Ryzen

And I know that this is a drop in the flood of Ryzen threads, but your feedback would be most appreciated.

>>59211784
i7 6900K has poor gaming performance too compared to i7 6700K and i7 7700K. There is a reason nobody recommended it for gaming, and that reason isn't only the high price. Same applies here, if you're a poorfag who only plays games then why would you buy a processor with 8 cores? Those only help in modern games with Vulkan and DX12 if they are properly optimized, and that isn't most games.

6850k should still be listed simply because of the PCIe lanes.

Ryzen is maxed out at 32 (16 from chipset + 16 from the CPU).

6850k has 40.

>>59211851
Who actually needs that many PCIe lanes though, AMD did it right imo. They may have marketed Ryzen to be the top dog but its not and they know it. They are looking for between average joe and enthusiast users, which will never use more then 20 PCIe lanes imo, 2 8x cards (or one 16x) and a m.2 ssd.

>>59211833
feedback about what, your diary entry?

>>59211851
I did not omit the 6800K because of its PCIe lanes. With PCIe 3.0, x8/x8 is fine.

>>59211784
I think that the performance of the games will improve once the games actually make use of the cores. It's really an unfair benchmark when the hardware is not being used to its potential.

On an unrelated note, please do consider releasing your JavaScript under a FOSS license, Falcon. I've tried viewing your website, but sadly, GNU IceCat blocks the loading of your nonfree JavaScript.

<https://www.gnu.org/philosophy/javascript-trap.en.html>

>>59211784

1700x - Worth a read:

http://www.hardocp.com/article/2017/03/02/amd_ryzen_1700x_cpu_review/6

>First and foremost, for most HardOCP readers, you are going to find that Ryzen simply has a huge clock deficit to overcome compared to Intel. If you are looking to build a system for desktop gaming alone, there is simply no way to suggest that Ryzen is your CPU as the Intel 7600K and 7700K still hold a great advantage especially if you are overclocking.

if it's only gaming don't do it. otherwise it's pretty reasonable with good TDP.

I also worry the 1800X really isn't meant for what it's being compared to (7700k 7600k),but we don't have i3/i5 competition yet so that's my speculation and speculation and those comparisons is all we got).

>>59211833
Drop the 1700X and 1800X, they make zero sense as the chips are mostly binned the same and auto-overclocking utilities on the motherboard will push a 1700 up to 1800X spec at the same voltage anyways. At which point the 1700 should be an option in exceptional tier as well. Use case will be a very important factor between either the 7700k or 1700, but I doubt you'll ever dive into that since your charts are and always have been shit.

>>59211851
Ryzen is 8 for the X370 chipset and 24 from the CPU itself. But I agree with that notion, some use cases do require 40 PCI-E lanes.

>>59211784
How about >>59211647 ?
It seems that the motherboard brand affects performance by a lot.

>>59211887
>Who actually needs that many PCIe lanes though
anyone with lots of storage, anyone with multiple GPUs AND storage. Anyone with a 10Gbe network card and storage, etc.

>>59211930
>I did not omit the 6800K because of its PCIe lanes
6800k has 28 PCIe lanes, so it's not really relevant to the 40 PCIe lanes given by the 6850k.


If you need lots of I/O, your only option is a xeon or HEDT intel.

>>59211903
Is this your first day on /g/, newfag?

>>59211936
>I think that the performance of the games will improve once the games actually make use of the cores.
That may be true. With new BIOSes that overcome the initial teething problems, Ryzen may prove to be a better alternative... in the future. But until that future becomes reality, I can only recommend what is good now.

The current code that runs the website is a bit shameful. I would want to tidy it up a lot before making it FOSS.

>>59211887
2 x 16 lanes
1 x 4
Already maxed out? jesus...

>>59211851
I know and it makes me sad. Still watching for skylake-x / x299 for this. hoping for some PLX on top of that.

Any ETA for the R5?

