Final Word on Ryzen 7

Well, it's time to face facts, Sup Forumsentlemen.

As an alternative to Intel's current HEDT offerings, Ryzen 7 is a clear winner, offering 80% to 90% the performance of comparable Intel chips for half the price or less. Anyone looking to build a workstation or streaming box without breaking the bank should be looking at AMD right now.

However, with limited single-core performance and a high price point compared to faster Intel chips, Ryzen 7 is a poor choice for gaming. Anyone who was looking for Ryzen for their next gaming build should either wait to see how Ryzen 5 fares or simply buy Intel instead.

>B-b-but muh 4k benchmarks
At high resolutions like 4k, the GPU is the bottleneck in the system, rather than the CPU, so gaming benchmarks done at higher than 1080p are not indicative of actual CPU performance.

>B-b-but muh optimization
While it is true that optimization may help boost Ryzen's gaming performance in the future, right now Ryzen appears to excel in workloads that are highly parallelized. Due to the fact that games are typically forced to wait for player input before deciding what tasks to perform next, they are difficult to parallelize and thus there is a limitation to how well they can be optimized for a higher thread count, lower thread performance CPU like Ryzen 7.

>B-b-but muh Joker Productions
Right now his benchmarks appear to be an unexplained outlier. However, if accurate, they may point to the Ryzen 7 1700 specifically being a compelling choice for gamers if heavily overclocked.

Other urls found in this thread:

youtu.be/BXVIPo_qbc4
reddit.com/r/Amd/comments/5x4rtc/ryzen_gaming_benchmarks_summary/
it.pcpartpicker.com/user/ConcreteShelter/saved/M8hK8d
twitter.com/SFWRedditGifs

fuck amd bunch of lying naggers

There are major errors regarding smt, among other things. If ironed out it will give ryzen a boost putting it even with Intel's offerings. It's only 10-20% behind.

>Ryzen 7 is a poor choice for gaming
What did he mean by this?

See bottom of the OP.

real Sup Forumsentoomen will make the switch.

>the Ryzen 7 1700 specifically being a compelling choice for gamers if heavily overclocked
You forgot that the 7700k is also heavily overclocked in that benchmark.

No, I didn't. 5GHz is not a heavy overclock for the 7700K.

>Final Word on Ryzen 7

>One day after hard launch

Stopped reading there mate.

Why is overclocking even an argument?
Most people never bother with it.

>5GHz is not a heavy overclock for the 7700K.
Delusional fags will perpetuate this meme till death, won't they?

It's not the CPU's job to be massively multithreaded, it's the GPU's: AMD kids are illiterate in technology.
They pretend a CPU is a GPU..

Brainlets don't overclock

he means that it has worse performance when benchmarked (1800X vs i7 7700k vs 6900K)

>ignoring parts of the posts you don't like

the performance gain is mostly minimal and is shit compared to the increased heat and power drain which comes with it.

Right now i feel a sensible option might be to get a Ryzen 3 or Ryzen 5 and if they really want a 7, wait for the platform to mature.

Hopefully AMD will stick with the same chipsets and firmwares will get updated for a t least 2-3 years.

Even if 5GHz was a substantial overclock at 5GHz, it wouldn't matter much. The 7700K shits all over the 1700 at stock clocks. And it is a minority of users who will be willing to overclock and not all 1700 chips may overclock as well as Joker's.

Derp, I said 5GHz twice.

I'd honestly like to see the ryzen benched with 460 or a 1050, after all it's what a typical AMD buyer can afford. Much fairer benchmarks

>gaming benchmarks done at higher than 1080p are not indicative of actual CPU performance.
So what you mean to say is that people willing to drop $500 on a new cpu don't have 1440p/4k monitors, thus making ryzen no good for gayming?
seems a little convoluted desu famalam.

...

Not what I'm saying at all. I'm just saying that benchmarking at those resolutions is bad, not that the CPU shouldn't be used to play games at those resolutions.

Nah, if anything I'll wait for Zen Server CPUs.

Legit

Finally a voice of reason

Joker's video was legitimate though. You can watch his results in real time and it puts the 1700 right on par with the 7700k.

Then you remember the 1700 is cheaper and comes with a stock cooler that isn't shit

The 7700k beats the 1700 by %20~, with all the issues. If it can close the gap to about ~5%, it will be a real contender.

Hi OP.

>WILL COMPLETELY BTFO INTEL AT CORE 2 LEVEL ACROSS THE BOARD
>B-BUT IT'S GOOD AT CINEBENCH
Amdfags are the most obnoxius pieces of shit and wonder why they're universally despised for the hype they build. People should have learned already from their past bullshit.

