Curious

[ChoiSiwon]

after doing research on techy stuff. and learning more. out of curiousity will 860k bottleneck a 280/x/290x for 4k gaming

[Fate]

There is not a single benchmark on the Internet testing this.

So to answer your question, maybe, maybe not.

 

Maybe not because an 8350 doesn't seem to bottleneck two 980's at 4k in the games tested http://www.tweaktown.com/tweakipedia/58/core-i7-4770k-vs-amd-fx-8350-with-gtx-980-vs-gtx-780-sli-at-4k/index.html

But maybe in CPU bound scenarios because, well, there's no benchmarks! 

[Guest]

None of those cards (except 290x but still not very well) will handle 4k gaming at all.

[CoolaxGaming]

None of those cards will handle 4k gaming at all.

The 290x does...

[Guest]

The 290x does...

You read before my edit.

[spartaman64]

it will probably be fine and not bottleneck

[Fate]

None of those cards (except 290x but still not very well) will handle 4k gaming at all.

290x can play any game at 4k, newer AAA titles will need medium/high without AA, but it can max out pretty much any older game, racing game, MOBA, CS:GO, MMO's, Arma, simulators, etc.

[ChoiSiwon]

290x can play any game at 4k, newer AAA titles will need medium/high without AA, but it can max out pretty much any older game, racing game, MOBA, CS:GO, MMO's, Arma, simulators, etc.

wat? a 8350 won't bottleneck graphic at 4k?

[Fate]

wat? a 8350 won't bottleneck graphic at 4k?

I'm not sure, judging by those benchmarks no, but again, I cannot find any other 4k benchmarks comparing CPU's and considering at high resolutions CPU starts to become irrelevant, it probably won't bottleneck any current GPU at 4k most of the time.

[Manny Calavera]

if you plan to play CS GO on 4K then ok.  Or other games like LOL, Dota etc...

 

Forget about gaming on 4k with single card and 70$ cpu.

[Enderman]

There is not a single benchmark on the Internet testing this.

So to answer your question, maybe, maybe not.

 

Maybe not because an 8350 doesn't seem to bottleneck two 980's at 4k in the games tested http://www.tweaktown.com/tweakipedia/58/core-i7-4770k-vs-amd-fx-8350-with-gtx-980-vs-gtx-780-sli-at-4k/index.html

But maybe in CPU bound scenarios because, well, there's no benchmarks! 

???

dude look at the benchmarks in the link you just posted...

the 4770k is always 5-10 fps ahead of the 8350 sometimes in just the minimum fps, but sometimes also in the average fps...

thats called a bottleneck

[Fate]

???

dude look at the benchmarks in the link you just posted...

the 4770k is always 5-10 fps ahead of the 8350 sometimes in just the minimum fps, but sometimes also in the average fps...

thats called a bottleneck

Indeed, but compare the 4930k to the 8350 and they're pretty much the same/8350 pulling ahead. 

So that would mean the 4930k bottlenecks, which I highly doubt. 

[Manny Calavera]

btw 860k hasnt got l3 cache, it has 2 less fpus or whatever they are

 

8350 with disabled 4 alu clusters in each core mutilates steamroller unfortunatelly. 70$ looks like a great deal, but if you are not on extreme budget keep away from it.

 

 

 

id suggest something like widescreen sub 1440p monitor a gtx 970 and an i5k

[Enderman]

Indeed, but compare the 4930k to the 8350 and they're pretty much the same/8350 pulling ahead. 

So that would mean the 4930k bottlenecks, which I highly doubt. 

compared to a 4770k, the 4930 IS a bottleneck because its last generation architecture.

Haswell has better single core performance than ivy bridge.

[Manny Calavera]

compared to a 4770k, the 4930 IS a bottleneck because its last generation architecture.

Haswell has better single core performance than ivy bridge.

integer performance IVY vs Piledriver is almost the same, that is true, haswell fixed this

 

have a good read

http://www.realworldtech.com/haswell-cpu/

[Fate]

compared to a 4770k, the 4930 IS a bottleneck because its last generation architecture.

Haswell has better single core performance than ivy bridge.

A bottleneck, depending on how you look at it, would be the 8350/4930k holding the GPU's back, which it does not seem to be.

If it was a bottleneck, then going from two 780's to two 980's in their test would provide pretty much no FPS increase, but it clearly did provide a massive fps increase for all of the CPU's.

[Enderman]

A bottleneck, depending on how you look at it, would be the 8350/4930k holding the GPU's back, which it does not seem to be.

If it was a bottleneck, then going from two 780's to two 980's in their test would provide pretty much no FPS increase, but it clearly did provide a massive fps increase for all of the CPU's.

thats not how bottlenecking works

 

you can have a shitty CPU, but there still will be a difference between a 970 and a 980

the "bottleneck" is the CPU taking X percent off of the theoretical performance limit

 

For example with CPU NB (for no bottleneck) GPU X would perform at 50 fps and GPU Y at 100fps

With CPU B (for bottleneck) GPU X would perform at 40fps (20% off max performance) and GPU Y at 80fps (also 20% off)

 

so in real life if you look at how a bottleneck works, it doesnt "limit" the flow to a certain value, it "restricts" the flow by a certain percentage of maximum value

you can put 2x the pressure through a bottleneck and still get more flow than 1x pressure

[Jarsky]

I submit to this discussion, that 'bottleneck' is a highly overused term and doesnt apply in most cases.

I wish it wasnt thrown around so much on LTT.

