>Ryzen 5 series performance is literally half of competing

>>59775957
>spic
no thanks

>>59775957
I don't think we're looking at the same chart.

>>59775957
it looks really nice considering it costs half as much

File: dt_131216_cherry_picking_fruit_250x188.jpg (36KB, 250x166px) Image search: [Google]

36KB, 250x166px

>>59775957
These cherries sure are delicious, mmm. Now, what were you saying?

>>59775957
Are you even trying? You seem lazy.

File: bench.png (514KB, 500x498px) Image search: [Google]

514KB, 500x498px

>>59775957
Hey, you cherrypicked the wrong chart. Let me help you with that.

>>59776151
Total Shill: Intelhammer is my favorite game.

File: IMG_20170406_133612.jpg (129KB, 1024x768px) Image search: [Google]

129KB, 1024x768px

>>59775957
>Battlefield
>Doom
>Metro Redux
>Rise of Tomb Raider
>Total War Warhammer
>>59776151
>Total War Warhammer

Okay, but now, where are the benchmarks if more serious applications? Like, Blender, Mathematica, PovRay, h.265 transcoding...

>>59776284
Can Ryzen render bewbs that big?

>>59776284
There's some benchmarks out from Puget Systems, but they are pretty old. Solidworks and Lightroom results were really poor on Ryzen systems so it may suggest a rebuild is necessary for this platform and it might be significantly affected by special optimizations.

File: 1491162294224.png (49KB, 754x939px) Image search: [Google]

49KB, 754x939px

>>59776151

Looks pretty good.
What's the problem?

>>59777171
hes an intel shill

>Novidya drivers

Discarded

>>59775957
>ROTR
>1600 gets 20 fps more

>>59775957
its compeating against in i7

Chapuzas?

File: 1473465786841.jpg (765KB, 2560x1440px) Image search: [Google]

765KB, 2560x1440px

>>59775957
at 3.6 with 2400 RAM

>>59777632
it costs less than i5

>>59775957
>literally half of competing Intel CPU

Oops, you forgot to post the chart showing this.

File: 1466029086789.jpg (385KB, 2560x1440px) Image search: [Google]

385KB, 2560x1440px

>Nvidia the way it's meant to be gimped!

File: proxy27.jpg (235KB, 981x1472px) Image search: [Google]

235KB, 981x1472px

>>59775957
why are they comparing a 1600 to 6700k?
the 1600 is over 100$ cheaper
almost 150 cheaper if you include the free cooler

what is this nigger shit?

i just want a reviewer to compare CPUs that are the same price? is that so fucking hard?

>>59777967
Really makes you think.

So >>59776151 is just a dumb shill faggot?

>>59777994
>why are they comparing a 1600 to 6700k?
>the 1600 is over 100$ cheaper
No, it's over $250 cheaper.

6700k+z270 motherboard+decent cooler is about $550.
1600+motherboard = $290.

>>59775957
>when crying intelshills lose touch with reality so bad they post images that don't even remotely support the 'points' they try to make

>>59775957
Is this the site that used 4GHz RAM for the Intel processors and 2GHz RAM for the AMD ones?

File: whycontainit.jpg (39KB, 324x291px) Image search: [Google]

39KB, 324x291px

How many times are we going to go through the whole "AMD releases line of products, lackluster launch, drivers improve, it BTFOs the competition" until people get it?

Ryzen just needs better drivers and when devs start taking advantage of Ryzen's advantages, it'll be GOAT (just like their GPUs)

>>59779222

Until AMD gets its shit together and launches a goat product from the get go?

Instead of having to wait for the performance increase that inevitably comes?

AYYMD RYPOO HOUSEFIRES

File: reddit.jpg (73KB, 620x465px) Image search: [Google]

73KB, 620x465px

>>59779254
But that'll never happen because they invest in future technologies

AMD Fine Wine technology is real for a reason.

>>59775957
fuck yes broheim, 1600 confirmed for 60fps 1080p ultra
>>59777967
works better on red cards
>RX580 1500mhz $200
75hz Freesync monitors are cheap
this is looking good

>>59779283

Then we will go through this every fucking time until AMD gets a fucking clue that building for tomorrow today doesn't meant shit to a customer who just wants the best right now.

>>59775957
Why does the R5 get such a massive increase over the R7 in tomb raider?

>>59779318

Because Tomb Raider is fucking broken since Nvidia got involved.

For gaming it really feels like the 1600(x) is the one to get.

Glad i waited.

>>59775957
How lame, they didn't even overclock the i7 to 5ghz or more. how is that fair.

>>59778010
you know some dev are not lazy and willing to optimize their game right?

>>59779170
>Is this the site that used 4GHz RAM for the Intel processors and 2GHz RAM for the AMD ones?
Pretty much.
3466 for Intel, and 2400 for RAM, iirc.

>>59779318
Because they didn't retest the 1700, they just used past results, for some of them. For that game it's just because the engine is retarded and scheduling threads wrong.
I'm sure there is lots more fucked up with their methodology, BIOS, Windows settings, etc., as well.

>>59779318
R7 is bottlenecked by AMD retarded infinity fabric design

File: 1490504795973.jpg (48KB, 438x436px) Image search: [Google]

48KB, 438x436px

>>59779313
Fair enough but for anyone that knows better and is faced between a 7700K or a 1700X, just go with the 1700X, we know how it'll end.

>>59776209
I literally got it for free with a fx8350

>>59779363

How's your fourth generation Skylake?

>>59775957

>Competing Intel CPU
>CPU that costs 50% more and have an average of 10% higher performance in games and 20% lower performance in multithreaded apps.

>>59779383

Selling promises is never a good business strategy. Sell results.

>>59779422

The i5 is going to be interesting. It will have higher maximums, but lower minimums and the averages will be in favor of the 1600 because of it.

>>59779350
Does that mean this test is more unfair to the i7 than any of the others because of optimisations made for it since release?

Fucking shills

>>59779433
Should average FPS in gaymes even be shown?

