Is your older CPU still relevant?

[Terry Rock]

I ask this question because I am on an older I7-3820 and have had zero issues with any game to date.
You'd have to go all the way back to the first quarter of 2012, that when the 3820 was released and yet, it still runs every sim currently released...quickly...and in VR.

 

[Case_]

I take it as a response to what I was saying, so if it wasn't meant to be, sorry for assuming it was...

In sim racing, you want as many frames per second as you can get, since this will allow your gpu to render and display a frame as close as it can to the current point in time. If you are limiting the framerate, you will always be slightly behind in this regard, as your frame time will be much higher. And when a game relies heavily on quick, decisive inputs, having as many fps displayed as your gpu can achieve is actually quite beneficial.

No, I really don't want as many frames per second as I can get, because there's the point of diminishing returns.

If I limit my framerate to ~60 fps, what I see is never more than 16 ms behind. Given that the average human reaction time is around 200 ms, it's pretty much a meaningless delay. I do run with Vsync off, because turning it on introduces a lot higher lag that *is* noticeable, but the 16 ms of delay with framerate limited at 60 fps is an absolute non-issue. Especially when there are other crucial delays you really can do nothing about, like when it comes to your input devices, FFB and all that jazz.

Would I run at 120 fps if I could run the games locked at that framerate with no dips at the level of visuals I expect? Yeah, probably. But I would be aware of the fact that it's really just to make me feel better, a placebo, really. Because for 95 % of simracers, those extra 8 ms gained would make exactly zero difference (apart from raising our PC power consumption).

And I definitely don't want to leave my framerate unlimited, because anything other than very close to an integer multiple of the monitor's refresh rate looks like crap to me, especially if the framerate is fluctuating (which it almost always is if you're not limiting it).

I know from experience that a gtx 1070 paired with an i5 4670k @4.4 GHz is heavily bottlenecked in cpu heavy games.

Well, obviously it will be bottlenecked in CPU heavy games. Isn't that absolutely obvious? The real question is - which of the current sims do you consider to be CPU heavy to the point of being bottlenecked on such CPU?

So to say that a 1080 with a lesser processor wouldnt be, is denying reality.

But I didn't say that, did I?

Will a 1080 be utilized fully with an Ivy Bridge? Probably not, and to what extent, that will depend on a specific game (and on the specific CPU).

 

[natedogg1867]

I take it as a response to what I was saying, so if it wasn't meant to be, sorry for assuming it was...

haha i was more just speaking generally, but quoted your comment because it related to a small degree.

About the high framerate... I suppose that goes back to the endless debate many people have on whether the lag from Vsync, or similarly low framerate is perceptible compared to a non-limited framerate or very high frame rates.

I will say that when switching from 60Hz to 144Hz, the difference in going from a Vsync'd 60fps to unlimited is night and day. The overall smoothness and thus more up to date rendered frames really matters to me, since I can notice it quite easily. Even before when I had a 60Hz display, I still played uncapped because the screen tearing was never close to being bad enough to be disturbing, although I know many people disagree there.

I think we can probably disagree about the benefit of having the most up to date frames being rendered and displayed though. People simply wouldnt care about 144/165/240+ Hz displays if a framerate above 60 didnt matter.

Well, obviously it will be bottlenecked in CPU heavy games. Isn't that absolutely obvious? The real question is - which of the current sims do you consider to be CPU heavy to the point of being bottlenecked on such CPU?

Well, sims are traditionally classed as cpu heavy, since they require so many physics calculations. So that is why I referenced it there. Quite different than playing Crysis or Call of Duty where the emphasis is much more focused on the graphics than physics.

Specifically though, rF2 for being unoptimized, RaceRoom for being an ancient platform that does not play well with Pascal gpus (in my experience), and AC with lots of AI or players in MP. AMS is probably the lightest on my system oddly enough.

But I didn't say that, did I?

Yeah sorry, wasnt referring to you specifically there. Probably should have tried to clarify some of my thoughts to not imply it was a direct response.

I have seen people elsewhere talk about gpu/cpu combos like this and make statements that there will be no bottleneck with the cpu, and that isnt backed up by facts/benchmarks from everything I have read and experienced.

I completely agree though, that a beefy gpu with a slightly older cpu isnt going to go to waste, and will certainly serve anyone who has a combo like that pretty well. I think my main point is simply that it just wont be fully utilized, or utilized to its potential.

Cheers

 

[Case_]

haha i was more just speaking generally, but quoted your comment because it related to a small degree

Yeah, I was actually just about to edit the whole comment above out, because I decided it was not worth arguing about and, but then the notification about your response popped up...

I suppose that goes back to the endless debate many people have on whether the lag from Vsync, or similarly low framerate is perceptible

Is there such debate? I've done plenty of comparisons and tests to know for sure the lag from Vsync is definitely perceptible. Whether you care about it or not is another thing, but its existence can be proved fairly easily.

