Cores vs frequency for muhgayms

Core count doesn't matter all you need is more MHz this is literally another bulldozer for gamers

videocardz.com/66354/core-count-vs-frequency-what-matters-for-gaming
computerbase.de/2017-02/cpu-skalierung-kerne-spiele-test/#diagramm-watch-dogs-2-fps

Other urls found in this thread:

cpu.userbenchmark.com/Compare/AMD-Ryzen-7-1700X-vs-Intel-Core-i5-7500/3915vs3648
tweaktown.com/news/56422/amd-ryzen-7-1700-4ghz-1800x-performance-329/index.html
techpowerup.com/230541/8th-gen-core-cannon-lake-over-15-faster-than-kaby-lake-intel
videocardz.com/66354/core-count-vs-frequency-what-matters-for-gaming
computerbase.de/2017-02/cpu-skalierung-kerne-spiele-test/#diagramm-watch-dogs-2-fps
twitter.com/SFWRedditVideos

You're a faggot

We don't know how well Ryzen will overclock yet. If the quad or six-core processors can hit 4.2-4.6GHz they'll keep up with Baby Lake. If not Ryzen might be a workstation chip.

Then why is a 5Ghz fx-9590 performing like a 3.4Ghz core i7 in your little unnamed benchmark? For many DX11 games or serial processing tasks in general performance is dictated by BOTH clock for clock efficiency and overall clock-speed.

those fx processors don't have hyperthreads, what the shit is this chart?

>Then why is a 5Ghz fx-9590 performing like a 3.4Ghz core i7
Because Bulldozer has very weak cores? Duh, it doesn't even have 8 FPUs so it's not even in the same ballpark.

Also these are like 17 games aggregate.

At worst we're looking at similar performance to current Intel processors, however you also get the benefit of more cores, which could still apply to games if you do other things at the same time, like stream/encode video.

4M8T. They don't have 8 cores.

4 modules 8 threads you halfwit.
Only Intel has 'hyperthread', it's fucking SMT and 'hyperthread' is just a marketing name.

>can raise mhrz count just be changing some options
>can't raise core count

AMD wins again

Exactly my point to OP, just because you have a high core clock does not mean your performance will be better then a lower clocked design that has a higher clock for clock efficiency. Its not just Ghz if it were netburst Pentiums would have been awesome at something other then space-heating.

Nobody needs more than 4 cores if they do they're just pretending to need them

Modern games are mostly garbage gameplay and story-wise but at least they started using more than 2-4 cores, it's inadvisable to get a 4 core now when an 8 or 6 core is similar to the 4 core's price and has per core performance within 5% of the 4 core.

>Core count doesn't matter all you need is more MHz this is literally another bulldozer for gamers

whynotboth.jpeg
Intel BTFO

So what you're showing is that there is a measurable difference between 4, 6 and 8 cores right now?

That's what I'm seeing.

>So what you're showing is that there is a measurable difference between 4, 6 and 8 cores right now?
Of course there is? Even in gaming.

Nobody needs more than 1 core, if they do they're just pretending to need them.

We do know one thing, the 6 core won't OC any better than the 8.

'Ryzen Master' allows you to disable cores entirely, so an 8core will always be able to OC to the same level as a 6core (since both are 2x 4core complexes, 6core just having 1 core per complex disabled)

Who knows about 4 core though, could be a single complex could be 2 complexes with 2 cores disabled per complex.

Games don't utilize more than 5KBps of Motherboard Transfer Core Power

>game doesn't use more than 4 core
>higher clocked 4 core wins by some 5%

>game uses more than 4 cores
>higher clocked 4 core gets its shit wrecked literally
Obviously, there's no way a some 500MHz can keep up with more cores when cores are actually needed, and they'll just be more needed in the future.

Now if you only tested 2010 games the 4 core would win.

The 8 core reached 5.2 stable on all eight cores when heat was removed as a factor (LHe cooling). That's about as good as Haswell-E gets under the same conditions.

If Ryzen can reach the frequencies Haswell-E can, Intel is fucked

I've been hearing some guy mention he got the 1700(cheap $320 8 core at 65W) to 4.1, but unknown if its with the new amd tool or through the BIOS

If so that's good indication the X versions should be pretty decent.

they're fucking listed as 8 cores everywhere. fuck off

With all 8 cores enabled.
It should do better than that with only 6cores enabled.

That's one thing Intel can't do on the fly, disable cores.

>intel fucked by AMD housefires barely closing the gap

>AMD housefires
Way to out yourself as either a shill or fanboy...

>barely closing the gap
They closed the gap, AMD has better IPC than Kaby Lake.
Moreover they're doing it cheaper than Intel.

>better IPC than Kaby Lake.
Hey hold your horses, this claim needs proof.
And proof of not just a single thread benchmark testing the FPU and leaving the int units to sleep, show me benchmarks where it does both better then you can claim this is true.

Well unlike me, you don't need to "out" yourself.

