CPU Benchmark in realistic condition (multiple programs running)

turn_pike

Senior member
Mar 4, 2012
316
0
71
#1
So I was reading yet another vociferous debate on the merit of intel vs amd cpus here on AT forum. Amongst the mudslinging there were a couple of interesting points being raised:

- Modern games requiring more and more thread; 4 years old i7 2600 can sometimes outcompete modern i5
- Most benchmarks are done in sterile fashion to prevent confouding variable so almost no other programs are running which can be unrealistic

Now I dunno about most gamers but I usually have multiple programs running while gaming. Firefox and Chrome with double digit tabs each or Torrent, IRC, IMs, Skype, SSH, Netlimiter, Virtualbox and so on. Most of the time they wont take much cpu power but they do add up. 20% or so cpu used is rather common.

Moreover streaming or encoding your play to upload it later to youtube is something that is becoming more and more common. Sometimes I even record myself playing in order to review it later and learn from my mistakes (playing mobas or RTS).

Just having a "dirty benchmark" in general would be more useful.
I hardly ever stop my other programs before gaming, at most I would set the cpu affinity when encoding / compiling. I would think many enthusiast here in AT would feel the same. Most people at least dont quit their browsers while gaming.

With that in mind, is there a website somewhere that test for things like this ?
I'd love to know if i3s and Pentiums are still be the cpu of choice if you want to game and stream for cheap ?
How would the benchmark results look if multiple programs are running in background and you're also streaming your play ?
 
Last edited:

Burpo

Diamond Member
Sep 10, 2013
4,050
25
126
#2
The new i3-6100 has respectable performance on a budget.. not many benchmarks yet tho..
 

turn_pike

Senior member
Mar 4, 2012
316
0
71
#3
The new i3-6100 has respectable performance on a budget.. not many benchmarks yet tho..
That is missing the point.
What I'd like to know is if they still are good while streaming and recording your gameplay. There's also 30 tabs browsers and multiple programs running in the background (unlike benchmark scenario which is usually clean windows install).
 

TheELF

Platinum Member
Dec 22, 2012
2,708
70
106
#4
- Every game comming out is a console game,even pc "exclusives" are being programmed on the same engines and thus have the same requirements,the only exeptions are MMOs/online and older (engine) games that run on less threads.

- Most sites do in game benchmarks which only test the GPU,they do not run the game code so they are pretty much useless.

- Most games don't have the same requirements throughout the whole game,depending on where you benchmark you will have drastically different messurements.
(Crysis 3 for instance, "Welcome to the jungle" is a completely different game compared to any other part of the game)

You don't need a lot of cores to do what you want,if the rest of your work uses up 20% then in the worst case your game will play 20% slower.
Since most of the times people play with gpu bound settings (way higher quality than their card can handle) they won't notice any difference.

Inquisition on g1820 with only recording running.
https://www.youtube.com/watch?v=XykMNHlYIOg

Inquisition on g1820 with recording,browsing and kodi running.
https://www.youtube.com/watch?v=uFPXTHm4Q1M

Yes you do loose speed,but if you have more then enough,or your GPU throttles you anyway,then what's the difference.

Afinity is a bad way to controll programs since it messes with the task scheduler,you can use priorities from task manager to tell windows which programs to favor over others.
 

Leyawiin

Diamond Member
Nov 11, 2008
3,191
1
91
#5
Now I dunno about most gamers but I usually have multiple programs running while gaming. Firefox and Chrome with double digit tabs each or Torrent, IRC, IMs, Skype, SSH, Netlimiter, Virtualbox and so on. Most of the time they wont take much cpu power but they do add up. 20% or so cpu used is rather common.
I would never do that. When I'm gaming its on a full screen with nothing else of any consequence going on in the background. Maybe you want benchmarks that show a cluttered process list, but the whole point of professional reviews is to isolate the performance of the given hardware they're looking at so the reader can see exactly its benefits and flaws with as little interference or demands as possible from OS related sources.
 

turn_pike

Senior member
Mar 4, 2012
316
0
71
#6
I would never do that. When I'm gaming its on a full screen with nothing else of any consequence going on in the background. Maybe you want benchmarks that show a cluttered process list, but the whole point of professional reviews is to isolate the performance of the given hardware they're looking at so the reader can see exactly its benefits and flaws with as little interference or demands as possible from OS related sources.
Do people have problems understanding what I've written. I just wrote that I understand wanting to isolate confounding variables.
The point is to ask if the benchmark would be representative at least for the portion of gamers who do run other stuff in the background.