>20%

You fucking delusional? They're neck and neck trading blows

youtu.be/BXVIPo_qbc4

they could replace the i5/i7 sandy/ivy in the $10-$40 sector. Pretty cheap, ECC & M2 SSD slots make them an attractive package for low end servers.

Yes, with jokers build that does not seem to have those aforementioned issues.

I'm giving the benefit of the doubt to Intel with the assumption that he got a golden chip.

Remember, reasonable thread is reasonable.

actually someone did a pretty good sum about some of the benchmarks
reddit.com/r/Amd/comments/5x4rtc/ryzen_gaming_benchmarks_summary/

seems like asus mobos and asrock and msi has quite a lot of problems on bios not to mention that they get wonky results

gigabyte mobos are by far the best ones amd send the worst possible mobos out there

>Ryzen is good
>Motherboards are shit

So this is why Joker's got such good results? I think he had the gigabyte motherboard

No benchmark is legitimate and should be taken with a grain of salt unless you stand in the fucking room when it was undergone. This fat brown shill has to be only one showing such performance and his results are stacked against 20-30 different reviewers that all show something else.

He could do 50 different things to skew the benchmark and looking how desperate he is with all those videos about Ryzen I wouldn't be surprised.

Keep in mind that the GPU core speed on his Intel rig is slightly lower which could be caused by so many things (thermal throttling, motherboard using some sort of auto OC, him enabling or disabling high performance profile in Windows or just straight out using different GPU or limiting it in some way). I don't say that 13 Mhz different matters but it's just odd.

Re read his post, friend.

13mhz is well within margin of error.

Keep it friendly, buddy.

>mfw when there are at least 2 Pentiums faster than the 1700

this is why everyone get different results..they were purely random from being -8% from 7700k to -30%

i explained it yesterday on some thread here that its quite a lot of combination doing wonky stuff..

the L3 is "broken" and shared along with ccx's and that wouldnt be a problem if the smt wasnt tapping on the caches

BUT the good stuff starts with windows because they treat every thread as a cpu and they load balance them incide the cpu and that is very BAD for ryzen because amd treats every ccx as a cpu and thus when windows throws the threads around it lets the data of them on the L3 randomly without knowning where it was they therefore start a new cycle->parking the core thus leading to the high perfomance bug (which is already solved with an update from yesterday on windows) BUT the smt will take months to fix (it took months to ms to fix the HT back when intel introduced it)

>Stacked against 20-30 reviews

So, gamers nexus? The one review out of 30 that was a blatant hit job?

Joker's is more reliable at this point. He even went out of his way to do a follow-up showing it REAL TIME.
youtu.be/BXVIPo_qbc4

Steve from GM hasn't shown shit.

Basically by introducing a GPU bottleneck you end up compromising the relevance of the actual benchmark. This is because the CPU isn't doing most of the processing, rather the GPU, meaning that even meager performers in the CPU market can be made to look good with a GPU bottleneck. That means that by relying on the bottleneck to meter performance you end up buying a CPU that will rapidly fall. That said the benchmarks do indicate that it doesn't, itself, bottleneck anything.

Furthermore this is the first iteration of the Ryzen series of CPUs and they are early into the release process so one might speculate that:
A)Future iterations, esp. 4c 8t, and 6c 12t will perform more adequately for a gaming load
B)Software optimizations may either marginally or substantially improve performance
C)A housefire

And additionally as far as I know, the more advance features of the series (de-coring) haven't been tested extensively and as such posture an interesting mystery until they're pushed adequately.

Personally I'm waiting to see what the 4c8t CPUs perform like before I pass a judgement onto the architecture itself, and I'm interesting in seeing what kind of results the collusion between AMD and Bethesda generate, if any, and whether or not devs will be opting to optimize for SMT and higher thread counts.

>LOOK AT THIS ONE COMPLETELY UNBIASED POSITIVE REVIEW BY A LITERAL WHO ON YOUTUBE WHO SELLS AMD MERCHANDISE
>THE REST ARE ALL FAKE AND WRONG

ah yeah the GN that suddenly got praised here because it exposed amd

so is it worth going from an E5-2670 to Ryzen?

>E5-2670
yes especially if you're going to do more single threaded tasks

GN was praised 24/7 thanks to this one retard who kept making threads and posts about potential reviewers Sup Forums can to trust before NDA got lifted. Many anons named gn among trusted reviewers.

I love how it's been less than 24 hours and you guys do a full 180 in damage control saying that gn cannot be trusted while posting retarded pictures like yours when clearly all major players in reviewing showcased how shitty Ryzen is for gaming. Ryzen is over, sorry bud but your mental gymnastics won't change it.

