Why is single threaded performance so important for gaming? Battlefield

>DX11
nice bait

>>61741077
The default core usage is not 8.
Most reviews doesn't know this because they don't play games often.
https://forums.battlefield.com/en-us/discussion/55246/gstrender-thread-maxprocessorcount-how-many-cores-unparked-i7-6700hq

>DX11 in 2017
>not upgrading your life with far superior DX12 performance

>>61741077
Take from a Newegg customer review:

>This give you a lot more headroom than something like a i7-7700k because where as that chip does have higher single thread performance, its also capped out for what it can be in the future. In other words its already being fully utilized where as Ryzen 7 still has unused compute power existing in its extra cores. I think over the next 10 years this is going to be quite a good chip.

So basically just wait and over the next 10 years there will be software that will let Ryzen surpass the 7700k.

>>61741103
Waiting for something to be good is a garbage argument.
Why not just wait for the new processors that'll be out in 2027?

Look up Amdahl's Law

That is why Rypoo can never outperform Intel Core

Guys is it even worth waiting for Coffee Lake? The 6 core would be nice but for a mainly gaming PC will it be much better than Kaby Lake?

File: bf1.png (688KB, 1920x1080px) Image search: [Google]

688KB, 1920x1080px

>>61741077
It's certainly strange how (((Gamers Nexus))) always manages to get significantly worse Ryzen results than other reviewers. I wonder why that would be the case?

>>61741077

>Gamer's Nexus

>>>Reddit

>>61741098
if you had read that thread, there's a post that explains that adding that line to the config file actually breaks the config and causes the game to lauch at reduced graphics settings, which then shunts more load to the CPU as it tries to handle the increased framerate, thus increasing % load. It doesn't increase thread count. The OP there even mentions that his graphics are fucked when he uses that line. Other posts around the internet point out that the graphics settings cannot be modified if that line is present.

It'll probably be next year before we really see games that take full advantage 8+ threads, but the cat is out of the bag as far as 8c CPUs go, so there's no turning back now.

>>61741077
>Why is single threaded performance so important for gaming?
A combination of "proper multi-threading is hard" and "game programmers are incompetent".

File: maxresdefault.jpg (27KB, 1440x1080px) Image search: [Google]

27KB, 1440x1080px

>>61741185
nono, everything is alright with GaymersNexus

>>61741099
>dx12
>both lose performance but ryzen loses less

lazy dx12 devs

>>61741222
>It'll probably be next year before we really see games that take full advantage 8+ threads

I've been building gaming PC's for 10 years and they said this every one of those years.

>>61741185
GayMansNextAss has a golden sample 7700K that overclocks to 5.1Ghz that's why.

>>61741077
Is there any benchmark that show performance on more than one task? I want to see how both perform in gaymes while playing video in HD on second monitor.

>>61741306
It's not even about that though. Their BF1 number for Ryzen is 30fps below what Hardware Unboxed achieved with the same CPU in the same game on the same settings. Even the result with shitty 2666MHz RAM is significantly better.

>>61741299
turns out 90% of people bought i5s

>>61741222
>It'll probably be next year before we really see games that take full advantage 8+ threads
Technically they propably are using all the threads available. Usually there is some kind of threadpool/executor that executes the tasks.
Problem is lot of stuff requires syncronation and using locks has overhead.

>>61741319
I don't think an HD video is taxing enough to make a significant difference.

>>61741299
Oh, I didn't realize that 8c processors had been made affordable every year for the past ten years. Where have I been?

>>61741299
Well, we're already seeing i5s shit the bed when it comes to minimum framerates.

Guys game devs are going to start making games for 8 core CPU's any day now. It will really help out AMD. Sure it will run like shit on the i5's that 90% of people have and they won't buy the game but game devs don't care about that. They will take the loss to bring PC gaming into the future and save AMD.

>>61741363
That was just an example I want to see how they perform with 4monitors each doing different task because that's how I use my pc.

>>61741430
Intel is releasing i5s with 6 cores, you do know that, right?

Never buy hardware for "future proofing". It's retarded and a waste of money. By the time future arrives there is new any much better hardware anyways.

>>61741481
That's dumb logic.

My i7 870 is still performing great for its age because more and more threads are being utilised in newer software. In dx12/vulkan it should be punching above an i3 7100 and i5 7500 thanks to it having 8 threads, which was considered massive Overkill for its time.

>>61741116
You're missing his point. The software isn't able to utilize the capability of the cpu. It's a software/computation/coding problem and not a hardware problem or 'bottleneck' as you /v/ users say.

