the lowest latency computer...

P1x3L

Weaksauce
Joined
Jul 31, 2012
Messages
81
Please checkout this amazing article, which is unfortunately a rarity today: https://danluu.com/input-lag/

As you can see the Apple 2e had much lower keyboard to screen latency than any modern desktop has today. I'm interested in building a modern computer that is as low latency as possible, with the highest performance possible as a secondary attribute. More work needs to be done on where the latency is coming from, but this article is a great start. The top 3 areas latency is coming from are: 1) keyboard 2) monitor 3) GPU - it seems. In terms of solvability 1) seems very solvable 2) seems somewhat solvable 3) seems difficult to solve... And to get it down to 30ms, I think one would have to find the other factors increasing latency before we can determine if they're solvable with today's parts.

Thanks - and I hope some of you find this as interesting as me!
 
According to the testing from this video, looks like with the right setup you can almost get a sub 15ms response out of a modern setup.
 
According to the testing from this video, looks like with the right setup you can almost get a sub 15ms response out of a modern setup.


It is clear to me that you did not read the article. No it isn't 15ms keyboard to screen.
 
Software wise, it looks like linux does a bit better than windows, but there are probably tricks to help out windows.

Hardware wise, I would like to see the differences in GPUs. Curious on how the APUs compare to Intel's iGPUs and how different GPUs compare with various drivers and APIs.
 
In his article he doesn't even say what "modern keyboards" or monitors he used. He just kind of assumes they're all the same at the same time saying keyboards have changed. Like he knows they make a difference yet didn't care because it was inconvent for his article.

His tests are all over the place measuring different things. There are so many details lacking I wouldn't even be surprised if the results are just made up.


What the article was showing is that modern applications have more input lag than primitive text only applications.

I bet the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2 if you actually measure something modern built for low latency.
 
Oh god please don't let this grow into an audiophile-esque obsession with people willing to pay hundreds of dollars for useless accessories.

Super Low Latency mouse and keyboard cables! Only $129.99/ft.
Super Duper Mega Magnet - just position it near your mouse to decrease input lag by up to 4ms, guaranteed*! $47.99 each, minimum order of 2 because why not, sucker.


*Not guaranteed. All sales are final.
 
Software wise, it looks like linux does a bit better than windows, but there are probably tricks to help out windows.

Hardware wise, I would like to see the differences in GPUs. Curious on how the APUs compare to Intel's iGPUs and how different GPUs compare with various drivers and APIs.

I'm leaving software out of it for now since it is more flexible. In Linux there are lots of options for lower latency at the cost of other things, for example. Any idea how to benchmark GPUs fairly? I guess the latency benchmarks of computer components (except peripherals like the screen) can all be done in software accurately?
 
In his article he doesn't even say what "modern keyboards" or monitors he used. He just kind of assumes they're all the same at the same time saying keyboards have changed. Like he knows they make a difference yet didn't care because it was inconvent for his article.

His tests are all over the place measuring different things. There are so many details lacking I wouldn't even be surprised if the results are just made up.


What the article was showing is that modern applications have more input lag than primitive text only applications.

I bet the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2 if you actually measure something modern built for low latency.

So much of this is not true. He talks about gaming monitors. His benchmarks are better in many ways because he is actually comparing the full latency from input device to change on screen. You're just degrading his work without adding anything. You have no evidence for your bet... Or even rationale. Why would, as you said, "the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2."
 
Oh god please don't let this grow into an audiophile-esque obsession with people willing to pay hundreds of dollars for useless accessories.

Super Low Latency mouse and keyboard cables! Only $129.99/ft.
Super Duper Mega Magnet - just position it near your mouse to decrease input lag by up to 4ms, guaranteed*! $47.99 each, minimum order of 2 because why not, sucker.


*Not guaranteed. All sales are final.

Silly. You're making a straw man here. There's plenty of benchmarking to see latency from lower latency devices is real and not akin to audiophile perceptions.
 
So much of this is not true. He talks about gaming monitors. His benchmarks are better in many ways because he is actually comparing the full latency from input device to change on screen. You're just degrading his work without adding anything. You have no evidence for your bet... Or even rationale. Why would, as you said, "the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2."

No, because one "gaming monitor" can have 50 ms of lag and another can have 8 ms of lag. He doesn't say which "gaming monitor" he uses. Same for keyboard and mice. And he doesn't even say how he tests or anything.
 
P1x3L, are you Dan Luu? Or at least close to him?

Internet gamma males... If you are the only supporter of a view they can find, then you must be the originator of the view or a friend of them... Absolutely not; this is self evidently important to those of us who can discern between useless and useful benchmarks.
 
No, because one "gaming monitor" can have 50 ms of lag and another can have 8 ms of lag. He doesn't say which "gaming monitor" he uses. Same for keyboard and mice. And he doesn't even say how he tests or anything.

I didn't say you were not welcome to assume the worst of him and his article. Sure... However, your primary claims have nothing to stand on:
"I bet the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2."
 
No I just think you are being way to defensive of the article. The speed of electricity hasn't changed over the last few decades and when those electrons have a lot less to do, naturally there will be less delay.
 
I didn't say you were not welcome to assume the worst of him and his article. Sure... However, your primary claims have nothing to stand on:
"I bet the lowest latency keyboard, mouse, and monitor on a modern computer have less input lag than an apple 2."

https://www.blurbusters.com/gsync/gsync101-input-lag-tests-and-settings/3/

A bunch of those input lag numbers are lower than 30 ms which was his Apple 2 latency measurement. And these are actual tests that show the exact methodology so anyone could easily reproduce them. And these are with 3D games displaying graphics instead of a super simple terminal like he tested.

All of his tests were done with terminals. So yeah if his tests are legit, terminals on new PCs do have higher latency. But that shows absolutely nothing about games. And if you want a lower lag terminal on a modern PC I'm sure it could easily be done and may already exist because he didn't specify exactly which terminals he tested. But latency for terminals isn't something the makers seem to care about which is the entire point of his article.
 
Hasn't this been posted before? Pretty sure I responded because I found that the "modern" keyboard tested was something like a standard Dell USB with 3-key interrupt and at best 110 Hz polling. The range of keyboards available makes this whole article moot. Seems like a guy who is just stuck in the past.
 
Back
Top