0.1% minimums, 1% minimums, and 10% minimums should be the only numbers anyone cares about.

It's meaningless if you get 250 max FPS when it still keeps dropping to 80 when you have a 144hz monitor.
It's meaningless getting 90 average when it drops under 60.

It looks like the 1600, especially with an OC to 3.8-3.9ghz, is not going to drop under 60 in just about any game when a 7600k OC'd to 4.8 or even 5 does on tons of them.

File: tmp_13980-Mount-stupid-–-Borgerlyst1754851986.jpg (23KB, 600x338px) Image search: [Google]

23KB, 600x338px

>>59777171

Either Intel shill, or at the peak of mount stupid

>>59779363
oy vey let's pay $1000 for 8 cores instead of $300
r5 also uses infinity fabric with 3+3 configuration

>>59775957
only shows that TW warhammer is apparently a huge pile of shit with that huge a discrepancy.

>>59777967
>blue: avg fps
>blue: min fps
ok

>>59779363
oiiiiiiiiiii veeeeeeeeeey

>>59780735
Shit, I just noticed this. He better delete that fucking video now.

>>59779383
The problem with selling promises, and being AMD, is the whole over a decade of failing to deliver thing m8.
Personally I'm waiting a month to decide on my build to give time for shit to come out, and to be improved. However not every buyer is willing to do that.

>>59779363
>R7 is bottlenecked by AMD retarded infinity fabric design

The R5 uses the IF crossbar too, dumbass. It's the exact same dual-CCX setup, just with one core disabled in each.

>>59777994
They shouldn't.
They should compare 1400X with 6700K
So it's on equal terms, 4C8T vs 4C8T

>>59775957
>gaymes are the only measure of peformance
who cares, any NEET with that much time to play games will buy a G4560 because it's all their saved allowance can afford

>>59781093
I'm going to downgrade my i7-2600K to a G4560 just to spite you.

>>59781033
1700 should be compared to the 7700k because they're most similarly priced!

1400 should be compared to the 7700k because both are 4c/8t!

EYE FIVES ARE THE BEST AND NOTHING CAN BE COMPARED TO THEM.

>>59781219
do it

i spent my neethood a few years ago with a 3770K

shit was cheap back then because strayan dollars were worth 10% more than those of the burger

>>59777967
>>59780735
>graphs are actually green and red

kek

>>59775957
why the fuck doesn't having 2 cores more to compute any effect on performance? fucking gamedevelopers.

so the 1600X will perform like the 1800X and be faster than the 7700K? looks sweet desu

>>59777967
Are you retarded?

1060 min. fps is way higher than the overrated RX480

>>59775957
>Doom (OpenGL)
>terrible

Welp I was planning to buy a Ryzen for my OpenGL work but nevermind then.

File: 1491092790355.png (128KB, 302x480px) Image search: [Google]

128KB, 302x480px

>>59781527
>1800X faster than the 7700K

Go kys delusional faggot

>>59781597
that's how he got fat hehe?

>>59781547
You are supposed to compare the RX480 on the 1800X vs the GTX 1060 on the 1800X pleb. The RX480 on the 1800X is higher than the GTX1060 on the 1800X, showing that Nvidia is gimped in DX12.

>>59781617
You tech illiterates are really suprising me everytime.

You don't understand jackshit and try to cover it up by using buzz/meme words.
Why don't you atleast read 1 article on that matter instead of showing how dumb you are.

Goes for the rest of you guys too.

>>59781565
Sure you were....sure you were

>>59781656
>hurr durr
>you just don't understand!

>>59780735
>>59777967
Why is "minimum framerate" even a stat, assuming it's literally the lowest the framerate was measured at? Surely a more useful measure would be the average framerate of say, the highest 2% of the frametimes.

>>59781741
>hurr durr
>muuuh gimping

Anyone using that word should be permabanned from /g/ as that person obviously doesn't even possess superficial knowledge about core tech.

>>59779482
>0.1% minimums, 1% minimums, and 10% minimums should be the only numbers anyone cares about.

I agree these should have priority, but I would not discard average completely

>>59781597
>muh 25% single core
It gets shrekt in anything even remotely multithreaded and that's facts. Faggot. p.s. kys

>>59781764
people don't like hiccups, lag, etc.

>>59781899
a single slow frame isn't a "hiccup".

It does not explain if it's the lowest group of frames over a second, 0.1% minimums, 1% minimums, or if it's just what the framerate would be if you took the single slowest frametime.

Why do you even comment when you don't know what that guy is asking?

>>59781764
Because it doesn't matter if you have 100 average fps if you go down to 30 min fps in the most extensive moments

>>59781944
>>59781899
You missed the point. I'm saying that the 'minimum framerate' could easily be a statistical outlier. If it were say average of the bottom .1% or 1% or whatever, that'd give a much better picture.

>>59781995
You're right, but we are not the reviewers. So unless it becomes a demand, we are not gonna see this become standard

>>59777967
That's an AMD game dumbass. Nvidia isn't humping anything.

File: IMG_1128.jpg (21KB, 220x300px) Image search: [Google]

21KB, 220x300px

>>59782197
>Being a phoneposting moron
>Getting cucked by autoccorect
>Being too dumb to read a chart
>Or correctly infer it's meaning
How embarrassing for you.

>>59775957
>looking at amd 5

found the poorfag

>>59782354
>H-he's going to get a cheaper and better CPU than me
>Better call him a poorfag!

>>59782354
but everyone buy a 7600k your computer and no applications ever use over 4 threads!!!!

>>59781933
>a single slow frame isn't a "hiccup".
it's not a single frame, don't be stupid
it's average minimum fps, meaning how low it will go on average

ideal measure would be frametime graph 1hour long for cpu performance
because just measuring averages is retarded and doesn't give you time relation when what frame happens, it can go from 190 to 70(or from 70 to 40 and average would be ~60) every other frame which would make it unplayable but averages would look like it's got 120 fps

>>59782354
fun thing with R5s is that you get 6850K($500) for $250

>>59775957
How is the 1700X losing to the 1600?