>>61741103
either they're being Intel shill, or everyone else is an AMD shill. pick either or both.

>>61741077
*tips tinfoilhat

They probably tested a different part of the game

Also Hardwareunboxed is a fucking AMD shill, he tests, he tested overclocked ryzen cpu's versus non overclocked ones in his R3 review even though there was a k cpu in there

>>61741077
is this another GN shilling video?
man they just cant stop

>>61741530
I think it just means that while it's not as good as the 7700k at this moment, it'll stay relevant for longer as more threads become utilised in future programs.

Similar to how my i7 870 aged >>61741511

>>61741103
Or, by that team, you buy the cpu that performs best then.

I got a 6700k for a cheapo and oc'd it to 4.6 ghz.

Whenever it will start to struggle I will just get what performs best then, at that moment.

Who the fuck buys a cpu that 'might' be better in 10 years, it's such a horseshit argument, by that time there will be much better cpu's.

>>61741584
That's nice but 7700k performs better than 6700k so you aren't following your own rule really.

>>61741584
I guess the fear is buying hardware that is obsoleted in 10 minutes

>>61741103
They said the same with the FX processors, I doubt game devs are going to waste time making their games use all cores.
But I still got a ryzen because fuck intel mayo paste.

>>61741602
a 7700k basicaly is a higher out of the box oc'd 7700k.

But it still only boosts to 4.5 ghz on one core.

My 6700k runs at 4.6 ghz on all cores

The differences become so marginal then that I was better off getting a 6700k which was sold for 100 dollars less

>>61741628
Ok but you can overclock the 7700k too.

>>61741584
ever wonder why gn never put 1600 against 7700k?
https://www.youtube.com/watch?v=UfNMn7RWgLw&t=260s
yeah thats why..they cant really shill for intel with a cpu that has less tdp less temp and more cores with less mhz and still perfoms the same as a highly clocked 7700k

>>61741628
> higher out of the box oc'd 6700k.

>>61741643
>still perfoms the same as a highly clocked 7700k

doubt.jpg

>>61741643
It doesn't perform the same lol, check the video
>>61741641
Right, but the differences are marginal

>>61741641
Intel said not to overclock you fucking amd shill.

>>61741658
At least you can overclock Intel cpu's.

Ryzen shits itself when going past 3.9 ghz for most people

>>61741653
You can oc your 6700k to 4.6 you can oc a 7700k to 4.9. That should net a pretty decent bump in performance.

>>61741676
Not really

>>61741653
half the games are in favor of 7700k
half in favor of 1600

ITS NOT THE SAME

>>61741709
What games does 1600 win? I've been watching for 5 minutes and 7700k wins every time.

>>61741730
so you basicly watched 5 mins literally 4 games only out of 30 and concluded that its not..

/typical intel user

>>61741306
>has a golden sample
It's clocked at 4.9ghz in most tests
Not a golden sample as 67% can hit 4.8+ GHz
His new 7700k is golden as fuck though
Able to hit 5.3ghz
He said it wasn't fair to include it though
That being said, 4.7-5.0ghz results in like a 2% difference. Kinds pointless to bitcy about it.
>>61741346
>Their BF1 number for Ryzen is 30fps below what Hardware Unboxed
>I'll take "What are different settings" for $500 Alex.

>>61741754
Name one game.

>>61741779
find the summary near the end
its simple
youtube a bar on the video that allows you to move the feed
since 2007

>>61741794
You can't can you.

>Battlefield 1 is a game that actually uses all of Ryzen's extra threads

It doesn't though, I really can't think of a game optimized to take advantage of 16 fucking threads. I've seen a few select titles use 12. But those are insanely rare. If the CPU market goes the way of extremely high core counts and threads you'll see games move towards that. But when for the longest time you saw consoles using less then 4 cores, that translated directed to the PC space as game devs are lazy as fuck. It was for this reason alone we saw the i5 dominate for so long in the gaming space.

But with the 8 core CPUs in the PS4/XB1 we've see the 4c/4t CPUs lag behind the 4c/8t CPUs in most major games. This is the way the market and industry moves. Ironically this move can hurt the 4c/8t CPUs in certain applications like game streaming because games are designed to use all 8 threads, so you'll find yourself resource limited as games want to make use of all the CPU resources. But that's just one aspect of the debate between the 7700k and the 1700.

It's a really interesting time for consumers. It's no longer 2011 where the i5 2500k completely shit on the FX 8150. Now your choice is all on your use case. Of course the debate means complete and utter nothing if you're just gonna be gaming on a 60hz monitor anyway.

