Ryzen 1700x + 3600mhz memory = 7700k gaming performance

Looks like regardless of the memory used it's still a smoother experience overall

>>59603063
A lot of games are memory bound if you remove the GPU bottleneck, so giving a 7700k faster memory will increase FPS there too. Gaming benchmarks are worthless for comparing CPU performance.

>>59603075
/thread

>>59603099
He already covered this though, apparently there was no benefit in increasing memory speeds above 3200mhz on the Intel system.

>>59603063

>Ryzen 1700x + 3600mhz memory = 7700k gaming performance

Only in well threaded, well optimized games. Any shitty console port, any game made by Ubisoft for example or any emulation still favors i7s.

Although! Ryzen is good enough to a degree that these shitty games are still very much playable with decent to good performance.

Right now core for core and clock for clock things look like this

>GAMIN'

non-OCed Ryzen = locked Intel core for core

OC'ed Ryzen is > equivalent priced Intel CPU EXECPT i7 where they trade blows but 4/8 is starting to become a ''corelet''

>Productivity

Ryzen curb stomping Intel all over the place.

>>59603063
That's cool. What about games? What is the source for that?

>>59603063
Appears better than it with the same 3200mhz memory, OP.

I'm not a retard that goes by averages.
0.1% and 1% minimum are the most important metric.

File: 1487255039638.jpg (117KB, 1440x1080px) Image search: [Google]

117KB, 1440x1080px

>No stability tests with the ram at 3600 or more
>using under powered GPUs

Oh boy another great test for some shit head youtuber.

File: 1489858911261.jpg (811KB, 2560x1440px) Image search: [Google]

811KB, 2560x1440px

One thing for sure, 4.0Ghz 1600x in gaming going to be fantastic for smoothness.

>>59603120
When all the console are running eight core AMD chips the "shitty console ports don't utilize cores" argument doesn't really hold water.

>>59603383

>When all the console are running eight core AMD chips the "shitty console ports don't utilize cores" argument doesn't really hold water.

You don't know much about optimization and CPU workloads don't you?

>>59603480
No

>>59603480
(This is the bit where you fill me in on what I've missed here).

>>59603063
how did this guy manage to fuck up his benches this hard? since when does ryzen with 2.6ghz memory match a 7700k with 3.2 ghz memory? we know for a fact that even with 3.2 ghz memory on the ryzen platform the 7700k with 3.0ghz is still faster in 9/10 games.

>>59603063
>poor gaymin perfomance
You do know it's actually not that poor when in comparison to a 5 GHz CPU, its only about ~5 FPS off.

>>59603480
Not going to? Do you actually know anything about it either?

>>59603533

Thanks for the honest answer anon.

>>59603616

Consoles have used multi core/thread processors since PS3, specially PS3 since it was sooo heavily threaded. Although since it was a different architecture and operating system, it requiered a tiring work to optimize for PC for but since new PS4/XboxOne uses AMD X86 arcihtecture and XboxOne uses Microsoft/Dx, it's easier.

BUT! There is a significant BUT! here. Console optimization works differently. It's not as simple as coding a program for general Linux/Microsoft computer. They optimize the software for THE HARDWARE ON THE CONSOLE and nothing more. Just that hardware alone gets optimized.

If developers were actually decent human beings who cared about their consumers, (PC side) it would be easy enough for them to optimize the game, specially on AMD hardware as Consoles uses 4+4 AMD Jaguar APU (which one of the reasons why (2+2/3+3/4+4 AMD CPUs aren't that effected by gaming performance compared to 4+0 core counterparts) but they are not. As Deskop CPUs are much stronger both by IPC and Clock, they just dump down the optimization process.

Hence the reason Far Cry primal that can run 1080P30fps on a shitty AMD laptop APU that is 4+4 from 2013 will struggle to get past 70 fps with Ryzen and a 1080ti and the reason clock monster/housefire i7 at 5Ghz can reach 100+ fps.

>>59603833
I see, thank you.

>>59603833
>>59603824
>>59603533
>>59603616


>They optimize the software for THE HARDWARE ON THE CONSOLE and nothing more. Just that hardware alone gets optimized.


To add that, as PCs have so much variation and developers are such pieces of shits, they pretty much either use Microsoft's compilier to do their work.

So in consoles there is this

Program directly speaks to hardware and talks to him it can undersrtand the best

But in PCs it happens like this

Program dumps all of it's requiered work on the Microsoft, Microsoft than translates that pile of knowledge and tries to distribute the workload between cores. Ofcourse, this dumping can be done more ''gracefully'' such as the case with GTA 5 or Star Citizen as these will let Microsoft use as much as threads as there are on the market right now.