>>59783670
Older benchmark before updates improved performance.

>>59777967
>>59776151
>>59775957
The nvidia single threaded driver + total warhammer first core usage is the main issue with the performance differences.

Total Warhammer (and many other) typically utilize first core/thread for the main task and the other threads as auxilliary systems.

Nvidia driver uses first thread/core as the driver.

What would fix this discrepancy is if nVidia can change the thread affinity to one that's used the least or change it to not use the default thread.

This would then bring both the performance in line.

Give it me straight.

will 1500x be any good?

Uhhh, isn't the r5 1600 $220 ?

>>59784585

Beat bang for the buck on the market.

>>59784593
yes

>people still testing BF1 DX12 on Nvidia cards

>>59784562
is it a cateter in the middle

>>59784597
>not spending that extra $40 for 2 more cores

>>59784672
Some people don't need the higher power draw for just gaymen.
Some people don't want to OC, and the 1500X boosts to 3.9Ghz stock.

It also has more L3 per core so it theoretically should have a higher IPC.
It's also better binned (seems confirmed from people getting 1400s and 1600s early doing OCs) so you can probably get another 100-200mhz OC out of it than the 1600 if you do OC.

As much as $30 more for 2 more cores is nice, someone that doesn't need it ought to not spend the money. Especially when they could just get a 2nd gen Ryzen Athlon next year that's a single CCX and can clock to 4.5ghz or so and just use the 1500X as something good enough for now.

>>59784742
If you have a functioning pc at all you should just wait for R2, don't be an early adoptard

>>59784631
>people still testing DX12 on Nvidia cards
Fixed.

>>59784812
>If you have a functioning pc at all you should just wait for R2, don't be an early adoptard
But this isn't like Intel where a socket only lasts 1.5 years.

I'm early adopting because I can get zen3 or zen4 further down the line.

For a new architecture, this has been a really painless loss. Biggest problem was some poor optimizations and faster memory not working for most on day 1.

>>59784812
rumor is R5 is a stepping
also can clearly see that 1600 at 3,6 beats 1700x at 3.8
8c is niche, cheap decent 6c though is closer to reality for most people

>>59784958
>rumor is R5 is a stepping
it's a bullshit rumor
>also can clearly see that 1600 at 3,6 beats 1700x at 3.8
because the 1700X result was from day1 and the 1600 results are from BIOS updates + windows updates + game updates.
tard.

>>59784950
really, 250 is nothing for 2 years of use
I would think this thrice over if it was intel "we change sockets because we can"
I can even see "zen4" on 7nm moving more of south bridge on CPU and still working on am4

>>59784998
>found a mexican

how do you know?

>>59785009
Right.
I mean I can afford a 1700 now.
Hell, I could even get the HEDT platform 12 core.
But this is clearly a server/laptop process it's on, even though the architecture is good, so I'd rather wait to drop more money later.
It's clear the 2nd gen will big significantly better for desktop, if not a lot better. And the 3rd gen pretty huge of an upgrade, as that's confirmed to be on IBM's 7nm process that clocks high, as well as being lower TDP.

A 1600X is plenty good enough and way, way better than the 2500k I have now out side of highly single threaded shit (which I don't use) with a big overclock.
Intel never gave an attractive option to upgrade. If I want 8 threads, I need a new motherboard, RAM. And then they'll come out with a new socket in 18 months.
Here, yeah I still need a new motherboard and RAM, but it's a much longer term platform, and I can get up to twice as many cores as well.

Intel still considering 6 cores as "high end desktop" that needs its own socket 2-3 years ago is a fucking joke.
Only recently since it was starting to turn out that Zen really looked to be good did they finally announce that the upcoming Coffeelake would bring 6core CPUs to a mainstream socket. Too little too late. Fuck Intel.

>>59785031
I know because tons of people have reported overclocks for the 1400 and 1600 and they're the same as the 1700 or worse.

It's very apparent that many 1400 and 1600 are not 1700s with 1-4 cores not working, but they're ones that weren't as stable at lower voltages. So since they didn't meet the TDP target for a 1700, they got cores disabled.
This is what I said weeks ago. And what do you know, someone that knows what they're talking about made a good prediction over your retarded one.

>>59784950
>>59785009
>>59785264


So what would you guys say then? Is it worth it to get Ryzen now, or should i just get a 7700k for gaymen then upgrade to zen2 when it has higher clocks and can compete with Intel in high refresh rate gaymen? Im looking to hit a comfortable 120fps on all games, (getting a 1080 ti) and from what i have seen so far ryzen can barely hit that in some titels while it doesnt make it in many others. 7700k on the other hand does a comfortable 150fps on most titles

>>59785536
>just get a 7700k for gaymen then upgrade to zen2
That sounds like a tremendous waste of money. Why not just jump on the Ryzen train now and swap CPUs when Zen+ comes out?

>>59785552

Because ryzen cant keep up with the 7700k when it comes to high refresh rate gaymen (from what i have seen at least), plus i have a significant backlog so ill be playing a lot of games released the last few years which are better optimized for Intel.

I want to multitask as well which made me consider ryzen in the first place but i just find it a bit of a waste to trade in gaymen performance for a secondary thing.

>>59785593
You're only losing 10 or so FPS between Ryzen and i7-7700K. I don't know where you're getting a 30 FPS difference from, except in a few unoptimized games, RTS games, or games that Ryzen struggles with (probably needs a patch).

>>59785593
It's your money, just know that by the time Zen+ comes out, that i7-7700K system will have aged like milk and you'll have trouble selling it off. It's on a dead socket with zero upgrade path to begin with.

>>59781995
This is an issue in Metro Last Light, the minimum recorded in the benchmark is useless since it always stutters at the 8th frame.