 

Bottlenecks used to happen back in the early 00's when we were having huge jumps in technology.

Single to Multi core CPU's, PATA to SATA, Multi-core (SP/CUDA) Graphics cards, North/South Bridge to On-Die, PCI to PCI-E Bus, etc....

 

I say that bottleneck with modern board design is really a misnomer in 90% of cases.

There are always going to be places to chase more performance on your setup. You increase the CPU, then your GPU becomes the bottleneck, you add another GPU and now your motherboard without enough lanes is the bottleneck. You upgrade your motherboard, and now your SSD's are the bottleneck.

 

I really think we need a new term to replace 'bottleneck' in these type of cases where you could make upgrades for extra performance, but there are no real 'bottlenecks' in the system.

[Fate]

thats not how bottlenecking works

 

you can have a shitty CPU, but there still will be a difference between a 970 and a 980

the "bottleneck" is the CPU taking X percent off of the theoretical performance limit

 

For example with CPU NB (for no bottleneck) GPU X would perform at 50 fps and GPU Y at 100fps

With CPU B (for bottleneck) GPU X would perform at 40fps (20% off max performance) and GPU Y at 80fps (also 20% off)

 

so in real life if you look at how a bottleneck works, it doesnt "limit" the flow to a certain value, it "restricts" the flow by a certain percentage of maximum value

you can put 2x the pressure through a bottleneck and still get more flow than 1x pressure

What would you say is no bottleneck then? CPU 'NB' (4690k) overclocked to 4.6GHz from 3.5GHz will still give you better fps than when it's not overclocked even when there's 'no' bottleneck.

 

In the 4k benchmarks I linked, would the 4770k be a bottleneck as well because I'm sure if you overclocked it or used a substantially faster chip, it would give you even better performance.  

And there's the issue, in my opinion, with calling the 8350/4930k in those benchmarks a bottleneck. 

If you took a very weak chip and paired it with a low end GPU and a high end GPU, it would still get similar performance to the low end GPU.

http://youtu.be/TGdo75gasaQ?t=5m39s

[Manny Calavera]

What we know now: you can use FX 8320+ cpu with 1886mhz mem and Nvidia gpu with no bottleneck

 

Id seen reports on web about 1600mhz ram causing frame drops... and ive read that AMD processors like Nvidia gpus more than their own.

 

EDIT: Fine, no bottleneck, performance degraders. Better?

Gpu is never free of "bottleneck" (performance degradation from cpu-time). Theoretically, if cpu was infinitely fast, that means it would take infinitely small time to compute, the gpus would see a massive boost and would work at their max outputting max frames.

Now thats no bottleneck.

 

 

And i already said IVY ~= FX in terms of integer compute.

In fact in some benchmarks that FX wins against 3570k ( look logans video) the FX is clocked higher, therefore more integer performance. Not 100% on that, but pretty sure.
 

[zappian]

I have no idea at 4k because at that res games arent very cpu depedant.

 

At 1080p hell yes a 8350 will bottleneck a 290x and there are many benchmarks that confirm this.

[Enderman]

What would you say is no bottleneck then? CPU 'NB' (4690k) overclocked to 4.6GHz from 3.5GHz will still give you better fps than when it's not overclocked even when there's 'no' bottleneck.

 

In the 4k benchmarks I linked, would the 4770k be a bottleneck as well because I'm sure if you overclocked it or used a substantially faster chip, it would give you even better performance.  

And there's the issue, in my opinion, with calling the 8350/4930k in those benchmarks a bottleneck. 

If you took a very weak chip and paired it with a low end GPU and a high end GPU, it would still get similar performance to the low end GPU.

http://youtu.be/TGdo75gasaQ?t=5m39s

Theres no such thing as no bottleneck. If you watched linus for long enough you would already know this because he talks about it so many times.

The best you can do is REDUCE the bottleneck as much as possible.

 

If there was no bottleneck that would mean computers could have infinite performance because there is nothing holding it back.

 

And the fact is that the 4770k is LESS of a bottleneck than the 8350, which is why it is better.

[Fate]

Theres no such thing as no bottleneck. If you watched linus for long enough you would already know this because he talks about it so many times.

The best you can do is REDUCE the bottleneck as much as possible.

 

If there was no bottleneck that would mean computers could have infinite performance because there is nothing holding it back.

 

And the fact is that the 4770k is LESS of a bottleneck than the 8350, which is why it is better.

Ok, now I understand what you're getting at, but what do you propose we should start doing when someone asks "will my FX-6300 bottleneck my 270x"?

Should we say yes it will bottleneck, just less than something like a 4690k? Or should we continue to say whether or not a GPU bottlenecks based on if it's at 99% usage or not? (what most of the forum does)

[Enderman]

Ok, now I understand what you're getting at, but what do you propose we should start doing when someone asks "will my FX-6300 bottleneck my 270x"?

Should we say yes it will bottleneck, just less than something like a 4690k? Or should we continue to say whether or not a GPU bottlenecks based on if it's at 99% usage or not? (what most of the forum does)

 

When someone asks if it will bottleneck, I usually tell them the average performance loss compared to a 4690k which is the "ideal" gaming CPU for most people

so for an 8350 it is about 0-10% less performance depending on the game, but for something like a 6300 it could be closer to 30-40% on average

 

the "will this bottleneck" question answer is always yes, but many people dont really understand that part, and I dont bother explaining it to the thousands of people that just want to know if their GPU+CPU combo will work decently