Microsoft is also at the blame. Since they are pushing their ''we are google play, applestore'' and Xbox is the gaming platform shit, they really don't fund Dx12 as much as they could or should.

Vulkan which is developed by a such a small company compared to Microsoft has better ''relationship'' with hardware in Microsoft developed operating system compared to Microsofts own drivers! That's how much Microsoft and developers screws us.

>>59603063
> tested on GTX 1070
oh, so another GPU bottleneck to make Ryzen catch up I see...

>>59603928

>10% fps difference
>GPU bottleneck

How the fuck do people like you even function anon?

>>59603928
That is not a GPU bottleneck there's too much variance in those results.

Also CCX latency doesn't seem to effect performance as much as previously thought, a few percent at best
https://www.youtube.com/watch?v=Rhj6CvBnwNk

>>59603063
you forgot the significant overclock on that R7

>>59603964

It would be ridiculus for games to be effected by X+X architecture to begin with, specially games that have also been optimized for consoles.

Consoles use 4+4 APU, I think it's one of the reasons why PC exclusive titles such as Total War Games have such shitty performance on Ryzen to begin with. Games simply doesn't know what to do with CCX architecture.

>>59604026
The 7700k is at 5ghz for comparison. Very fair

>>59604026
The significant overclock on the i7 isn't worth mentioning?

>>59604026

>Significant

You can reach 3.8 on 1700 with stock coolers anon.

3.9 is a guranteed clock for 1700x as much 5.0 is on 7700k

Oy vey, you stupid nigger cattle don't see its all bottlenecks?!

Buy Intel, the bester for goymes

File: C7YCtn6XUAAv67A.jpg (25KB, 528x306px) Image search: [Google]

25KB, 528x306px

>>59603063
>>59603964
>>59604044
>Games simply doesn't know what to do with CCX architecture.

It's worth saying that there are reports about last Windows update did fix Ryzen scheduling to some degree which benefited gaming workloads in addition to others.

https://twitter.com/BitsAndChipsEng/status/843864982320267265

That would explained the huge performance uplift seen in OP video compared to old results and disabling a CCX core making small difference now in the new hardware unboxed video. With all evidence, windows scheduling was broken for Ryzen imho. With the new microcode and faster ram it helps Ryzen even more, Zen was designed as a server architecture, faster RAM benefits it a lot

>>59603063
>muh memory speed
That shit dwelling pajeet is full of crap and he's literally the only one among dozens of reviewers out there to get such results. Just yesterday DigitalFoundry already proved that Ryzen still performs like dog shit even with increased RAM speeds.

https://www.youtube.com/watch?v=TDvk9_iTq6Y

>>59603063
7700k is still cheaper, is there a significant benefit to paying more for a 1700x?

>>59604143
t. shlomo

>>59604143
Called bottle necking the system. Why the youtuber in OP got those results.

Every test I've seen with a competent GPU at it's stock settings or OCed has shown the Intell CPU beating AMD.

>>59603099
It's not the same. Basically, AMD's infinity fabric runs at half ram speed. So the frequency your ram is running at has a greater effect than on Intel systems.

>>59604044
Total war doesn't run well on anything
They only moved to a 64bit engine with their most current game, and supposedly they're implementing DX12 which has been a beta feature for 10 months (despite being a game that AMD supposedly helped with, although the DX12 implementation is better on AMD cards than nvidia ones)

File: 1488205393827.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>59604143
DELETE

>>59604143
>DigitalJew

Yes, good Goy, our reviews are completely unbiased hehehe :^)

File: 1488223097723.jpg (540KB, 1100x1002px) Image search: [Google]

540KB, 1100x1002px

>>59604256

>>59604242

>Total War

It has been an Intel supponsored game through the ages, I have played that game since the first shogun and I can't remember a time that it didn't have Intel logo on it.

For such a CPU bound game, they sure as hell ''hate'' using more than 4 threads. I don't think it will get better anon, same story with WoW and same with Starcraft.

To this day, if you trick the game engine with ''geniune intel'', AMD performance improves by 10-20%.

>>59604143
see>>59603353

>>59604361
https://www.youtube.com/watch?v=Dl5s4LWHJXU

about that.

File: no bro.jpg (68KB, 1280x872px) Image search: [Google]

68KB, 1280x872px

>>59604384
>AMD propaganda

>>59604174
Rendering, streaming/encoding, anything that isnt gaymeing and even with gaming its still more or less the same as the i7 but more efficient

>>59604361
the newest total war is amd sponsored

CA threw that intel shit in the garbage

>>59603063
>you have to buy a much more expensive mobo + memory to match 7700k performance
Wow great

>>59603063
I just want a thinkpad with mobile raven ridge, please make this happen chinknovo

How retarded do you have to be to not realize that the text there needs outlining?

