cpu "bottleneck"?

[DarkSmith2]

4 minutes ago, mr anderson said:

 

i know this video, telling only half of the story...

[mr anderson]

1 minute ago, DarkSmith2 said:

i know this video, telling only half of the story...

whats the other half?

[DarkSmith2]

5 minutes ago, mr anderson said:

whats the other half?

Frametimes, Frametimes = / = Framerates.

Bad frametimes = Stutter 

 

take a look at the post i made earlier showing graphs.

Edit: But also, in every heavy CPU bound game, like CSGO or WoW, or d3 or anything that dont need alot of GPU power.. 8700k = 2600k x2 FPS wise.. btw..

[mr anderson]

2 minutes ago, DarkSmith2 said:

Frametimes, Frametimes = / = Framerates.

Bad frametimes = Stutter 

 

take a look at the post i made earlier showing graphs.

so you're suggesting that i just get a better resolution like 1440p or 2160p and stick with my system besides gpu?

[DarkSmith2]

13 minutes ago, mr anderson said:

so you're suggesting that i just get a better resolution like 1440p or 2160p and stick with my system besides gpu?

i think your GPU isnt capable at all for anything higher than 1080p. Without taking money in your hand you wont achieve anything. Either GPU+Monitor(+2600k) or CPU+Board+RAM.

 

i played with my 2600k since 2011, upgraded GPU to 970 and monitor to 1080p 144hz in 2015.  Had stutter... couldnt get rid of it. Played on it until ~3month ago. Bought 8700k+board+ram -> no stutter at all... very smooth gaming experience still using the same GPU until something better gets released.

 

FPS in GPU bound games didnt changed alot, FPS in CPU bound games got doubled >.<y But the reason that there is no stutter is, much better frametimes.

[Streetguru]

1 hour ago, DarkSmith2 said:

They benchmark using frametimes, then turn that into fps for graphs, and they also list .1% and .01% frame rates.

[mr anderson]

17 minutes ago, Streetguru said:

They benchmark using frametimes, then turn that into fps for graphs, and they also list .1% and .01% frame rates.

i could simply not get a new cpu and go for a new monitor + gpu something like an 3440x1440 monitor with a new 1180

and be set for god gaming or no?

[DarkSmith2]

9 minutes ago, Streetguru said:

They benchmark using frametimes, then turn that into fps for graphs, and they also list .1% and .01% frame rates.

what they do is :

Average FPS, 1% low, and 0.1% low times are measured. We do not measure maximum or minimum FPS results as we consider these numbers to be pure outliers. Instead, we take an average of the lowest 1% of results (1% low) to show real-world, noticeable dips; we then take an average of the lowest 0.1% of results for severe spikes.

[Streetguru]

Just now, DarkSmith2 said:

what they do is :

 

 

If I"m not mistaken, PresentMon outputs the frame times, which they turn into fps for simplicity using an internal script
 

 

[DarkSmith2]

16 minutes ago, Streetguru said:

If I"m not mistaken, PresentMon outputs the frame times, which they turn into fps for simplicity using an internal script
 

 

and it does matter what, if they dont publish the right graphs showing the stutter all the older CPUs are producing? Im not even going to watch your video, because there is literally 0 difference in lying or not showing important data that can affect the whole picture of a products performance.

 

I quoted their testing metholodogy if they use AMD OCAT overlay or MSI aftburner showing 0.1% and 1% lows doesnt make any difference for the reader/viewer of the review because they are only showing an average out of X 30sec. tests without pinpointing anything. An indepth frametime analysis shows like the graphs i posted that severe stutter excists which is not even closely mentioned or visible in the graphs they are showing. ¯\_(ツ)_/¯

[mr anderson]

at a high res. would i need to even upgrade my cpu, for example 3440x1440p, i'd figure at that point i should just get an 1180?

 

[DarkSmith2]

7 minutes ago, mr anderson said:

at a high res. would i need to even upgrade my cpu, for example 3440x1440p, i'd figure at that point i should just get an 1180?

 

1080p is fine bro, if you would have atleast a 6GB version of the 1060 you could easily just buy a 1440p monitor and play on low refreshrate. Since you have only a 3GB version just buy new CPU+board+RAM ... 

 

But a 2600k alone would be a waste of money. Wouldnt solve your problem.

[mr anderson]

1 minute ago, DarkSmith2 said:

1080p is fine bro, if you would have atleast a 6GB version of the 1060 you could easily just buy a 1440p monitor and play on low refreshrate. Since you have only a 3GB version just buy new CPU+board+RAM ... 

alright thank you, and I have never been able to game at higher resolutions, is it that much better?

[DarkSmith2]

3 minutes ago, mr anderson said:

alright thank you, and I have never been able to game at higher resolutions, is it that much better?

dont even try, you get addictive and see yourself buying GPUs every new release because your FPS suck.. 