Many people whom I observed while they're gaming in real life dont close their browsers and torrent when gaming. Neither their skype, whatsapp web or IRC or whatever communication they prefer. More and more are picking up streaming. More and more are learning the value of recording and reviewing their gameplay to learn from their mistakes.

If as mentioned in the threads many games require more and more threads, modern games being optimised for console having 8 cores. Would gaming in 2C/4T i3 be compromised in a dirty setting (many programs running in the background, at least encoding/streaming the game).
 
Last edited:
Aug 11, 2008
10,457
66
126
#7
I would never do that. When I'm gaming its on a full screen with nothing else of any consequence going on in the background. Maybe you want benchmarks that show a cluttered process list, but the whole point of professional reviews is to isolate the performance of the given hardware they're looking at so the reader can see exactly its benefits and flaws with as little interference or demands as possible from OS related sources.
Exactly. Why in the hell would you have 30 tabs in a browser for instance open during gaming? Nobody is that good at multitasking!!!! Personally, I dont see the point of having more than a few tabs open even when not gaming, but that is just me.

I can see maybe having encoding or other such tasks running, but one does not game 24/7. The logical thing is to close out as much as possible to get the best gaming experience, and leave the background tasks to run overnite or when one is at work and not using the computer.

Edit: OP, it is not that we dont "understand" what you are saying. We just think you are setting up an illogical scenario. I think even streaming/recording can be offloaded to the gpu now, but neither of those interest me, so I am not up on the most current software. Even then, I would pick the system for the most optimal performance on a "clean" system, and if it is not performing well, turn the background s*** off. Simple. As someone else said, no website can possibly know or test all the possible combinations of conditions that could exist with multiple background tasks running.
 
Last edited:

coercitiv

Diamond Member
Jan 24, 2014
3,063
336
136
#8
Just having a "dirty benchmark" in general would be more useful.
No it wouldn't.

If you want to stream while gaming, benchmark your streaming solution, then add the necessary resources on top of base gaming resources. Streaming alone can bring down a system if it's not properly configured relative to system specs.

Adding 30 tabs in your browser of choice does not necessarily increase resource usage, other than maybe some RAM. In fact browsers try quite hard to not use CPU resources while minimized. Adding such a variable to a test would likely end up benchmarking browsers/websites/webapps rather than CPUs/GPUs.

Would gaming in 2C/4T i3 be compromised in a dirty setting (many programs running in the background, at least encoding/streaming the game).
Yes it would.
 

Erenhardt

Diamond Member
Dec 1, 2012
3,251
0
101
#9
I like to update my other games while I'm playing those that are updated. I sometimes install games when I'm playing.
There is antivirus software running in the background.
Chrome with 10+ tabs is always opened, but that doesn't use CPU much - RAM on the other hand... :)

And the most important - win10. It does so much stuff when you are not looking it is not even funny.

But I have 8 cores and half of the CPU is always idle, so I don't care as long as my hard drives can keep up.
 
Last edited:
Feb 6, 2011
1,768
78
136
#10
Exactly. Why in the hell would you have 30 tabs in a browser for instance open during gaming? Nobody is that good at multitasking!!!! Personally, I dont see the point of having more than a few tabs open even when not gaming, but that is just me.
was just playing Bf4, 36 tabs open across chrome and firefox and a 1080p Mpeg 2 stream from my iptv server( so i can listen to the cricket). Why wouldn't you?
 
Apr 22, 2012
20,395
0
106
#11
Having tabs open or watching streams doesn't really add any usage (Unless you done have hardware decode). Also it affects the GPU just as much if not more than the CPU due to decode, flash and other GPU acceleration etc.

Unless you do something extreme. Then the difference would pretty much be nonexistent in terms of CPU load. And if you do something extreme while playing. Then you belong to a 0.1% crowd.
 
May 19, 2011
12,207
329
126
#12
Just having a "dirty benchmark" in general would be more useful.
It would only be useful if the load caused by "stuff you've left open in the background" is a) somehow definitive of everyone else with the same habit and b) the loads on the various parts of the system are "near-identical in relative terms" (which is pretty much a meaningless term).

Through my line of work I've seen it enough times that browser A works better than browser B on customer Z's computer, and completely different findings on customer Y's computer. Implementations of hardware acceleration differ. Some users may be blocking Flash entirely, some only playing Flash on request, and some playing all Flash clips. The implementation of Flash also differs depending on the platform.