>>59604730
>expensive mobo
but the anus ones are the shit ones causing all this bullshit

>more expensive memory
all memory is expensive, welcome to the 2017 ddr4 shortage friendo

>>59604783

>wanting a ThinkPad made past 2012

>>59605010
>being poor
they are just as durable as the old ones, its just the functionality and layout they fucked up, and even then modern thinkpads btfo any other laptop in those regards anyways

>>59605036

>they are just as durable as the old ones

Nice joke. Its not about the money at all. Modern ThinkPads are shitty MacBook clones, I already have a MacBook around. Their construction is low grade. Reminder, IBM barely made any profit on their ThinkPads, as the years went buy Lenovo found ways to cut costs which is why we have cheap ThinkPads today still going on sale for $1,000+ despite costing maybe around $350 to build.

Can someone with a fucking brain please just buy a 1700 and do extensive testing. How does oc headroom change with cores or smt disabled, full ram scaling, etc. I just feel like there is still a bunch of missing information. With games using so few threads maybe going down to 4 cores same ccx no smt for increased oc headroom with top tier ram will make it a monster?

>>59605137
>build quality is shit
rossman already debunked this, they still have the magnesium alloy rollcage, they have carbon fiber lids, everything else is a high quality plastic (it doesnt feel cheap)

>>59605192
25% of 1700 can reach 4ghz so its a bit of a gamble, higher clocks with one CCX disabled is a given

>>59605286
But how much higher, is it worth it? Where is the sweet spot? How much variance in overclock from board to board? How much does smt impact headroom? Not expecting answers but I'm just saying that I haven't seen this explored in depth

>>59605319
the 1700 is worth it, the 1700x and 1800x are just for the people not willing to bet on the silicon lottery, if you wand a almost gurrented 4-4.2 ghz get a 1700x or 1800x

>>59605374
>>59605319

I could see why someone would get the 1700x.

on the 1700, all cores run at a max of 3200 with I'm not sure how many reaching 3700 (and having an XFR boost of 3750) the 1700x has all cores reach at least 3.4 with whatever number of cores reaching 3.8 with XFR of 3.9
and 1800x has them all run at least 3.6 and reach 4.0 (4.1 with XFR)

I meant is it worth turning off cores for headroom not is it worth buying
>>59605319
>>59605374
>>59606102

>>59606128
In my case it would be, the first thing I'm running into is a heat problem. I intend to buy water cooling because I don't feel like gimping my hardware just to get manageable temps.

>>59604190
Yeah it being bottlenecked by the GPU is why better CPU MEMORY SPEEDS make it faster, and why the 7700k has shit min frames in half the games.

Kill yourself, retarded shill.

Guys what mobo do you recommend for R7 1700, that is around £150?

>>59603182
>Intel has huge framerate dips
Enjoy your St-utter

>actually getting 3600mhz stable

Credit to him for doing so but not many people are having such luck.

>>59607413
Use a decent motherboard and it isn't an issue. GSkill's F4-3600C15D works fine.

>>59603063
>benched with a GTX 1070


HAHAHHAHAHAA THIS POOR "REVIEWER"

>>59607446

My ram is on the QVL and on the crosshair hero I really struggle to keep it stable over 2666mhz.

File: Ryzen-Rise-of-the-Tomb-Raider-1080p-Very-High-Preset.png (22KB, 800x600px) Image search: [Google]

22KB, 800x600px

>>59607482
Just going to go ahead and leave this right here

>a 4.5 year old game is the most used and respected benchmark for PC gaymin

pc gaming is dead

File: Screenshot_20170328-030659.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

>>59607503
Nice try.

>>59607491
ASUS motherboards are junk for AM4

>>59607386
>misinterpreting the post this hard

>>59607548
>no source, doesn't specify anything besides ram speed and processor.

Oh, okay. I'll just take this and toss it into the trash where it belongs

I hope the four core will be good enough to upgrade from fx-4300

few weeks to go

>>59607595
>lying on /g/
Stop it

>>59607609
>fx
It gets shit on by ryzen.

>>59607600
It's digitalfoundry, Mr. Rakeesh.

>>59607619
nigga you dumb, here let me clarify that post
>[1700x] [a]ppears better than [7700k] with the same 3200mhz memory, OP.
>as opposed to OP saying that it was equivalent with 3600mhz memory instead
then he want on to make the same point you were about the dips being the more important metric.

