So will the R5 and R3 series be better targeted at 'gaming'?

So will the R5 and R3 series be better targeted at 'gaming'?

They will have fewer cores, and higher clockspeeds for greater single threaded performance.

They're very, very cheap too.

why would anyone buy them for gaming? the ipc is still dogshit. The cores are the same as r7

cores of Kaby lake are the same as Sandy bridge, must mean that Kaby will perform same as Sandy!

What? The IPC of Zen is almost identical to Intel's, plus they're much, much cheaper.

i think by "ipc" being "dogshit" he means cowadoody at 1024x768 runs at 280 frames instead of 290

>1600X
>6C/12T
>4GHz
>$260

4C/8T variants even cheaper and higher clocks.

>playing games when you're over 15 years old

This

It's based on R7, it's going to be a complete disaster

Really?

The only reason R7 doesn't do as well against the 7700K is because it doesn't have the same single threaded performance, games don't utilise the 8 cores properly.

R5 will drop the number of cores and increase the clock speed.

can u post link for image

1600x has same clocks as 1800x?

but i thought you liked dogshit user?

>watching anime when you're over 12 years old

They'll be good alternatives to the Core i5 and i3, and probably age a lot better. If you're expecting something better than a 115x i7, that's not forthcoming, and anyone who knows what they're talking about could have told you as much 5 years ago.

the r3 1100 is going to be the de facto standard budget gaym box cpu with the 1400x being a compelling budget alternative to 7700k

DESTROYED!

But your picture shows at the 1600x has the same clock speed as the 1800x

>wait fags
Just buy 7600k

>and higher clockspeeds
No they won't. The top 6 core part has the same clocks as the 1800X.
Voltage is the limiting factor in clock scaling. None of the Summit Ridge chips are going to clock high.
The Raven Ridge chips are made on a different process, and will have a wider range for clock scaling.

Raven Ridge will have even higher voltages.

Check out the voltage on the Bristol Ridge APUs on AM4.

that nigger looks like an ayylmao

look for it yourself fatfuck

Thats a retard false equivalency.
Raven Ridge is not made on the same 14nm LPP Vt as Summit Ridge.

>Sup Forums babby
RMA yourself

The r3 1100 has to be either cheaper or faster than the G4560 for it to be any use. Both seems unlikely from what we have seen so far.

If you think the voltages are going to be different or improved somehow, beyond the node shrink, you're going to.be SORELY disappointed.

What process will it use? 14nm HP from GloFo?

>I have no idea what I'm talking about: The Post
Cute, kid.

No, I doubt they'd be running bulk and SOI concurrently. The Raven Ridge parts will likely be 14nm LPU.
10% frequency uplift, and it scales down much further.

But does it have better ICP is the question, if it's exactly the same chip then r6 is going to be pretty shait for gaming.

>higher clockspeeds
That's not higher, same clock as 1800X.

R5s: Same single core speed, half the price.
R3s: Same single core speed, quarter of the price.

Where were you when intel was BTFOd forever?

...

>Where were you when intel was BTFOd forever?

With a 1700 awaiting motherboard stock.

No.
Its already leaked the maximum frequency will be around 4ghz.

Different from intel, that run at least 2 different lines to bin (2011 and 1551) Ryzen from r7 to r3 will be the same production binned, so its not a different architecture like kaby vs broadwell e.

R5 & R3 when?

Honestly is there a release date for them?

ANOTHER MAJOR DISSAPOINTMENT

next quarter

so before june

Just need a silver painting kek

Don't tell me what to buy fuck you

>R7 can't OC very well, even on just a 2 cores
>R5 clock speeds the same as the R7
>less cores

actually it's about 10% lower and they clock 10-20% slower ergo performance is significantly worse

>higher clockspeeds for greater single threaded performance.
Uh, no. It's just going to be fewer cores for cheaper prices.

1600X is what you should be look at if you're into gaming, but you won't get better single threaded performance than the 1800X.

At least, not before overclocking is brought into the equation.

No CPU on 14nm LPP can into muhghz.

But a quad core could be clocked higher while maintaining the same tdp, so it should be much better.

Personally I'm going to get a 6c12t ryzen if they can reach 4.3ghz since the 8 core ones seem to peak at 4ghz.

>But does it have better ICP is the question

Depends on how much Faygo it gets

You can't clock it higher. Heat is not the issue, voltage is. 14nm LPP is not suited for high clocks, it's sweet spot resides at 3.3ghz.

Overclocking on Ryzen is not being limited by power. Its voltage and heat.
The 1800X wants 1.475v for its 4.1ghz XFR boost.

Lower tier SKUs aren't going to change a fundamental characteristic of the process.

Holy fuck how hard is it for you guys to understand they won't be clocked higher?

At best, they might XFR to 4.1ghz on 4 cores instead of 2 on the 8 core parts.

The design is simply not made to clock high. They're made for servers. They almost surely wont' clock over 4.1 stock.

The majority of games release the past 4 years do use 8 threads.

It's getting to the point where 4 threads is definitely not enough.

>The r3 1100 has to be either cheaper or faster than the G4560 for it to be any use. Both seems unlikely from what we have seen so far.
this
exactly

The IPC is fine, its whether or not AMD will figure their shit out when it comes to their SMT code. Either way, a r3 zen chip with 4 cores clocked higher, will make intel drop their 7700k prices again. So it's good for everyone.