>>59785615
>and you'll have trouble selling it off
That isn't true for any Intel CPU, people routinely sell Sandy and Ivy Bridge i5 and i7 CPU's for $100. Hell, some people still sell C2Q's on Craigslists.

File: firefox_2017-04-02_15-49-52.png (72KB, 772x854px) Image search: [Google]

72KB, 772x854px

>>59785536
If you get a 7700k now you're spending
$340 for cpu
$140 for motherboard (those $115 z270s are dogshit and much worse than the $85 B350 Ryzen boards)
$55-$90 on a cooler

So.. about $550 wasted. You can go through the hassle selling it used to get some back.

If you just get a 1600X now
$250 cpu
$85-$95 motherboard
$20-$55 cooler.

$360-400.
And when you upgrade, it's just the $250 CPU down the drain which will probably depreciate less.

1600X should hit 120fps in most games that aren't just shittily optimized and do badly on the 7700k too.
With the BIOS and windows updates so far, and 3200mhz RAM, the 1800X is beating the 7700k in 30% of games. Even in ARMA 3.
The 1600X won't perform much worse than the 1800X in games since the 1800X is beating the 7700k in 30% of games with tons of CPU utilization left over while the 7700k is at 100% since most games are optimized mainly for 8 threads now days.

It's not as behind as /g/ shitposters make it out to be.

>>59785599

All benchmarks i have seen (even with the updated bios, windows and fast ram) show ryzen lagging behind anywhere from 20 to 30 fps behind the 7700k. I wish it was only 10 fps, but so far thats not what im seeing.

Take a look at the last ryzen video from GamersNexus

>>59785615

I know, and i wasnt planning to do so either. Im willing to take the loss on this of it means better performance for my uses. Also, im not argueing, just bringing up points that i believe are relevant in making the decision on a new CPU. I wish someone could convince me to get Ryzen but from what i have seen so far it looks a bit disappointing in terms of gaymen performance only.

>>59785705
>GamersNexus
Fucking hell, you know they're full of shit, right? Their latest benchmarks have an i7-7700K OCed to 5.1 fucking GHz which is impossible for anyone to hit without a massive cooling solution, super expensive motherboard and cherrypicked CPU.

https://www.youtube.com/watch?v=Rhj6CvBnwNk Look at this vid for some more sane results, has the i7-7700K clocked at 4.8 GHz which is much more in line with what you're likely to get.

File: firefox_2017-04-02_15-29-47.png (55KB, 764x954px) Image search: [Google]

55KB, 764x954px

>>59785697
Also, it does better paired with AMD GPUs.

It might be more like 50/50 on Vega compared to the 7700k, if not better.

>>59785705
>shill's nexus
The guys that put tape between an SSD and its heat sink and cried about how poorly kept the SSD from overheating and throttling?
Some experts.

Anandtech, Phoronix, computerbase.de, and Tech Alter are the only ones that seem to know how to test probably. They go in detail with multiple tests and clearly write out their methodoly.
not
>xD replaced the motherboard and CPU on my intel machine with a windows install and configuration done on intel and it sucked.
>i also reloaded game profiles optimized for 4c/8t intel CPUs with this new one lol wow it sucks!
>btw I still haven't updated my BIOS since launch.
>Oh and he's Ryzen with 2400mhz RAM while Intel has 3600mhz
>hey see this 7700k at 5.1ghz? Yeah everyone can do that. just get an $85 z270 motherboard, one of those sweet $60 LIQUID KEWLERS and click the auto overclock in your BIOS! golden silicon for everyone!

>>59785800
>>59785831

So what are some good benchmarks then?

Also, motherboard and cooling isnt a problem, im willing to pay out the ass since im already getting a 1080 ti and want to push it as far as i can.

>>59785831
>tech alter
>methodoly

my sides migrated

bless you anon

>>59786045
I've been up 20 hours. Gee I mispelled a word.

Can you point out any flaws in tech altar's testing?

They tested one of the largest variety of games.
They tested with the same RAM speed on all computers.
They tested with both AMD And Nvidia GPUs
They tested with various settings that were thought to negatively affect performance when set to their defaults.
They tested without (significant) GPU bottlenecks unlike Tom's Hardware and Tech Power Up did in any game that favored AMD.

>>59786185
Sorry anon, my laughter was genuine.

I agree with your post.

>>59786041
>Also, motherboard and cooling isnt a problem, im willing to pay out the ass
I really don't think you'd be disappointed in getting a 1500X, Geil Evo X 3200 ram or better that's compatible, and an ASRock Taichi, then.

But Nvidia drivers can cause problems so you should really wait until May for Vega and see if it's competitive.

The 7700k still wins on average if you use your PC as a gaming console, but for CPU bottlenecked games it's going to drop frames just when you get a PM on Discord and shit like that.

>>59786262
And for "enthusiast gayming", don't forget that you get 1ms lower input lag with Ryzen.

That's a much bigger deal than going from 120fps average to 125fps average.

>>59787433
Only if you use the USB controller from the CPU instead of the chipset one.

File: IMG_20151109_WA0000.jpg (55KB, 460x507px) Image search: [Google]

55KB, 460x507px

>>59776316

>>59787443
Yes. If you use the CPU i/o USB, of course.

>>59786262

>but for CPU bottlenecked games it's going to drop frames just when you get a PM on Discord and shit like that

Is it that bad in multitasking?

>>59787494
Gonna have to bring up the motherboard diagram to know.

>>59787497
It's not bad at multitasking, but when 4 cores are at 100% something happening outside the game, will cause a context switch and drop frames, 6 and 8 cores simply have more room to breathe.

>>59787497
No it's not BAD at it.
It's just most game engines have no problem 99% utilizing a 7700k now days.
When your CPU is that maxed out, and you're CPU bottlenecked, ANYTHING else trying to use resources is going to slow it down.

You can also simply cap your framerate and not have that problem with the 7700k, because a consistent framerate will look smoother than it going up and down, but then you're playing at lower FPS and it's not "winning".