>>59607566
Tell me about it. Sitting here with a Prime B350 Plus.
Just updated the newest BIOS 0513, overclocking is still nonexistent. You cannot set base clock, or multiplier on this board. It is either locked to auto, or it is locked to normal base clock, and a multiplier of 33. It'll run at a fixed frequency of 3.3ghz, I have no other control than that. It'll show fields for certain settings, but you cannot alter them whatsoever, some of them display readings of 0 which is impossible.
Apart from the Ryzen Master tool there is no control over clock speeds whatsoever. On a board that is supposed to support overclocking.

ASUS is so absolutely shit I can't even fully articulate it.

File: 15b.jpg (700KB, 7680x4320px) Image search: [Google]

700KB, 7680x4320px

>>59607656
>nigga
nothing you said is even remotely relevant to the matter at hand
fuck off pham
>>59607623
>digitalfoundry
>still doesn't link to the specific page
HAHAHAHAHAHAHAHAHAHA, NICE ONE!

>>59607700

See

>>59604143
It was already linked but it's probably hard for you to scroll up on your Athlon x2.

>>59607269
b350 tomahawk

File: 1489469160974.jpg (199KB, 600x610px) Image search: [Google]

199KB, 600x610px

>>59607734
>still linking digital foundry
Okiedokie mister, that's enough Jewish propaganda for you today

>>59607828
He's clearly protestant, you islam.

>>59603063
you can't even afford 3600mhz memory

>>59607269
Asrock b350 pro4

>>59607548
do you even read your own graphs?
ryzen 78/77
i7 114/117 and a dip right on your screenshot
also this >>59603353

>>59607843
>protestant
Stop protesting this schlomo

>>59607886
Slow day at the tech support scam hotline today?

God damn it, AMD. You just HAD to turn. high end ram into not a meme. I don't have a fucking APU, I shouldn't have to buy 3600 RAM for decent gaming performance.
Before you all sperg out, I have a 380x and am considering a 480.

>>59607922
Slow day Jewing people Krzanich?

>>59607656
Yes, this person summed up what I meant.

Not that it should need summing up. It seems perfectly clear.

>>59603063
why would they make just dual channel for ryzen, if it's that memory sensible? wouldn't quadchannel prevent, that 2133/2666MHZ suck so much and make it faster than dualchannel 3600MHZ ram? that's retarded

>>59608276
>I don't know anything
>I'll express a retarded opinion anyway

The system isn't starved for DRAM bandwidth. The data fabric is set to a 2:1 ration with RAM speed, faster memory equates to faster data fabric. Faster data fabric means lower latency and higher bandwidth between CCXs.

Adding more memory channels would be completely useless.

File: 14884382954907.png (84KB, 600x587px) Image search: [Google]

84KB, 600x587px

>>59608331

File: 3600c14.jpg (442KB, 1920x1080px) Image search: [Google]

442KB, 1920x1080px

>>59603063

Artificially reducing Skylake's performance with 3200 14-14-14-34 2T vs Ryzen's 3600 16-16-16-36 1T.
http://imagizer.imageshack.us/a/img922/7465/vqLZz0.png

File: IMG_2802.png (538KB, 1279x720px) Image search: [Google]

538KB, 1279x720px

Poojet's upcoming cpoo won't get any faster cores, just fewer just as shitty cores.

File: Shadow President.jpg (62KB, 569x762px) Image search: [Google]

62KB, 569x762px

>>59603120
>not having a 32 core Xeon for the price of an i7
>not ENJOYING the vPRO botnet.
If you're going to get VPRO.COMed, might as well get the one with more than 4 cores.

Wish AMD had a cheap CPU with so many cores as cheap as used XEONs.

>>59608422
Kill yourself.
https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/

>>59608582

>5.4ghz 7700k

That has to be be upper percentage for overclocking, even for kabylake.

>>59608814
thanks buddy, will read later

>>59608600
>trusting farcry primal, A literal rehash of a previous farcry game cept with cavemen n shieet
If they couldn't be bothered to make a new map do you really think they bothered to code it properly

Plus those benches are literally that of a shill you phoneposter

ITT: Benchmarks that doesn't favor Ryzen, is not counting

Some of these 'reviewers' don't even do a fresh instal of Win10 and games. It's already been shown that some games hold over their settings from the previous install(s). Like if they ran on an Intel chip and them switched to a AMD chip it would perform worse because the settings for the game are still thinking a Intel is installed and/or pulled the settings from online. Witness F1 for instance.

>>59607862
I can only find an AB350. Is it the same thing, or is it something inbetween a B350 and an A320?

If it's the latter, then >>59607799 would probably be better.