Not if you already have one.

Oh I get that, but from everything i've seen, it's still not even a CPU bottleneck. I know it's going to take time for the programs that are used for testing to mature. But I'd like to see at least steady 65-75% utilization before I feel judgement can be passed.
Moar cores is the future, game vendors are going to have to cater to it.

Objectivly false. If GN leaked reviews they would logically be negative.

Fat spic productions intentionally made the gpu the bottleneck in order to hide cpu performance.

>whatthefuckamireading
i have the most basic overclock of all time on my 2600k, literally only changed the turbo multiplier to 46 in the bios and every other setting is on auto and it is basically on par with a stock 7700k

i find it really hard that not a single typical intel shill site brought this up (because ALL OF THEM did those tests) and only gn did and in such a way like amd actually tried to force them to do it..

the gpu is pegged at 100% almost constantly so that is a gpu bound benchmark and not really indicative of the raw performance of a cpu.

that aside 1700 will obviously perform better everywhere else so if the res and quality he's running at fit your use case you'd be stupid to not get the 1700

you forgot that soon the Ryzen 5 and 3 should come out and it might be just the same case with intel that their 4core have better single core performance therefore better for gaming. we need to wait and see

Did you even read their review? They never claimed that AMD forced them to do it, in fact they actually stated that they contacted AMD first trying to ask for tech support to find out if there is something they can do to make AMD look better. AMD's response wasn't anything along the line of "do x to fix it", no. Their response was "do benchmarks where GPU hits limits way before CPU gets to show any performance gains". GN disagreed with this advice since it does absolutely nothing for benchmarking the CPU and hides the ugly fact that those frames will also be lower in 1440p and 4K when we get our hands on better GPUs. That's it. Now AMD and Intel shills try to spin it into some sort of unholy crusade against AMD/GN.

the single thread performance is lackluster, but I did some research and could just upgrade to an E5-1650V2 and 1866MHz memory for significantly cheaper than buying a Ryzen + motherboard + DDR4 memory

>At high resolutions like 4k, the GPU is the bottleneck in the system, rather than the CPU, so gaming benchmarks done at higher than 1080p are not indicative of actual CPU performance.
This is retarded. When you saturate the GPU, CPU variation is more pronounced instead of compensatory.

Yes. That's why every CPU including i3 are within 1 fps margin error in 4k right shill?

Not true.
Massive parallelism on the GPU is only suitable if the task at hand involves doing the same operations on a lot of data.
If you want to do different computations in parallel then CPU cores are better, as they are way more flexible and actually independent.
GPU "cores" cannot execute different instructions independently. If you have branching in your GPU code you sort of serialize the execution of those different code sections.

More CPU cores is great, because we can use them for things we can't use the GPU for.

b...b...but our customers DESERVE to know that our Ryzen(tm) will perform that one fps BETTER than Intel(tm)

>thumbnail says raw performance
>both cpu's at 50%
>gpu at 100%

fat spic productions at it again

I don't even know where to begin with your broken braincells.

Let's try this:
You have two people. One person makes a left-foot shoe, and the other makes the right-foot shoe.
We'll call left-shoe person the CPU and the right-shoe person the GPU.

Now the left-shoe person can make 10 shoes a day. That's great, but GPU man can only make 5 shoes a day, but they need a complete pair, ready at the same time for the same customer, to sell the shoes at all.
Now, if left-shoe man only makes 6 shoes a day, or 15 shoes a day, how will this effect how many shoes per day they sell (IE how many frames per second you see)?

You sure know about broken braincells. You should take your condition more seriously.

Obviously if one piece of hardware has constant output (i.e. saturated GPU), the variations of the other (CPU-bound processing) become clearer.

how long until we see motherboards with AM4 sockets with DDR4-3600 speed?

Seems like Ryzen was made to be a server in mind,

Sooner than you'd think if this bios fuck up turns off consumers.

Hashing performance is extremely good as well. Ryzen should be a good hit with people looking to upgrade for servers.

I hope in the next 3 months, famalam

You're wrong. is a good analogy. At high resolutions, the GPU has far more impact on FPS. Therefore, high res benchmarks aren't a good indicator of CPU power relative to other CPUs.

Bought Kaby Lake three days ago.
Balls of steel or brains of mush? We will see.

Mostly went for maturity of platform/chipset etc, (Kaby Lake is certainly derivative but there is another side to that).

Main point of build wasn't really performance, its quiet (PC is now utterly silent) and chipset features.

Old gamers exist, time > money.