This is what people mean when they subjectively say that gaming seems smoother on Ryzen, even coming from 4c/8t Intel CPUs. Its average FPS may be lower, and in some games the minimums are in a sterile benchmarking environment, but other shit going on with your PC doesn't phase it at all. Shit that doesn't happen in sterile benchmarks.
That's why I'd really just in general recommend 6c/12t to anyone at these prices.

Think of the normalfags for a second, do you think a 4 core can handle good gaming with a system filled with junk , useless services/bloatware and malware normalfags usually have on their PCs?

>>59787497
Plenty of games (and the nvidia driver) will eat up those four cores so anything else happening is going to hurt performance.

six core cpus are nice there for having just a little bit more to give

File: 1453283037961.gif (41KB, 600x394px) Image search: [Google]

41KB, 600x394px

>>59787567
Even fucking consoles have cores dedicated to OS usage so you don't suffer performance loss while gaming. It's like back in the 90's when you restarted then closed everything before even loading a game up. What a fucking joke.

>>59787631
>normies
>buying computers

>>59787694
They're not normies they're normalfags.

Mindblank tech with the run down on memory overclocking for Ryzen on the Asus crosshair VI Hero.

https://youtu.be/d0MmDrRUeeQ

>>59787728
I'm not touching ANUS motherboards ever again.

>>59787692
the new "Game Mode" in win10 finally does this.
It will put the game on n-1 cores. Which makes

Apparently this gives +10% higher minimum FPS for Ryzen in some games, but lowers the average. Imo, and what should be the opinion of everyone, lower averages at the cost of higher minimums is always preferable.

No idea how that mode affects the 7700k. I would think it does less since only 4 cores, or might make averages more worse.

>>59787567

Im looking to hit a smooth and comfortable 120fps at all games. I'l have discord, a browser and maybe some background music open usually. Would i be better served by a 1700x or a 7700k in that case? From what i have seen Ryzen seems to be struggling to reach 120fps in most titles, sometimes barely hitting it and in other cases not reaching it at all.

I wish Intle just had a 6c/12t Kaby lake/skylake, that would be the perfect CPU for my uses. Now i have to either trade off multitasking and will have to upgrade soon (6c will become mainstream soon) or i will have to give up on high refresh rate gaming. The 6800k was looking real good but they managed to fuck that up as well with the stupid 4.2ghz voltage wall. Fuck this shit.

>>59787728
>Memory with Samsung B-die NAND chips
>NAND chips

Please kill yourself you clueless shit.

>>59787631
When a single core performs the same as 2 from AMD, yes.

>>59787764
You need to be 18 to post here.

>>59787478
Wouldn't that overheat like a motherfucker?

>>59787757
Don't get the 1700X.

Either get the 1700 (comes with a stock cooler that's good for a 3.75-3.85ghz overclock usually) or get the 1800X (good clocks stock, and can OC 100-250mhz better).

You could also simply get the 1600X to tide you over until Zen2 or Zen3, while the 1800X is a much more long term investment unless you have $500 blow each year.

>>59787752
4c/8t performance would get rekt even losing one fake SMT core. Intel's SMT isn't that good to begin with.

>R5 1600X performs worse than a four core Skylake with hyperthreading
>R5 1500X performs worse than a four core Kabylake without hyperthreading
Tell me, why should anyone care about Ryzen anymore?

>>59787779

1700x is only €50 more than 1700 and will guarantee 4ghz. I also heard that some 1700s will have problems overclocking memory past 3000 mhz

I like the wraith cooler though, and since zen doesnt overclock past 4ghz anyways, i might get the 1700 after all, but im not sure if its worth it to risk getting cucked out of a stable 4ghz

File: good goy.png (604KB, 800x523px) Image search: [Google]

604KB, 800x523px

>>59787810
shoo shoo stinky shill

>>59787810
>R5 1500X performs worse than a four core Kabylake without hyperthreading
No it doesn't.

>>59787818
I've seen a number of 4.1GHz 1800X and a total of one 4.2GHz with a AIO, the 4.2GHz one is probably one out of twenty at best.

Wonder if improved process will increase the overclocking room, did the same for Haswell/Devil's

>>59787841

Why did they not focus more on getting to at leas 4.5 ghz stable though. If Ryzen was ocing to 4.5 then Intel would be completely and utterly destroyed, as even the 7700k would barely have an edge on it then.

I understand they have a limited budget but this should have been top priority.

>>59787752
AMD's been priming the console market for years with their 8 core Jaguar CPUs. People are used to this now. All Intel's got left is single core performance. It's like they're on a tiny island that's slowly getting washed away and they're bragging that it's happening slightly slower than usual.

>>59787764
*chases shekel down the street*

>>59787859
Because they're using a low clocking process which is very efficient at lower clockspeeds, as in the chip uses half the power at 3.3 than at 3.8GHz.

This is good for mobile, servers and embedded, which are made from the same die current Ryzen uses, but less than optimal for gaming, which favors clocks

>>59783149
>>59784593
>>59785697
>>59778023
>>59777994
Get on it lads.

Make a review chart of all the CPUs cataloged by price at release.

>>59787841
Someone got a golden 1800X hitting 4.12ghz at 1.325v yesterday.
But yeah, even then, going to hit a wall trying much over 4.2. It's a server laptop/process even though the architecture is good for desktop.

Usually 1700s hit only about 3.7-3.8 at that voltage.

>>59787869
Same with Nvidia.

It's clear how much more powerful GCN is when games are actually optimized for it.
Look the PS4 and Xbone, and the visuals they get with a CPU that's HALF as powerful as an RX460.

But because that half powered RX460 has hardware schedulers and controllers that don't use up any CPU.
Obviously there's other shit going on like shared memory/cache, lower latencies, but it's still only about 1TFLOPS of GPU compute compared to the 2+ TFLOPS on the RX460.

But for some reason, even console ports to PC get more optimized for Nvidia, and get a rendering pipeline that favors Nvidia GPUs.