I will say that when switching from 60Hz to 144Hz, the difference in going from a Vsync'd 60fps to unlimited is night and day.

Well, if you have a 144 Hz monitor, then it obviously makes no sense to limit the framerate at a lower value unless you have a really good reason for that, and of course the difference between 60 fps and 144 fps would be *very* noticeable on such display. But it would be a visual difference only - the animation would appear a lot smoother, but there would be very little (if any) difference in your response time, which is what we were talking about.

People simply wouldnt care about 144/165/240+ Hz displays if a framerate above 60 didnt matter.

See above. Framerates above 60 certainly matter (though I'd still say going above 144 doesn't really matter anymore), but mostly just for the visual reasons, not because they significantly lower your response time.

Anyway, that's it from me, sorry for being annoying.

 

[natedogg1867]

Yeah, I was actually just about to edit the whole comment above out, because I decided it was not worth arguing about and, but then the notification about your response popped up...

Anyway, that's it from me, sorry for being annoying.

I think you would be one of the last people that would be considered annoying on this site. I always see you share good info, so I hope my comments above werent dismissive.

Is there such debate? I've done plenty of comparisons and tests to know for sure the lag from Vsync is definitely perceptible. Whether you care about it or not is another thing, but its existence can be proved fairly easily.

haha I suppose I should stop browsing reddit so much, as well as some other places that maintain that belief. Im not sure who the people actually are that hold that belief though. Whether they are just being contrarians to troll, or that phenomenon where you rationalize the setup you have... even if there is evidence that your thoughts are a bit mislead.

Well, if you have a 144 Hz monitor, then it obviously makes no sense to limit the framerate at a lower value unless you have a really good reason for that, and of course the difference between 60 fps and 144 fps would be *very* noticeable on such display. But it would be a visual difference only - the animation would appear a lot smoother, but there would be very little (if any) difference in your response time, which is what we were talking about.

I seem to have been conflating the effects of Vsync lag with the differences in frametimes from a locked framerate to a higher one.

After having played at higher framerates now for a number of years, there have been a number of times where I tried playing at the normal 60 fps again, and have always noticed a stark difference. Although, Im not sure how well I would be able to notice the differences in frame times based on my reaction times to what I see. I suppose this is more of a feeling than something that is provable with numbers.

In the end, I suppose I am just a proponent of playing at high framerates because the pros outweigh the cons for me. And the hardware choices matter in this regard if you want to get the most out of your setup.

 

[RasmusP]

Nice discussion guys!
I won't take part for now but only to the "you can't see/feel/whatever more fps" :
My second hobby is mixing and recording audio. Mostly I do some metal/rock instrumental covers just for myself but sometimes someone wants to record something.
If you want to have your voice (voice is the most noticeable for this) processed by the PC and back on your headphones you will have a latency shown to you by the mixing program (I use cubase).
You feel the difference between 2.5ms in-out and 5ms in-out and you certainly hear the difference between 5ms and 10ms.
Above 10ms it not only feels weird but you really start to hear the lag between what you speak/sing and what you hear.
I'm not sure if the human eye has slower receptors that the human ear but afaik the eye sensors are some "Constant sending" things so every refresh of the frame will be noticed at least by some of the nerves at every point up to the level where all nerves are at the maximum "I can send something again to the brain" - pace.
I think this limit is around 200hz for my eyes (tested with stroboskope until it was just a light without flickering from all angles).

That said: yes, I was talking about 60hz/fps in my posts. Most people want to slap everything on ultra max so a 2600k still runs fine with a 1080ti.
Sure, if you aim for max fps with lower settings this whole discussion shifts to new perspectives!

 

[Andy_J]

Well I always have a plan in my PC building and upgrading time lines. The reasons I have paired a new GTX 1080 to my old i5-5670k (OC to 4.5) is two fold, one is that I wanted to see the performance increase I might get and two, the GTX 1080 will be going in to a new mega VR build next year so was purchased with that in mind.
This is the one I purchased: https://www.gigabyte.com/Graphics-Card/GV-N1080AORUS-X11-8GD#kf

But I love the increase in frame rate, the way I can now max to ultra every single sim so far on my system (and that includes my flight sims and space sims) using a 2560x1080 34" monitor. And that has surprised me greatly. So much so that the build might be later rather than sooner.

ps, great discussion by the way. Oh and I rarely use v-sync except on some games that actually need it to perform satisfactorily.

 

[Case_]

You feel the difference between 2.5ms in-out and 5ms in-out and you certainly hear the difference between 5ms and 10ms.
Above 10ms it not only feels weird but you really start to hear the lag between what you speak/sing and what you hear.