>>59211973
>Who actually needs that many PCIe lanes though
anyone with lots of storage, anyone with multiple GPUs AND storage. Anyone with a 10Gbe network card and storage, etc.
My point was thats a super small portion of the market.
>>59211981
>Because 16x PCIe 3.0 is so necessary over 8.0
Most users will have one GPU and one M.2 SSD.
You could run 2 8x GPU's and still have plenty of bandwidth left.

>>59211980
Very understandable, until game developers can successfully make use of the cores, and the BIOSes are improved, it is only fair to recommend the Intel chip.

>>59211943
Will need to collect all these reviews and read them a second time. Thanks!

>>59211903
he is part of the people who run logical increments dot com. which sprouted out from /g/. in my opinion it is the best PC building site.

>>59212018
>Most users will have one GPU and one M.2 SSD.
"most" users. yeah, because a high performance non server cpu is targeted at "most" users. Its the Ryzen 5 marketed at most users, it just has a later release obviously for brand image.

>>59212073
With the lower cost there will be much higher market share in the higher market imo.
Seeing as Ryzen will sell well (imo) and intels cutting prices.

>>59211784
Is there a good reason to place a more expensive cpu to a lower tier? $350 vs $330 I mean.
Or do the prices need an update?

>>59212089
thats not even relevant to the discussion

>>59211784
>>59211833

nope, everything looks good.

>>59212126
Howso, their marketing is affordable HEDT?

>>59211784
I wonder how the Zen architecture will perform in the server market, that's a huge market that's pretty much 99% Intel right now.

>>59212051
i liked the Phoenix Linux gaming reviews as well

>>59212093
7700K should go down to $330 soon. It is $340 on Amazon... and I should update that! Thanks!

Thanks!

>>59211784
if you cared for pure gaming perf you obviously didnt know a lot of tech only idiots and shills started to hype it to 7700k levels creating stupid hype for no reason

if you cared for decent gaming perf WHILE being able to multitask at the same time then you are in for a treat

but it has a lot of problems with the motherboards and the os's till they get ironed out we will see some wonky stuff

>>59212155
were talking about how a typical and non typical customer uses their cpu, cost and marketshare have nothing to do with this

>>59212275
My point is more of your normal typical customers are going to get into the higher market since its cheaper on the chips and the boards. Alot of people wanted to get into the highend market but didn't want to pay that crazy intel tax.
And those typical customers wont need anything past 20x

>>59212233
AMD themselves did not hype it to 7700K levels in gaming. But they did compare it to the 6900 stuff. I was fooled by the fact that AMD's benches were mostly 1440p and 4K.

File: 5yaiddpbu1jy.png (900KB, 4000x2000px) Image search: [Google]

900KB, 4000x2000px

>>59212324
why?

look if we go down the road of the shills and compare it to the 7700k then yes

but in reality the 1800x was competing with the 6900k
the 1700x with the 6850 6800
and the 1700 with the 6700...

now afaik the 1800x is 13 %(on average) and 8% behind 6900 on mins and then 7700k is 10% behind 1800x on mins..... and that is on a system with a lot of problems that suprisingly its not even amd's fault lol (apparently only the taichi asrock mobo was the one giving a decent perf the rest had a lot of bios problems)
the 1700x is only 4 to 5% behind and the 1700 is on par...

as of the rest well the results speaks for its self it beats them in everything

there is a lot of problems to be fixed from both sides..

mobos with their horrendous bios and their memory latency
windows and linux with their smt patch and the L3 patch (also smt patch for linux on 5.10 and a non important sound driver for the new sound chip )

File: ultimate_gentoo_cpu.png (52KB, 609x765px) Image search: [Google]

52KB, 609x765px

>>59211784
Windows gayming is one thing. Linux/BSD workstation is another.

http://www.phoronix.com/scan.php?page=article&item=ryzen-1800x-linux&num=5

>>59211887
>Who actually needs that many PCIe lanes though
Anyone who actually has a use for 8 cores?