>>59787883

Well whats stopping them from using a different process for their consumer/desktop chips and a low power process for their server chips? Same shit with the ccx thing. Use ccx based chips for your server lineup and single die chips for your consumer lineup.

Im sure there is some kind of reason why this wasnt done but it seems like a top priority to me, it doesnt make sense to me to cut corners here.

>>59787883
Hopefully next year AMD has money to afford a HP process for the PC market and a LP process for others.

>>59787833
Yes it does
>AMD Ryzen 5 1400
>Passmark ST: 1685
>*4 *1.2= 8088
>Geekbench ST: ~3830
>*4 *1.2= 18384
>userbenchmark: 376 Pts
>Intel Core i5-6600K
>Passmark ST: 2147
>*4 = 8588
>Geekbench ST: 5096
>*4 = 20384
>userbenchmark: 474 Pt

>>59787917
>Well whats stopping them from using a different process for their consumer/desktop chips and a low power process for their server chips?
Simply put, they lacked funds and needed Zen to be a one-size-fits-all savior to all the company's needs. Zen+ should solve some of these issues.

>>59787917
Money, making lith masks is very expensive and time consuming, moving from one process to another takes time even in the same foundry.

>>59787927
Are you seriously just multiplying single core scores up? Are you on fucking crack? Those shill pills must have had some LSD mixed in with them.

>>59775957
>makes reply

graphs says it's ok

coststs /2

jewel shill again

and i'm a guy hwo got 6100 for free, so I'm not a drone, I no lies,

>>59787927
>pissmark

>>59787927
You forgot.

>6600k
$250 at 3.5/3.9
>1400
$160 at 3.2/3.4

Then you count in cooler and motherboard prices, etc, you're retarded.
And you're doubly retarded for using synthetic benchmarks.

Seems to be another case of AMD with the billion good cores vs Intel with a few great cores. AMD rocks performance tests and professional use while Intel is slightly but notably faster for gaming.

Granted the gap has closed a bit to the point where if you do a lot of both pure processing and games Ryzen is not a bad choice.

>>59788049
No it's not, a Zen core only has 6% less IPC than Kabylake, nothing compared to a Bulldozer core that was 35% slower than a Sandybridge core.

The difference this time is that Intel can overclock higher, but that pretty much relegates them to to the two unlocked K SKUs

>>59788049
It's a good choice for your wallet too, especially since AMD has said they aren't changing sockets any time soon. Imagine actually buying an upgrade for your CPU and not replacing your motherboard and RAM. It's like some sort of dreamworld fantasy.

>>59788076
Three if you count the 7740k for which nobody knows why it exists

>>59787935


Didnt know that its that expensive to switch processes/build of two different processes instead of one. The more you learn i guess.

>>59787941

>lith masks

Got any links where i can read more about this

Also, isnt this done by the foundry? Or does AMD make them and them send them to the foundry for them to use?

>>59788110
Semiwiki is your best bet, search there.
Or just google lithography mask.


AMD is the one making them, it's basically something like a mold you can reuse.

>>59788128
Correction, AMD+foundry makes them, that's half the reason it costs that much, you need no small amount of people working on it.

Aside that, you also need to tune the CPU for the new process, fix signal clarity and integrity, make the buses handle the clocks, fix synchronizations, etc, this is all done early in the process.
Heck, maybe Ryzen 1 couldn't even benefit from a HP process due to early implementation and library decisions, we have no way of knowing this.

>>59788213
>>59788128


So lith masks are like a really expensive die, if we compare it with casting for example. I guess it makes sense that its that time consuming and expensive as you basically need to map out every part of a cpu containing billions of transistors (if im correct).

Lets hope that Zen+ will be fast enough to beat intel in single core performance, that should really shake up the market.

>just a 20fps difference

why are you gaymerfags so fucking retarded? measuring absolute differences without context is worthless. Use percentages instead

File: 1491465079591.jpg (30KB, 647x594px) Image search: [Google]

30KB, 647x594px

I'm going to buy a Ryzen just because of all this autistic anti-competition spamming.

I listened to 4chan about bitcoins, and didn't buy any.

Whatever you people say is true is the opposite of reality.

File: 1484686763369.gif (36KB, 265x200px) Image search: [Google]

36KB, 265x200px

>>59787927
>passmark
>and only single thread scores
Wow. Could you make your shilling any more obvious?

Ryzen 5 1400 passmark:
8700
i5-7400:
7450
and PASSMARK IS FUCKING AWFUL.
In real world applications and games, the 1400 heavily outperforms the 7400 and even 7500 on average.

Wait. Holy shit. You compared it the 7600k that costs like $300 more for a build. That's even worse shilling than initially thought.

File: download (1).png (9KB, 239x211px) Image search: [Google]

9KB, 239x211px

>>59787927
what this anon said >>59787949

>>59787918
>>59787935
Yeah. That's many's hope, but it's obvious why it's this way to start with.

Bigger money and safer bet with enterprise contracts.

>>59788109
+100mhz base!
possibly accurate TDP this time!

>>59788110
lith masks are like screens in screen printing.
But they use acids and lasers to etch instead of ink.

Think about how those tiny little die has billions of transistors, and how they ought to make that much circuitry and transistors in that small area

>>59787904
I will be making a fair comparison chart when consumer chips have been tested and prices have stabilized.
>>59784958
Rumor I heard is the APUs might have some cpu improvements. Makes you think.>>59784950
I heard possibility of a AM4+ for R3 with backwards compatibility for R1 like AM3+/AM3 boards.

>>59788109
to prepare our anus for all consumer K cpus to be on a different socket from the rest

>>59788560
dont call 3rd gen r3. that's fucking confusing.

Anyway, AM4+ is likely to come with DDR5 and/or some improved PCIe standard.
Potentially that could come in 2019 with zen3. But the CPUs would still work on an AM3 motherboard, just without the DDR5 support or whatever else is new.