I think Ryzen could still be cool, but I'll leave the beta testing to others.

Yeah that' why for the last 15 years cpus were tested at low res. It was the same even when amd was on top. Sorry I forgot you were still in your diapers 5-6 years ago.

A shame it only has 4 DIMM slots though.

This limitation might be mitigated if it does support registered/multi-rank DIMMs which I'm not certain it does.

Assloads of RAM is what servers need.

Ryzen doesn't like fast RAM.

>1700 compelling choice for gaymurs if heavily overclocked.
Fuck, I was gonna use it as a heavy-duty code machine and clock it stupidly high. Are you saying it's the compelling choice due to the Wraith Cooler reportedly being able to overclock at the same level as a Hyper 212?

Wow, this looks like a great workstation chi-
>no dual CPU motherboards
>IOMMU groupings place PCIe slots in with SATA and everything else, preventing you from passing through a GPU
o-oh..

> 2017
> Upgrading for gayming purposes

Enjoy choking on that marketing dick. Does the mind tendrils feel good? Yes master I will buy new gayman gears, MOM I NEED TOP SCORE IN BUTTFIELD h.256

>for half the price or less.
maybe in usa
in EU were gonna get fucked no matter what we wanna buy.

Also seems like the SMT bug is found. Fixing it mostly on Microsoft's end.

>Taking the bait
I've got an R9 290X (for computing everything and driving primary monitor) and R7 260X (for driving left and right monitors)

And then you imply I can't afford anything better than a 460/1050, and I am frankly and personally offended by this poor attempt at an ad hominem attack.

What?

He is a meme spouting retard that forgots that at one point Intel also suffered from HT overhead and solving this issue didn't give them magical performance.

AKA: he is suicidal because Ryzen is mocked for shit performance in gaming.

But Ryzen wasn't intended for gaming with high IPC but shitty clocks and a massive amount of cores.

>Ryzen is mocked for shit performance in gaming
What. It performs well enough for gaming.
Where it shines is on workstation loads.

Let's say I have a 9 years old PC, I use it mostly for gaming, and I have 1250€ to spend on a new one.
My preferred choice would be the 7600k, since it's the best CPU for gaming for the buck, but everyone said WAIT FOR RYZEN. So I waited, and it seems it's not as good.
What should I do? Should I make the order?
I'm tired of waiting and not being able to play shit.

it.pcpartpicker.com/user/ConcreteShelter/saved/M8hK8d

>Bottlenecking a gtx1080 at 1080p
You are pretty fucking retarded

Unless Ryzen 5 makes more performance per core, I'd say just go Intel.

If you do other stuff on the side that can profit from more cores, get the current Ryzen 7 1700 probably.

Wait for ryzen to hit the market so the price drops retard.

and how long will this take?

That's the thing. How much until Ryzen 5? More months? It probably will come out during my exams session (June/July), and I'd like to make the order before the end the month. this way I'll be able to play at least for a month before I have to get back to studying. Even if Ryzen 5 proves to be much much better than 7 in terms of gaming, that doesn't mean that my 7600k will be obsolete in less than 3 years, since I also plan to overclock later. Right?

It already hit the market, I don't see any price dropping. At least here in Italy.

Ryzen is still pulling really good numbers. It's up to you if you want to support the blue Jew or not.
There's not much in it between them(at this point in time), but with how many threads and cores ryzen's got, and how long you keep your PC's for, it should be the obvious choice.

Q2 is Ryzen 3 and Ryzen 5.

Just get a 7600K and throw 5GHz on that bitch, AMD will likely not reach any similar gaming performance until ZEN2 or something when they reach higher clock speeds or get even better IPC.

PCPP hasn't updated with Ryzen yet, but for gaymin go with a 1700 and a B350M motherboard. Ditch the aftermarket cooler, replace with Noctua when you have the dosh, and replace the 1070 with an R9 Fury if possible.

Get Zen. Unless you have a 144hz screen. In which case you should buy Zen and OC it. If you plan to keep the PC relevant for at least 5 years, you wouldn't buy a 4 core processor.

Oh -
This is given if you do nothing else than gaming and the sole purpose your PC exists is games.

>look ma I posted my favourite photoshop again

>poor choice for gaming
>intel has a ~10% advantage at manlet resolutions only

>B350M
Isn't that a MoBo that doesn't overclock?

Guess what, when we have GPUs capable of high refresh rate 1440/2160p, Ryzen is still shit but current intel CPUs are fine.

Our expectations were decently hedged it was false flagging by fags

Bleh, there's still no AM4 brackets for my AIO so I can wait said months. A year even.