>>59211981
1 x 16 is maxed out. The 6850K has 40 lanes from the package, not using a PCI switch.

I just bought a 7700k that my local retailer was pricing $20 below the 1700 today after reading the review.
I only ever encode video once in a blue moon so the only real thing I use my pc for is gaming.
I sure hope AMD keeps improving on Ryzen and 3 years from now when I buy a new CPU that might be AMD.

File: t3_26s7g1.gif (283KB, 652x349px) Image search: [Google]

283KB, 652x349px

>>59211784
NICE TRY INTEL

It seems to me that the Ryzen 8-cores are only really suitable for the server market and not the desktop. You could build a workstation with it but with the motherboard teething issues and the limited I/O, why would you unless budget was your primary consideration.

My guess is they led off with the 8/16 to sell to fanboys and harvest what premium they could before they have to introduce the 4/8 and 6/12 to what will be a value-minded market. Now that the cat is out of the bag, they are going to have to price the lower bins very competitively to draw any interest from the gaming community.

I have been interested in the 4/8 all along so today wasn't too much of a letdown. I think /g/ forgot the MOAR CORES meme for a bit there and got overexcited with the numbers against intel's 8/16, which no one uses for gaming anyways. AMD did claim 40+% IPC gains and they surpassed that but somewhere in the past few months, /g/ forgot that this projected somewhere around Haswell's IPC. Couple that reality with Ryzen's low clocks and here we are--a really good architecture improvement that just doesn't clock where Skylake and Kaby Lake can.

>>59213257

They also pretty much market it to streamer back then in December.

>>59211784
BIOS updates you fucking faggot.

>>59213257
>It seems to me that the Ryzen 8-cores are only really suitable for the server market
Yea, that's probably because you don't know anything about the enterprise market.

>>59212675
So their only potential customers are OS developers? Good luck with that, Another Massive Disappointment.

>>59213523

True, or at least I don't claim expertise as you apparently do. Do you foresee building a lot of these for enterprise? Why so?

>>59213634
>only OS devs compile code

hehehee stupid goy, buy intel

>>59211784
>"Hur dur 7700k is king of 720P gaming 1080p and higher is irrelevant because gpu bottlenecks that's why I play at 800x600 plus human eye can't see more than 4 cores anyways".

>>59213692
What? Are you retarded? Ryzen won't be seen anywhere near the enterprise market. Maybe the new Opterons could actually give some competition to Intel in that market, but Ryzen died when it didn't have 40 PCI lanes and no quad channel memory.

File: gnu linux and windows - PC.webm (2MB, 640x360px) Image search: [Google]

2MB, 640x360px

>>59211784
I expected Haswell/Broadwell perf. I got exactly that.
I still don't see any need to upgrade until Zen+, my haswell refresh will serve me until then.

I believe that the title of the review fits nicely

>AMD Ryzen R7 1800X Review: An i5 in Gaming, i7 in Production

and for ages we all said that i5 is enough, and only multithreading is worth the extra money for i7

and ryzen got that covered with future games

>>59213793

Oh, so you meant server when you said enterprise.
Fair enough. I should have said Zen core instead of Ryzen.

when will you have an updated laptop guide?

>>59213849
I meant literally any enterprise use. Maybe low end desktops will see APUs in them, but something that actually needs 4+ cores in enterprise is going to have Intel stuff at least until AMD figures out they shouldn't bottleneck themselves.

File: autism.jpg (38KB, 371x480px) Image search: [Google]

38KB, 371x480px

>>59213804
>wearing autism sweater
>moves windows around spastically
>plays some shitty ubi gayme on one monitor in a VM

oh yeah behold the power. doing something any modern system can do

>>59213872

Okay, so we agree then. Not sure why you jumped down my throat.
>Ryzen not suited for desktop
>maybe cheap workstation

>>59213823
people are going to be shocked once the R5 and R3 show up to chew up the i5's and i3's

>>59213901
I think you're confusing what a workstation is. It's not simply the shitbox you sit your secretary at which could employ Ryzen, but won't. It's the machine you sit your engineers at and have them make you money. Ryzen will never be there. Doesn't have 40 lanes, quad channel IMC, 2S scalability.