>>61741077
single core programs are much easier to create and manage
in fact any multi-core program requires a lot of additional code that slows down performance
obviously this only means that multi-core programs will be less efficient per core, but with multiple cores it gets faster despite the additional code

>>61741077
>DX 11
Thanks for holding back technological progress with your retarded obsolete software.

So should I get Coffee Lake or is that going to be worse than 7700k in games?

>>61742024
Coffee Lake is just Kaby Lake with two extra cores. An Coffee Lake i3 will be equal to a 7700K.

File: 1492367786158.jpg (138KB, 751x749px) Image search: [Google]

138KB, 751x749px

>>61742024
Most likely worse.
The 7700k has an edge in games because of high clocks. If they add 2 more cores while retaining the TDP they have to clock it lower, there's just no way around it.
It most likely won't overclock as well, too, since the mayo they put under the heatspreader won't be able to dissipate the increased head effectively. Unless you take the risk and delid.

pic unrelated

>>61742097
Is it completely out of the question that they'll solder it?

>>61741077
games dont parallelize well
it's alot of extra work for minimal performance gains, or even perfomance losses in some cases

>>61742141
You mean that the current tools used to build games don't parallelise well. There's no reason to tick each actor sequentially for example. There's no inherent reason why you couldn't parallellise the workloads, other than the engines not supporting it well yet. You'd have to make games programmers think in new ways though, and these are people who write in C++ because that's what they use.

File: 1497745287609.jpg (21KB, 480x480px) Image search: [Google]

21KB, 480x480px

>>61742118
Yes. Even their shiny new HEDT line has mayo instead of solder.

>>61742047
Do you think a Coffee Lake i3 will be cheaper than a 7700k? Worth waiting for the price drop?

>>61742188
yes

>>61742188
It will be cheaper but I doubt they are going to let their i3 perform better in games than their i7. It will be gimped somehow.

>>61741541
He's only following official Intel recommendations that tell you not to overclock your k CPUs goyim.

>>61742165
no I mean the fundamental nature of simulations doesn't parallelize well
Take physics as an example, objects that all need to move and collide with each other need to be done in serial or else the whole system is unreliable and things will start intersecting with each other and bugging out. Any tightly connected system is not going to multi-thread well and a game is really one big tightly connected system

>>61742211
>It will be gimped somehow.
You bet. It's called "less cache" and it has a noticeable impact on performance.
Also, no OC unless you buy a board that costs double the price of the CPU.

>>61742257
intel will start selling HT unlock keys again like its 2010 again

>>61741077
It's really easy to write single-threaded applications and really hard to write multi-threaded applications. That's It.

File: 6C9F0D46-6D50-4DE7-ADDB-E4D0F4DBB568-1665-0000033875155093.jpg (14KB, 443x332px) Image search: [Google]

14KB, 443x332px

>>61742188
Price drop? Good joke goy!

>>61742169
Whats with his face...?

>>61741185
That's not a gays nexus benchmark though

>>61741765
>I'll take "What are different settings" for $500 Alex

They're both running the Ultra preset at 1080p you fucking retard. It says so right on both graphs. Get an eye test.

File: f533f7efc00996b728b9f8ca375640d5.png (26KB, 160x160px) Image search: [Google]

26KB, 160x160px

>Multithreading in a workload where the whole experience depends on the coherence of the world

What are you going to multithread? You can't multithread collision or hit detection or entity or projectile movement.
It's just an inherently single-threaded application unless you add constant costly thread synchronization which destroys all benefits of multi-thread operation

>>61743697
>You can't multithread collision or hit detection or entity or projectile movement.
You actually can. You just have to be smart about it.

>>61741077
Because 8 cores is of little use when there are only two tasks maybe three. Ryzen only works well as a rendering chip, everything else it loses to Intel.

>>61743974
Actually its quite the contrary, high fps gaming is the only scenario in which intel pulls ahead by a little margin.

>>61742188
>price drop
AHAHAAHAHAHAAAHAHAAHAAHAAHAAAHAAHAAAAAAHAAHAAHHAHHAAHAH

File: aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS84L0svNjYyNTY0L29yaWdpbmFsL2ltYWdlMDQ4LnBuZw==.jpg (37KB, 600x450px) Image search: [Google]

37KB, 600x450px

>>61741077
Why do these results look so fucked up? Even Tomscuckware looks better.

>>61742687
Look at where his right eye is holding the camera and then look at where the viewfinder is.

>>61746082
Oh... Though he was deformed or something. Kek.