>>59788612
>DDR5
>2019
Not happening
Expect in in 2020/2021 when Intel finally releases their new Arch

>>59787810
Here's the cheapest parts for a decent build, ranked by GAYMING berformance. AMD of course can't challenge the lake i7 in ultimate performance but just look at the numbers

7700k
https://pcpartpicker.com/list/XtRBJV
$567

6700k
https://pcpartpicker.com/list/LvCJZ8
$533

R5 1600
https://pcpartpicker.com/list/JjhMgL
>$414

7600k
https://pcpartpicker.com/list/vwK8vV
$462

6600k
https://pcpartpicker.com/list/wVRBJV
$434

Hmm.

>>59788612
>dont call 3rd gen r3
Shit you're right. I'll do zen 1-3.

>>59788777
You haven't seen the news the last 3 days, have you?

https://fossbytes.com/ddr5-ram-2018/

File: 2139145 - Hilda Porkyman.jpg (102KB, 1000x1360px) Image search: [Google]

102KB, 1000x1360px

>Tfw amd can't into competing and Intel have no interest in increasing performance
>Tfw my 4690k will remain relevant for the next 40 years

>>59788837
>6 phases for the 7700k motherboard

Your opinion is worthless.

>>59788837
Point sort of made except those are garbage.
You are hardly going to be able to OC those unlocked CPUs at all with those motherboards.

Including the 1600. That is the worst motherboard that only has 3 phase VRMs.

You need at least 6 phase for Ryzen and 8 for intel.

The 1600 build should be $15 more and the Intel ones should be about $50 more.

>>59788924
it's cheap as possible, you can shave a few bucks on the ram but it's bottom dollar, I also wouldn't use a deepcool hsf and I'd buy faster ram. Not the point of the comparison

>>59788949
With such a shitty motherboard, you should buy a locked CPU.

Comparing a locked 7700 to the 1600 or 1500X actually makes more sense, anyway.

>>59788949
You don't understand, for STOCK CPUs a 4 phase motherboard is recommended, not optimal, but recommended.
If you're overclocking a 7700k, which can use TWICE the power at 5.0 compared to stock, 6 phases isn't gonna cut it, even if they're quality 20A ones, you're literally playing with fire and hoping your mobo, PCIe devices and PSU explode.

Here's my advice, don't build anything for anyone ever again.

>>59788837
Your mobo choices are worthless for overclocking

>>59788990
>>59789072
If only most other people realize this when they try making a 7700k build and don't realize it's going to be $550+ for the cpu+motherboard+cooler alone.
They just look at the CPU cost.

>>59788908
Wishful thinking.

>>59788906
>He did not read the article he quotes

'But, as processor and SoC-makers need to make appropriate changes, to use the DDR5 RAM in your PCs you’ll need to wait for a few more years.'

>>59789158
DDR4 standard final specification
>Sep 2012

Haswell-E (first platform with DDR4 support)
>Aug 2014


So, two years, also considering x86 performance has slowed to a crawl, we'll get it sooner because Intel needs a new socket to sell and needs more reasons for people to upgrade.

K processors don't even have VT-d

Lol

>>59789127
You forgot the delid job, so voided warranty to top it off, and $20 more if you can't do the delid job yourself and need a tool.

Meanwhile all a 1700 needs is the stock cooler and an Asrock AB350 Pro4 to get 3.8 and 3.9 if you're lucky with the silicon, that's $90 mobo.

>>59789290
ASrock really has the best motherboards for AM4, by far.

What the fuck were MSI and ASUS doing? Celebrating new years?

>>59789337
The Asus Prime actually has 10 phase VRMs. People seem to be getting good overclocks with it and getting RAM to 3200 fine. For some reason, it's the fuckexpensive Crosshair that's had problems.
But it seems most of the others are at least 8 phase anyways.

Those $85-$95 B350 motherboards are literally batter than the majority of $140 Z270 boards.

>>59788963
>>59788990
>>59789072
I trust pcpp more than you shills

>>59789444
That's because you're about as dumb as pcpp is, similar people flock together.
But hey, place your trust in a motherboard designed to handle up to 150W load with a CPU drawing 200W, wonder which color that magic smoke will be?

>>59788908
>Tfw my 4690k will remain relevant for the next 40 years
That doesn't really sound like a bad thing.

File: IMG_2788c-678x381.jpg (44KB, 678x381px) Image search: [Google]

44KB, 678x381px

I'm set until Zen 4 comes out

>>59789430

What would be the best mobo to get now? The titanium, though way too expensive, looks really good. I wish it had base clock overclocking though

The crosshair seems to be the hot thing now, i would get it if it didnt look like dogshit. Yes i know its pleb tier to involve aesthetics in the decision but this particular board looks really bad imo.

>>59789290
>Meanwhile all a 1700 needs is the stock cooler and an Asrock AB350 Pro4 to get 3.8 and 3.9 if you're lucky with the silicon, that's $90 mobo.

Is there a good tutorial on this?

t. Noob with a 1700, stock cooler and AB350 Pro4

File: Screen Shot 2017-04-07 at 9.44.51 PM.jpg (370KB, 943x1262px) Image search: [Google]

370KB, 943x1262px

>game fps charts

>spic wetback charts

>french charts

show me the SPEC, all these gay-ass non benchmark lame ass butt fuck shill benchmarks for gay men are lame

>>59776284
>using your cpu for work and productivity

gtfo my board, this is the >>>/v/ tech support board, not tech nerd board faggot

>>59790512
https://hardforum.com/threads/ryzen-7-1700-b350-overclocking-tidbits.1926296/
this is the best info on b350-1700 overclocking I;ve seen so far.

File: 668295727417.gif (263KB, 480x479px) Image search: [Google]

263KB, 480x479px

>>59777967
>them minimums
Intel + Nvidia is proven to be the masterrace once again.

I think Intel's jewish tricks are getting to me, it actually makes more sense to me to buy Intel because i can get a cheaper motherboard and ram than if I went for ryzen because i dont expect them to last as long thanks to the socket change meme.