Not that guy, but it's pretty close.
cpu.userbenchmark.com/Compare/AMD-Ryzen-7-1700X-vs-Intel-Core-i5-7500/3915vs3648

They aren't 8 core processors fuck boy.
They share cache and can't effectively compute at the same time

>buy 4c/8t Ryzen CPU for $199
>overclock it to 4.2Ghz with stock cooler
>performs exactly the same as 7700k
problem solved

Doesn't count.

Can you provide a source to a report of a modern (last 10 years) AMD CPU or GPU actually being determined to be the root cause of a lost property fire or disaster?
Thank you.

Or better yet buy the 6 core for $250
Overclock it to 4.2
Performs better than 7700k

IPC depends on the workload, and should therefore be averaged over a wider range of benchmarks.
I'm not saying that Zen's IPC isn't good or that this benchmark is wrong, just that it won't be ahead of Kaby Lake in every workload you throw at it.

Why?

Can you deny AMD's over ten years of just forcing more power and always having shit thermal performance all around?

Ofcourse you can, wouldn't be a real fanboy if you coudln't.

That graph is percentage of a baseline chip, but doesn't have the 100% chip listed for some reason. Oh, and Computerbase has mentioned they have a R7 1800x in for testing, but has to wait for the NDA to expire before they can say anything about it. ;)

It's not reliable.
Only after you put a benchmark that hammers two main compute parts of the core (fpu/alus) can you claim the IPC is better than kaby lake.

Personally I'm expecting AMD to lose the FPU by a few % with ops that aren't higher than 256b but win the integer race, so tl;dr depends on the workload, cinnebench for example is more FPU centric.

I love that bench so much as an AMD customer.

exactly, that's an even better suggestion

just need to stop letting people meme the idea that the 8c/16t 65w R7 1700 is worth buying for gaming when a 6c/12t 95W 1600x is 10x better

>Performance[in nazi](FPS)

Clearly it's not percentage.

Anyone thinking the 1700 is for gaming and not a budget workstation chip is going to be disappointed.

1700 with a B350 is a steal for a great all around desktop though.

Userbenchmark does, it tests Int and FP separately and then in a combined test.
Not to worry, only 7 days until we can see Sandra results along with Linpack and sysbench.

>Data in percent

So Sandy Bridge is 50% slower than new chips in gaming?

SANDYMANGINA KEKS ON SUICIDE WATCH

Depends on the game, but newer games compiled with recent compilers - yes.

Intel's perf hasn't really moved much in 5 years, (you can see as much if you test with older benchmarks that are provided binary only) - but it also has, a lot.
Newer instructions cut down on clocks required to the get the same result and there are more clocks too.

The 4 core versions are, in modern games.

But I was told 4 cores is all you need

No shit

Ryzen in real world is going to be a disaster, completely useless.

AMD keeps doing this shit over and over again.

>50%

Looks like 100% to me

100FPS is 100% of 50FPS, not 50%

2 cores is all you need tbqh. the pentium is the best budget cpu right now and only has 2 cores. it runs modern games perfectly fine according to all the reviews i've watched.

It is.

But why settle for what you need, and get what you want?

I feel like someone's lost in time and space.
Or just very good at selective filtering of unwanted information.

>100FPS is 100% of 50FPS
wew

IF YOU GOT 50FPS
AND YOU WANT 100 FPS
YOU NEED ANOTHER 50 FPS
BUT THOSE 50 FPS ARE YOUR ENTIRE FPS
SO THAT MEANS YOU WOULD HAVE TO NEED 100% MORE FPS

100% more = 200%

Why get 4.5Ghz 7700k instead of getting 3.7Ghz r7 1700 for cheaper? It's easy to overclock it to 4.0Ghz on all cores or 4.5Ghz on 4 cores reaching kaby lake single thread scores.

That's why in your graph 6900k, 6950x and 6850k are all shitting on 7700k and 6700k while still having a much lower clock frequency, right?

because the 1700 is a 65w part and is total shit at overclocking

you need a 1700x to achieve stable OC over 4ghz

this is just a prediction but it's well informed and probably correct

It's sarcasm.

Some people here like to pretend games are still in 2009, or they're just salty as fuck because they recently bought a 4 core for $350

Because the 1700 has a TDP of 65w and won't go past 4Ghz.

If you want to overclock, just grab the 1700X

>because the 1700 is a 65w part and is total shit at overclocking
If the 1700 can hit 4.0 on all cores then it's great, that would obliterate any Intel 8 core and put the 10 core in spitting range.

I can sacrifice some 200-300MHz for cheaper price, it's no problem, I won't sperg over 6% FPS

You can exceed the TDP you know...

4.0Ghz on 1700 already confirmed
tweaktown.com/news/56422/amd-ryzen-7-1700-4ghz-1800x-performance-329/index.html
You just need a good phased motherboard to supply all the voltage needed.
If you want to overclock over 4.0Ghz turn off 4 extra cores.

Don't see how higher power consumption and thermals in the context of non flash point temperatures matters, especially not in the minor difference between AMD and Intel thermals. Regardless, don't see how it's a house fire without a report of actual house fire. Please help me understand your viewpoint in a clear and constructive manner.