>But a quad core could be clocked higher while maintaining the same tdp, so it should be much better.
hey look it's the muh binning argument all over again heh
funny how that worked out for you guys

It will certainly be faster than g4560 by having 4 physical cores.

BUT AMDFRIENDS TOLD ME THAT AMD 14nm IS BETTER THAN INTEL

You can't clock 4 cores higher than 8 cores. Read the fucking thread.
It's better until it hits 3.4ghz.

>10yrs of refinement of an architecture

VS

>first attempt

My memory is hazy as i'm a bottle of vodka deep - how well did the pentium 4 perform again?

It does appear to have some significant advantages in the range of 2.1ghz to 3.3ghz. Zen shows totally linear scaling in this range.
Part of this efficiency is Zen being a smaller core, another part of it is the process itself simply having more favorable electrostatic characteristics.

Not enough to bottleneck a GTX 1080 at 1080p, that is.

An i5-7600k is fine for 60-120fps in those games. But the hyperthreading in the 7700k gives a 20-45% performance increase at the same clocks in almost every game that's come out the past 4 years except for horribly programmed indie shit.

SMT is better than HT in synthetics and workstation tasks. Something is seriously wrong with how games or win10+games is utilizing it.

>It will certainly be faster than g4560 by having 4 physical cores.
stands to be seen honestly, even the R7 is only 5-10% ahead of the G4560 in many games

CA is working on fixing SMT in TW:W so we have to wait and see how well Ryzen actually performs in CPU-bound gaymen.

Because it's fucking bugged you fuckwit. Let them fix their arch before Naples.

>But the hyperthreading in the 7700k gives a 20-45% performance increase at the same clocks
ehhhhhh

> But the hyperthreading in the 7700k gives a 20-45% performance increase

>Intel shills believe this

Because Intel cannot into SMT the gain is closer to 25% even in synthetic workloads. Leave it to AMD and IBM to (once again) show Intel how its done.

>stands to be seen honestly, even the R7 is only 5-10% ahead of the G4560 in many games

This is obviously due to bugs and shit. You can't really be so stupid, can you? Some of these gaming results don't line up with all the other results and it's a brand new architecture.

HT was getting 5-45% performance drops with HT enabled in games in the beginning as well.

The 1100 is surely going to be better than the 6100, 7350k, g4560. And the 1300 is surely going to be better than the 7400, 7500, and the 7600k.

This is the average of over 20 different CPU tests.

I doubt 1300 will be better than 7600k performance-wise.

>voltage and heat
Well it should be easier to manage with less cores. It won't be much, but a higher overclock should be possible.

Are you suggesting that 4 cores will use just as much power and put out just as much heat as 8 cores?

No but if you can hit 3.9ghz stable across four cores, rather than just one or two running at 3.9 and the rest running at 3.2. That alone makes a difference for gaming, especially if they can get their multi threading working.

It would be marginally better. Again, the process is not suited for muhghz. Leave housefires to Intel.

So my point stands.

>mostly bugged SMT gives 28% gain
>decade old mature SMT gives 25%

So at face value AMD's un optimised (be it software and/or hardware) SMT has slapped the shit out of Intel. Still lags behind IBM but then again, IBM is the final solution for x86 and POWER.

I don't think the r5 with outperform Intel but if the value is good then it'll probably be worth getting instead (since most games are bottlenecked by the gpu anyway).

I know this has been asked 100+ times but are the SMT and Bios bugs fixed now?

No. Wait a month or so.

>Because it's fucking bugged you fuckwit. Let them fix their arch before Naples.

Thanks dood

>This is obviously due to bugs and shit. You can't really be so stupid, can you?

Oh look, a live specimen of Sup Forumstard.

>over various tests
How fucking stupid are you? You think the varience is only +/- 5% per application?

No in some tests AMD SMT only gets a 16% gain due to SMT. In others it gets a 68% gain.

Why do you bother wasting your time typing and others time reading your dumb shit when all you have is your stupid preconceived notions of something but no actual data or facts behind it?

There was negative SMT yield in some applications.

That was the point, just providing data to support your claim.

AMD's SMT implementation showed a major performance hit in Linpack. Were it not for this result then the over all average would be even higher in AMD's favor.
Looking an averages of numerous specific workloads is the only way to get a real picture of generalized performance.

Which should never ever happen. It rarely happens with HT because developers test with VTune to see where they should make sure threads are assigned to a physical core instead of using HT.

So the reality is that when developers fix using AMD SMT in places where it negatively impacts performance (or it could be a Win10 issue), then those gains of AMD SMT over HT will be even higher.

Yeah, you beat me to it with your post.

Go look at benchmarks that aren't GPU limited that compare the 7600k and 7700k.
They're roughly the same clocks. Difference is hyperthreading.

Lower clocks but with SMT will win on the majority of games from the past 4 years.

>Using the exact same "insult" over and over and over
you aren't clever kid

We have no evidence that the R5 and R3 parts will clock higher. Apparently the R5 1600X has the same clocks as the 1800X. The benefit of the R5 series is similar performance in games to the R7 series, but cheaper.

RMA yourself