>>59211784
I think you should give out some more free games

reading a few more reviews I want to know a few things besides fps

frametime graph at different resolutions, 8 core leverages that apparently well enough compared to 4c. Basically I want full videocard review treatment.
second thing is test with AMD card on amd drivers compared to intel cpus, i'm not convinced that nvidia drivers just not ready for new arch more than likely and happened before

select few graphs show that ryzen has much less frame jumps, it a thing i'd consider future proofing more than max fps

>>59213922
The Ryzen chips hit the market for 'youtuber content creator' spot on. Those guys who edit video to make their stupid '10 ways to serve apples' vlogs or whatever would get some serious punch out of a high-end Ryzen.
Or anyone who needs the power of a $1600 cpu for $500

>>59213944
also I want to see more worse running games like dishonored 2 which showed improvement in fps at 1440p against 7700K.

this shit is way too complicated, and nobody had time to make complete research.

>>59212675
>A shitty APU beats the 7700k by 300%
top fucking kek jews blown the fucking fuck out

>>59213922

No, I get that, which is why I pointed out the limited I/O when I said maybe a cheap workstation--after all, it is half the cost of the 6900X. I wouldn't do it if I had the budget but the possibility exists. >>59213959 posits a good use case.

>>59213441
How are we to know what a BIOS update will bring? When it happens, we will know. Before then, we can only go by what we have.

>>59213983
Yea, the YouTube video maker is probably the only person who could actually use Ryzen to do anything.

>>59213983
It ms definitely a solid use case.

I just wanted to make it clear that real businesses aren't touching this. Maybe Zen Opterons will be interesting, though. I kind of hope they are. Xeon pricing is out of control.

>>59213942
Will do that after I update the full guide. This is just a Ryzen update.

The big elephant in the room is that the R7 is meant to kill the "Extreme Edition" platform.

While people touted the 7700k as a good processor, it does it at the cost of the "prestige" or the LGA2011 platform. that already caused a shift in intel's plans.

The fact that Ryzen is outpacing the flagship skylake i7's is good enough, even at how green this platform is.

I think people are overestimating this "blunder" and not thinking ahead to what the actual mainstream chips will be.

>>59213990
Hopefully the R3s and R5s fill out the lower CPU ranks with some competition.
Maybe some aggressive binning will squeeze some good clockspeeds out of the R3s- considering that they have half their cores disabled they could prove to have much better cores.

>>59213959
Streamers as well. And gamers who use their PCs for productivity.

>>59214012
>Maybe Zen Opterons will be interesting, though.
It sure looks like it on paper. The performance is there and the efficiency looks very good. I wonder why they didn't lead off in that market.

>>59214051
"""""""""Productivity"""""""""" needs 8 cores?

>>59214054
Su says it's testing not ready, server market is much more complicated, they dropped this with unfinished bios, and memory profiles

>>59214054
It's going to be about another year before Zen Opterons come out. Even Intel is always a year behind.

>>59214093
naples in q2-3
market branding unknown

>>59214041
Right. Hopefully, with half the cores disabled, AMD can get much better speeds than the R7 stuff.

>>59214059
If the program you are using benefits from more threads, yes.
>anyone who needs the power of a $1600 cpu for $500

>>59211784
What about ryzen if you are not a gamer or something like that? Regular midrange PC for occasional gaming and mostly browsing and coding.

>>59214123
With $330 for the cheapest current Ryzen CPU, no one looking for something "mid range" will be going for it. After the R5 and R3 come out, perhaps.

>>59211851
I don't think anyone who knows they need more than 32 PCI-E lanes is going to use logical increments.

>>59213967
You are reading the graph wrong, lower is better.

>>59212006

~3 months. probably around the same time as vega launch.

>>59213257
>It seems to me that the Ryzen 8-cores are only really suitable for the server market and not the desktop.