As someone who is also guilty of doing some recording and mixing occassionally, even given what I said above, I have to confirm this is indeed true. This is in fact one of the reasons I currently have two soundcards in my PC, because even though I wouldn't really need them, the soundcard I would generally use has some processing delays hardcoded in its Windows drivers (it's not there in Linux, so it's obviously not a hardware issue) when it comes to audio input monitoring, which makes playing guitar incredibly awkward and often downright unusable - especially playing fast solos and stuff like that, it can really mess with your timing.

That being said, I still believe this to be a somewhat different issue than what we're talking about here, to be honest.

But...yeah

 

[Insert Coin]

One more thing to add: some racing games use the FPS as their 'physics calculation clock' - at every frame the physics are calculated. Old games like Race 07 and GTR2 do this, R3E did this (maybe still does, don't know). If your FPS dropped well below 60 FPS some strange physics errors started to occur. When R3E was in early beta it performed horribly, so the FPS was often very low and the cars started to behave strangely. At the same track, same situation, with 60 FPS or higher (with low graphics settings) the problem was gone. I think most games have a separate fixed physics calculation clock now, independant of the FPS. Correct me if I'm wrong.

 

[Case_]

I belive iRacing still locks its physics engine at 60 Hz, so framerates above 60 are definitely just visual enhancement only. But I sure hope they're the odd one out.

 

[RasmusP]

@Martin Fiala
To the comparison: it surely is different. It was just to all the "the human body doesn't see/feel/hear/recognise at all anything faster than XX.xx ms".
It's a big difference between being able to differ between 60 and 120 fps or hear the difference between a boost of 110hz or 120 Hz in an audio track but that doesn't mean that your body doesn't recognise the difference!
I honestly have to say I won't buy a 144 Hz gsync monitor just because I won't ever be able enjoy gaming on my TV again

 

[Furnace Inferno]

Nice discussion guys!
I won't take part for now but only to the "you can't see/feel/whatever more fps" :
My second hobby is mixing and recording audio. Mostly I do some metal/rock instrumental covers just for myself but sometimes someone wants to record something.
If you want to have your voice (voice is the most noticeable for this) processed by the PC and back on your headphones you will have a latency shown to you by the mixing program (I use cubase).
You feel the difference between 2.5ms in-out and 5ms in-out and you certainly hear the difference between 5ms and 10ms.
Above 10ms it not only feels weird but you really start to hear the lag between what you speak/sing and what you hear.
I'm not sure if the human eye has slower receptors that the human ear but afaik the eye sensors are some "Constant sending" things so every refresh of the frame will be noticed at least by some of the nerves at every point up to the level where all nerves are at the maximum "I can send something again to the brain" - pace.
I think this limit is around 200hz for my eyes (tested with stroboskope until it was just a light without flickering from all angles).

That said: yes, I was talking about 60hz/fps in my posts. Most people want to slap everything on ultra max so a 2600k still runs fine with a 1080ti.
Sure, if you aim for max fps with lower settings this whole discussion shifts to new perspectives!

Our audio and visual systems have very different reaction times, that's why they have an audible beep in F1 for shifting rather than just rev lights.

 

[Full_Ninja]

In addition to Intel, looks to have the Marvell 9128 with ASMedia1061 for eSATA controllers.
Check to make certain the boot and game drives are connected to the faster Intel 6 GB headers.
You may also want to download and install the PCI-e registry hack.
It allows PCI-e 3.0 to work on the board. (You will have to re-enable it after each video driver update.)
Most X79 boards were listed as PCI-e 3 'ready'' but it was not actively deployed at delivery.
It's been stated by some game reviewers, that they saw little difference versus PCI-e 2.
I beg to differ.
For most games it won't.
In RF2 and a few other racing titles, we discovered an almost 30 percent framerate increase...more importantly there was a massive improvement in smoothness.

I checked my PCI-e and sure enough it was 2.0. I ran the registry hack and now it's 3.0. Thank you so much for pointing this out.I was surprised to learn that nvidia disables 3.0 on x79 by default even if you had it enabled in the motherboard bios. Thanks again and have a good new year

 

[Eddie Ellis]

As someone who is also guilty of doing some recording and mixing occassionally, even given what I said above, I have to confirm this is indeed true. This is in fact one of the reasons I currently have two soundcards in my PC, because even though I wouldn't really need them, the soundcard I would generally use has some processing delays hardcoded in its Windows drivers (it's not there in Linux, so it's obviously not a hardware issue) when it comes to audio input monitoring, which makes playing guitar incredibly awkward and often downright unusable - especially playing fast solos and stuff like that, it can really mess with your timing.

That being said, I still believe this to be a somewhat different issue than what we're talking about here, to be honest.

But...yeah

I can also confirm what you and RasmusP said about audio. This is the exact reason I had to give up using the Jam Kazam app. In spite of it being a great idea, no interface through a Windows sound card could go below 9 MS. That also includes using ASIO drivers directly. Using a Behringer UM2 as recommended by the Developer, it never worked for me.
Some got it working by having custom DAW input processors built that were stand alone units that sent audio signals directly to a network card without an operating system.