[mr anderson]

1 minute ago, DarkSmith2 said:

dont even try, you get addictive and see yourself buying GPUs every new release because your FPS suck.. 

lol thank you for your help

[DarkSmith2]

1 minute ago, mr anderson said:

lol thank you for your help

youre welcome ^^

[Streetguru]

2 hours ago, DarkSmith2 said:

 

That kind of stutter is going to show itself in the 1% and .1% frame rates, like it does in some games under crossfire/SLI that they've reviewed/benchmarked.

The first question in this Q&A video has some more detail
 

 

[DarkSmith2]

4 hours ago, Streetguru said:

That kind of stutter is going to show itself in the 1% and .1% frame rates, like it does in some games under crossfire/SLI that they've reviewed/benchmarked.

The first question in this Q&A video has some more detail

Not at all.. if this would be true the 1600x would stutter like a 2600k, but it doesnt.  The 1% and .1% framerates are just useless in this case.

 

There is a reason AMD made public gameplay tests with Ryzen.  Where people werent told which system is AMD and which is not.

[Streetguru]

7 hours ago, DarkSmith2 said:

 

Except neither of those CPUs were stuttering in his testing of watch dogs 2?

Here's what it looks like when there's poor frametime performance, which translates directly into low .1% and 1%

[DarkSmith2]

1 hour ago, Streetguru said:

Except neither of those CPUs were stuttering in his testing of watch dogs 2?

Here's what it looks like when there's poor frametime performance, which translates directly into low .1% and 1%

Doesnt make sense either, Steve himself mentioned that he is filtering lowest drops as "outliers" in his testing meholodogy. Meaning: He is filtering bad frametimes, manipulating testresults and showing only a span of an average of again 30second tests... in other words. Lying. As far as personal experience goes close to every CPU bound game stutters at 1080p with a 2600k, unless your graphics card is so bad that it bottlenecks it beforehand.

 

But im tired to discuss with you saying all the things over and over again just because you cant accept facts. Have a nice day.

[Streetguru]

8 minutes ago, DarkSmith2 said:

Doesnt make sense either, Steve himself mentioned that he is filtering lowest drops as "outliers" in his testing meholodogy. Meaning: He is filtering bad frametimes, manipulating testresults and showing only a span of an average of again 30second tests... in other words. Lying. As far as personal experience goes close to every CPU bound game stutters at 1080p with a 2600k, unless your graphics card is so bad that it bottlenecks it beforehand.

 

But im tired to discuss with you saying all the things over and over again just because you cant accept facts. Have a nice day.

I'm not sure you understand what stuttering is exactly, if that were the case every single CPU would stutter at 1080p in high fps games like CS:GO, which they don't.

Outliers would be things like the first frame of Metro Last Light which was like 0-1 on every system due to how the benchmark worked loading in.
Or GTAV going above like 180fps which causes stuttering. Because clearly that isn't normal behavior.
 

 

[gamertaboo]

So just to hop in here on this topic if I may.... 

 

I've done some testing on my own. I decided to build my 9 year old a budget gaming rig in the last two years or so, and I had some hardware lying around that I figured would be plenty, but found out it was absolutely not....  I started off with a GTX 970 by the way, which to me seemed like plenty of HP for a 9 year old at 1080p.... But I decided last minute to put it on ebay to see what it would sell for and it sold for $215. So I turned around and bought a 6GB GTX 1060 the same day for $230 shipped, and so it cost like $15 to upgrade the GPU. (hopefully pricing gets back to this norm at some point)

 

As for the CPU I had an FX-8320, gigabyte 970 mobo and 8GB 1600Mhz DDR3 lying around (which I understand is quite different to the hardware you guys are currently talking about). But when I paired it with the GTX 1060 it gave me the same type of performance that @DarkSmith2 is talking about. At 1080p60, even though it could fairly easily hit 60fps+ in most games there was very often some stuttering, or such horrible framerate fluctuation as to never give a smooth experience in almost any games. 

 

So, keeping budget in mind I was able to get an H110 motherboard, an i5 6500 and 8GB of 2133Mhz DDR4 for super cheap. (See- Super cheap)

 

This immediately showed drastically different results, it easily way outperformed the FX-8320 in everything. Not necessarily min/max framerates, but more importantly in stability of framerates. 

 

Having said this, the i5 6500 isn't enough in all games now days, and I would really suggest getting a 6 core whether you choose AMD or Intel, and if you can swing it get one with hyperthreading or SMT. I would suggest Intel personally, since you are trying to reach high framerates on a high refresh rate monitor. Ryzen 2 does seem better at this, but Intel is still ahead for max framerates. 

 

Anyway, if I were to do this budget build again right now, this would be what I buy right this second, and it costs $328.11 TOTAL - 

Motherboard and ram - $133

i5 8600 - $195

 

That hardware with a GTX 1060 should make for a decent gaming rig at 1080p for quite a while to come. It's too bad you don't have the 6GB 1060 though...

(also faster memory would be worth it if you can swing it, if not the above will do the job)

-----------------------------------------------------------------------------------------------------------------------------

As a side note I would disagree that Steve removing the outlier frametimes being "lying." They are outliers, meaning they don't represent what you're really actually seeing while playing, and would only serve to skew the benchmark results.