Zen is DOA in the server market. AMD has a product that only competes in terms of price/perf and not perf/w.

>>59214286
what are you on about? it draws less than intel 4 cores not to mention 30% difference in draw from intel 8 cores.

>>59214286
How so? In everything but gaming it has better perf/w than Broadwell E

>>59213001
You know like 9001 of us have that as a check box option. Can even add in the (You) if it p[leases us/interests us enough that we want to see the response easier.

File: power.png (235KB, 602x685px) Image search: [Google]

235KB, 602x685px

>>59214294
>>59214297

?

if you draw the same amount of power but perform worse then you have lower perf/w.

an argument can be made that it's not total dogshit like bulldozer but not having perf/w that at least matches or beats intel makes zen an unappealing product in the server/enterprise market.

>>59214286
Naples ? Since when amd says zen is for server too ?

>>59214323
it beats stock 6900k and draws less pwoer in worklaods = higher perf/w
OC argument doesn't work in server enviroment

>>59214495
>it beats stock 6900k and draws less pwoer in worklaods

except the 1800x doesn't beat the 6900k

>OC argument doesn't work in server enviroment

nobody mentioned overclocking

>>59214157
I'm still waiting doe those two. A new build is not that urgent right now, I'll wait until summer I guess. At this time there should have been some price cuts already and some long term reviews.

>>59211784
The reason to suggest Ryzen is if you need a CPU that is not only for games, meaning streaming, content creation.
Problem is that this will only apply to a minority, most people are better of with an i5 or i7.
Fuck AMD for not releasing R5 and R3, it's their own fucking fault if they now crash and burn.

>>59215778
This is why it fits into only one tier, for now.

>>59215778
>unlimited hype train with no brakes
>whole bunch of neets pre-ordering before benchmarks are even out

Why wouldn't they cash in on that hype with their most expensive part?

>>59211784
falcon please change your website on mobile, the navbar takes up half the screen

>>59214119
I can't think of a single application the average person uses that can even utilize 8 cores, never mind actually be accelerated by them.

IF YOU WANT TO GAME ON RYZEN WAIT FOR R3 AND R5 YOU TARDS

>>59218519
alternatively, go for the r7 1700
1700 at 3.9ghz seems to be neck and neck with 7700k at at 4.5ghz. If you won the silicon lottery, you're golden. If it weren't for the limited OC ability, this chip would have demolished the 7700k.

>>59211784

I'd switch the 6850 and the 6900 as well. Put the 1800X only in the most expensive build, 1700 and 1700x in the rest

>>59211784
you realize TWO rx480s cost far cheaper and even out perform the 1080 right?

File: 1484605260603.gif (1MB, 326x300px) Image search: [Google]

1MB, 326x300px

>>59211784
I've always liked logical increments (and consulted it many times) but I cannot help but feel that you guys place too much of an emphasis on gaming.

I understand that it is first and foremost a website for gaming pc builds, however that does not mean that gaming should be the sole concern to the exclusion of all others.

The frame deficit in Ryzen can be as low as 1-2fps and as great as 15, yet in all other computational tasks it excels. While you might not get the "objectively best" gaming experience, you gain the ability to run more programs simultaneously, steam, and a slew of other things which are more difficult on intel machines.

To be frank, I think that the controversy over gaming benchmarks detracts from meaningful conversation about the processor and that while they are important, they are not everything.

>>59218635
I agree. Nobody owns a high-end PC just to play games on it and plenty of them have multiple tabs open with youtube, facebook, etc while gaming, they even watch streams of their favorite online game on a second monitor while gaming. Additionally, these benchmarks are performed with nothing running on the system except for the game. No crap installed with background processes and so on. And these are not realistic scenarios for most people.
Blindly following reviewers advice that i5/i7 is a better purchase for gamers at this moment is foolish in my opinion.