The "background load" you mention could result in sudden resource usage spikes that show up completely differently. Also, there could be a minor issue for example with application X and your graphics hardware that only surfaces when attempting to do many things at once including a GPU-demanding application. Or say your NIC driver might not like a particular application very much and it causes excessive processing time which isn't noticeable while outside of a game, and while in the game graphics performance looks fine, your ping time has gone to pot (or goes to pot intermittently).

Benchmarking already isn't that useful unless one sticks with the same hardware configuration, games, OS version and game configuration as the reviewer. It's always a case of saying "well, my hardware build is / will be different from the reviewer's in x, y and z respects, but I think I may approximately get higher or lower performance for reasons a, b and c". Face it, you're on your own if you decide that you want to leave a load of stuff running on your computer while also running a demanding game. There are far too many variables to consider for your idea of a "dirty benchmark" to have any meaning whatsoever.
 
Last edited:

Erenhardt

Diamond Member
Dec 1, 2012
3,251
0
101
#13
^That would imply you have to disable network connection when playing games under win10.
What if you play online games?

Win10 does a lot of CPU heavy work in the background. There are ways to prevent it, but none official or supported by MS. It often requires 3rd party tools to disable it.
 

cyclohexane

Platinum Member
Feb 12, 2005
2,837
0
81
#15
The whole point of a benchmark is to standardize as many vairiables as possible. It is counterintuive and pointless to do a "dirty" benchmark.
 
Apr 22, 2012
20,395
0
106
#17
telemetry, auto update everything, tracking and a lot more.
I dont see such usage. Can you document it?

Its not hard to count CPU usage after say 24 hours and see how much the OS uses. And that's not much.
 

Dresdenboy

Golden Member
Jul 28, 2003
1,730
0
136
citavia.blog.de
#18
Tens of open tabs are pretty common. My gf is journalist and usually keeps >100 tabs open over multiple days (w/ restore previous session option). If I play a game I usually don't close them even though Firefox can be a drag.

I'd also vote for user experience benchmarks, documenting GUI responsiveness under loads like browsing with 50 tabs full of flash/HTML5 ads or video editing (use case of a friend). That might matter more than encoding, as bigger part of time is editing, while encoding is a smaller fraction of time or done overnight. I heard, Apple has such GUI performance/user experience tests.

Game fps values are kind of a experience benchmark, as this usually inversely correlates with lag.
 
Last edited:
Aug 11, 2008
10,457
66
126
#19
was just playing Bf4, 36 tabs open across chrome and firefox and a 1080p Mpeg 2 stream from my iptv server( so i can listen to the cricket). Why wouldn't you?
Why *would* you? Do you browse 36 tabs and listen to music while playing a game? Each to his own I guess, but when I am playing a game, I give the game most of my attention.
 
Oct 18, 2015
87
0
36
#20
Why *would* you? Do you browse 36 tabs and listen to music while playing a game? Each to his own I guess, but when I am playing a game, I give the game most of my attention.
I pause pandora, leave everything else open right where i left it. I get what OP is saying here.

I don't reboot to a blank desktop for gaming, i just keep photoshop or lightroom open with ff and chrome tabs open and launch Fallout 4 and play for an hour or two. then quit out, go back to what i was doing before playing.

When i do benchmarks, i at least keep chrome and firefox open since i rarely have either closed whenever i'm on my computer these days.
 
May 19, 2011
12,207
329
126
#21
Why *would* you? Do you browse 36 tabs and listen to music while playing a game? Each to his own I guess, but when I am playing a game, I give the game most of my attention.
This. I can think of two scenarios where I've done otherwise, one being where my computer's hardware is way beyond even the recommended specs of the game, the other when I'm using a walkthrough or something similar. The rest of the time I want to ensure that my computer has maximum resources to devote to the game so it plays as smoothly as possible.
 

myocardia

Diamond Member
Jun 21, 2003
9,226
0
76
#23
I don't know about you guys, but I always launch both Prime95 and Furmark, before I start gaming.
 

Erenhardt

Diamond Member
Dec 1, 2012
3,251
0
101
#25
I don't know about you guys, but I always launch both Prime95 and Furmark, before I start gaming.
No, but I often instal, unzip stuff, update etc while playing games.

Why the difference betweenfx6300 vs phenom x6?
http://www.pcgameshardware.de/screenshots/original/2013/06/CPU-Multitasking-Test-03-pcgh.png

Also, it shows how much underutilized fx cpus are:
http://www.pcgameshardware.de/screenshots/original/2013/06/CPU-Multitasking-Test-01-pcgh.png
 
Last edited:


ASK THE COMMUNITY

TRENDING THREADS