Or just 2 cores even.
Or 6 cores if you're playing some old ass game that doesn't multithread at all on PC (like GTAIV) (graphics driver still needs a render thread, so don't go crazy and disable 7 cores)

Of course that is possible too. AMD new overclocking utility and (possibly UEFIs too) allow for turning off cores and tweaking per core voltages so you may not even need to turn off anything, put higher turbos on some cores and lower on others for optimal balance

doesnt the 1700x/1800x have a higher tdp than all the other ryzen variants?

BUT BUT BUT BUT BUT I WAS TOLD NO GAMES WILL EVER USE MORE THAN 4 CORES

amd falsely advertised them as 8 cores, and got called out on it. it was everywhere, have you been living in a cave?

do you see the literal 5 fps difference. they fucking don't.

this. amd poorfags really want to believe that they can get a cheap 8 core that will be better in all situations even though it's common knowledge that you can't utilize that many cores outside of specific workloads such as video encoding

Obviously they do since the 8 core at the top of the chart is clocked 700MHz lower than a 7700k and on a older architecture.

it's specifically in watch dogs 2 which is one of the few games that can use more cores, and even then it's not a huge difference, and it looks like they ran the 7700k on stock clocks, with 5 GHz overclock it would be more fair

> Processor Base Frequency 3.40 GHz
>Max Turbo Frequency 3.60 GHz

Lmao so just overclock it by 200MHz and it's already faster than your 4 core?
Not to mention when it's actually at 4.3-4 most BDW-E can do it would obliterate your shit 4 core.

They do but besides, you shouldn't post a console with that comment. The target audience of consoles is entirely different from "gaming" hardware. You know, consoles are for people who actually have a job and just want to relax and play some games while PC gaming is basically for NEETs who get their money by screaming at their parents and defecating in the living room.

they're max overclocked 5 GHz 7700k vs 4.3 GHz 6800k. good luck beating them with ~3.7 GHz mid tier ryzen

Most PC games are console ports though, only the biggest shitports like Batman would use 2 threads.

techpowerup.com/230541/8th-gen-core-cannon-lake-over-15-faster-than-kaby-lake-intel

AMD tards BTFO cannon lake already 15% faster

I like the fact that Kabylake was 15% faster than Skylake.

Really.

Canard PC get 5.1 Ghz on single thread with air already and that was engineering sample with a shitty motherboard.

>single thread
lolololol

One more reason to stop being a NEET and get a console so you can focus on the more important things

OCing depends on a lot of things, cooling, motherboards, silicon, patience.. Wait a few days.

>moar cores
AHAHAHHAA

there's a sweet spot in single-thread vs multi-thread performance vs cost. the typical user doesn't need 8c/16t

The typical user doesn't need more than 2 cores and a console.

These aren't typical user CPUs, for fucks sakes they're over $400 only those who need those will buy them.

>the typical user doesn't need 8c/16t
Two years ago users didn't need more threads as well, remember how i5 was recommended instead of i7? Trend has moved to 8 threads now, and even more threads will become mainstream in not so distant future. Even Intel sees this, new coffee lake will be 6c/12t. With Ryzen 1700 being cheaper than 7700k it's hard to argue against it.

>only those who need those will buy them

No, there are enough dumb kids with money, buying anything because it's produced by brand X.
There are enough people to pre-order hardware without having even read 1 review.

There's no point in sacrificing massive multi-threaded gains for 2-4% FPS in games, well.. unless everyone's suddenly a "competitive Counterstrike and DOTA/LoL gamer"

4 cores is more than enough for watching anime and shitposting

2 cores are more than enough for that.

Don't post this or Intelfags might start stabbing and burning you alive.

>>videocardz.com/66354/core-count-vs-frequency-what-matters-for-gaming
>computerbase.de/2017-02/cpu-skalierung-kerne-spiele-test/#diagramm-watch-dogs-2-fps

Sorry, don't buy it, I need actual tests from reputable sites in America instead of nazi shit and chinks

>(LHe cooling)
nobody cares about this childish bullshit

this. let us know if you can even do 4 GHz on all cores on air. and supposedly ryzen has around 93% the IPC of kaby lake so 4 GHz is only equivalent to 3.7 GHz on kaby lake

These benches clearly demonstrate that MOAR COREZ does not matter whatsoever. The determining factor is IPC and clock speed--it's that simple. If you guys can't determine this from these charts then I have no idea how you function in day-to-day life because you are simply unintelligent.

>Gibbo explained: "We just tested a 1700, it hit 4.0GHz stable in everything, but ONLY in the Crosshair mainboard, the lower-end boards it was hovering around 3.80GHz as the VRM's were cooking with extra voltage. It however was maxing around 4050MHz, so I'd say 1700 can do 3.9-4.1GHz, of course the 1800X will probably do 4.1-4.3 as no doubt better binned, but if your clocking the motherboard has a big impact on the overclock and so far ASUS Crosshair and Asrock Taichi seem the best two".
lol you need to spend extra on a fancy motherboard to even max out at 4 GHz

Then why is a 6900k, you know a 8 core with 800MHz less base clock leading the pack?