>>59214495
Now Ryzen is targeted for server environments HAHAHAAH

>>59218818
Naples is. The 1700 exists to show that the architecture can be used for servers

>>59218770
>B-B-BUT THERE ARE OTHER USE CASES FOR RYZEN
Yea right, that's damage control arguments on day 2.

>>59218858
Calm down. It's the same "for gamers" use-case. I just described what I think are more realistic usage scenarios.

Don't even have plans to purchase Ryzen. I'm fine with my cheapo Xeons.

>>59218635
The usage cases where 1800X is a compelling buy are much slimmer than people make it out to be. Rendering in graphics workstations is significantly done via GPU acceleration. You don't need an 1800X for that.

The CPU is good exclusively in software/CPU work, such as compiling Linux kernels. Which is fine. But people walk around thinking rendering is done on the CPU. And it kinda is, but mostly isn't. Though I assume the people who actually do this work are informed enough to know the difference.

I feel you need to double the RAM on exceptional and upwards but then again I do use a lot of RAM. Otherwise looks good.

>>59211784
to be honest the gaming perf is just ok its not like you gonna loose any fps on any game

the platform is new it will need some ironing out from ms and for bios we really dont know how good can it be (well we do since the gigabyte motherboards are the ones that actually got it right it seems)

for what it is now its really really nice (the 1700 and x the 1800x is just what the 6900k is a dick waving chip)i just hope that microsoft will fix the L3 and the SMT quickly and not after 8 months like they did with intel HT

also in linux SMT doesnt work yet (only on 4.10) and the sound chip doesnt have yet a driver(its very new)

>>59219432
>also in linux SMT doesnt work
anything post 4.8 works, maybe even 4.4. 4.10 and the latest 4.9 do however have fixes to make the SMT thread juggling work correctly
>yet (only on 4.10)
which any distro worth using already has
>the sound chip doesnt have yet a driver(its very new)
sound chip on the cpu? some mobos may have some newer shit but that's not the case for all the mobos

well at least its official now
http://support.amd.com/en-us/kb-articles/Pages/am4-chipset-driver.aspx

ryzen supports windows 7

>>59211784
You do realize most of the Ryzen hate threads are just nvidia fanboys reaching for ways to discredit the better bang for buck, right?

>>59213767
>all code compiles the same

stupid kid, let adults talk pls kys

>>59214039
People are just in denial.
It's a new model. Of course it will have some compatibility issues with current games.

>>59220852
Based.

Windows 7 is going to live forever.

>>59220852
That would be great news, but considering MS has to fix the SMT, I'm sceptical about them doing it for W7.

If you want a ryzen chip for gaming, you buy the 1600x for the best price:performance on the market, or you buy the 7700k because you have more money than sense.

If you want multithreaded performance on top of gaming performance and still cheaper than the 7700k, you buy the 1700.

If you want to go maximize your possible multicore performance and are willing to pay a premium for it, you buy the 1800x.

The only competitive chip intel has is the 7700k, and it's a 4-core has-been. Even the 1600x is going to have better multithreading performance.

>>59215778
They're pairing the R5 up with vega's launch.

R7 comes first, they iron out as many performance bugs in 3 months as they can.

Then R5 releases alongside vega - it'll be around 1800x gaming performance (but not multithreaded performance) for ~$270 vs ~$350 for the 7700k. e.g. 10% difference in performance, 30% lower price, AND better multithreading on top. Throw in the cheaper AM4 motherboards and you'll be looking at 40% savings going AMD on motherboard+cpu vs Intel, for very nearly the same performance.

And then there'll be ryzen+vega bundles to cut the price down a little further compared to going intel+nvidia.

Basically, if you want "good enough" and not "the best" for gaming, where your $1000 AMD PC is performing 10% worse than a $1400 Intel/nvidia PC; you'll want to wait until may.

This is basically how it's always been with AMD.

>>59211887
>Who actually needs that many PCIe lanes though

I do, GPGPU is very much bandwidth limited.

I also understand why AMD didn't put that many lanes though, it doesn't matter to most people.