>>59775957
>$220 R5
>$340 i7
>competing

>>59777130
>>59775957
How the hell are there games being bottlenecked to 60 FPS by a $300 CPU? Did optimization go to complete shit in the last few years?

>>59777967
>heavily underclocked Intel CPU beats the shit out of Ryzen


welp

What would be R5's Phenom II 965 BE from 2011?

>but amd is being held back by Nvidia!
https://www.youtube.com/watch?v=zmOglA32uRU
Watch shills try to deny this video.

>>59779504
>opineon
haahahhhahaahahahhahaahha

File: amd nvidia dx12 hardware unboxed.png (40KB, 589x389px) Image search: [Google]

40KB, 589x389px

>>59793223
>Hardware Unboxed being indirect as possible to not make Nvidia look horrible; admits to doing so.

Nvidia gimping shit as usual. Or just having crap DX12 drivers that they can't do anything about.

>>59793259
>opine: "hold and state as one's opinion"

>>59775957
>1 game on intel side is doing 38% better than AMD
>"Literally half" while other games are 15% or less better

>>59793048


https://www.gamedev.net/topic/666419-what-are-your-opinions-on-dx12vulkanmantle/#entry5215019

Its been shit for years. Its why Raja was talking about a good 50% of a gpus performance being on the software side (as well as smarter memory management being a big thing for fiji).

Why the fuck is that chart comparing Ryzen 5 with an i7? Shouldn't it be compared with an i5?

For anyone sane, it was an expected outcome.

It doesn't matter whether it has 4+4, 3+3 or 2+2 Zen-cores, the max. for allcores is around 3.8Ghz with reasonable voltage set.

File: C5h1GFdXMAAERfK.jpg (116KB, 768x1024px) Image search: [Google]

116KB, 768x1024px

>>59775957
>literally half
dr. pepper chugging pork snorts at you're edumication

>>59779300
>the 5xx Polaris rehash will be cheaper

File: r5 1400 vs i5 7500.jpg (912KB, 1920x1080px) Image search: [Google]

912KB, 1920x1080px

>intel corelets will defend this

>>59788259
>being this fucking retarded

>>59794016

I wish GF's process was at least a bit better. This was supposed to be using 14LPP and yet it can't clock into the 4 GHz range.

At least power efficiency is great. But I am hoping for Zen+ to rectify the power scaling and performance. It should become the new Sandy Bridge.

>>59792512
Minimum frames only count on AMD.
On intel minimum frames mean NOTHING!
You hear me Sutterfucks scholmos!

File: The lake, the lake, the lake is on fire.png (1MB, 1110x5110px) Image search: [Google]

1MB, 1110x5110px

>>59794165

>ryzen
>more cores
>and runs cooler

I mean we all knew kabylake is a nuclear reactor but this is silly. Can't wait to see /g/ defend kabylake murdering your power bill.

>>59789736
It's a good value for poorfags but that's all it is. I don't see it lasting until than end of year before it starts CPU bottlenecking everything even with a 1050 or similar poorfag GPU.

>>59792739
You can't OC on that shitty motherboard and RAM, so it defeats the entire purpose. Without massive clock advantage from OCing, Intel loses.

>>59794337

>B350 master race

While they aren't that great for the ryzen 7 chips you should be able to get a decent overclock on the lower core count ryzen chips on the lower tier boards. AMD may not have performance crown but it is the secondary and tertiary features they are aboslutely nailing.

>>59794190
Yes, Zen+ will be the good ones, if AMD do their homework.

Zen 1st gen is a beta platform, and only for fanboys who like buying stuff and having to deal with all sorts of problems.

I am going to buy a cheap Intel mainboard and a cheap Intel CPU now, so that I can switch to a new platform at the end of next year or so.

>>59793086
If you actually watched the video, he mentions the entire point is testing the GPUs.

File: 1477758235807.jpg (5KB, 249x180px) Image search: [Google]

5KB, 249x180px

>>59775957
Who here 7700k master race forever ?

>tfw no one tests undervolting and underclocking stability
I don't even know why Intel sells their T and S SKUs. They don't make any sense. Their idle and near-idle power consumption is exactly the same. The only difference is their load power consumption which is artificially throttled by lower clock speeds, not voltages.

>>59794432


https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/

>>59794337
i dont mean super cheap, something like a z270-A instead of the taichi id get if i went AMD. Also 3600mhz memory is pricey.

>>59794359
>I am going to buy a cheap Intel mainboard and a cheap Intel CPU now
That's really just wasting money. Get a good B350 mobo and a cheap R5 (i5 performance but cheaper) instead and upgrade to Zen+ or Zen 2 when they come out. AMD will be using the same chipset for at least 3 or 4 years.

>>59794491
That's fairly neat. I've tried testing my own 7700K and 6700K at lower voltages, but my 6700K seems to need much higher voltages to sustain lower clocks compared to the 7700K, which doesn't seem to make a whole lot of sense (I'm talking about difference above 0.1V)

>>59794581

Ryzens voltage scaling opens up a world of options for laptop and server parts. Even going beyond that ryzen could do very well in a lot of semi-custom designs depending on the needs of the client.

I would wager serious money that ryzen (in one form or another) will be the backbone of the next generation of consoles. A ryzen + vega APU designed for a console would be an absolute monster given the sort of TDP you could squeeze out with some low(er) clocks.

/g/ often overlooks at how powerful (for the TDP) AMD's igpu architectures are - its been the cpu that has been holding them back (well, that and a lack of memory bandwidth).

>>59794572
>70$ CPU+50$ mobo
>this triggers the amd shill
>NOOO BUY THE RYZNE!!! WHAT ARE YOU DOING?!

>>59794572
>good B350

There are not good boards out there.
Also I need a µATX or Mini-ITX board.

Too many issues atm, especially the long boot up time would piss me off.

>>59795393
Asrock ab350 pro4 is the best board out atm.
There's a mAtx version as well

>>59794405
aye