ZEN MASTER RACE! 15% faster than a 4790

>>56022644
>a 2016 CPU can beat a CPU that's literally three years old

Wow its another AMD shill thread
My i7 2600k still gives me good performance in all the games I play
AMD processors are a joke
Ashes of the Singularity is an "AMD" game meaning it will be optimized better for AMD stuff anyway.
This whole brandwhore shit is killing the tech and games industry. One game gets optimized for AMD and it runs like shit for Intel/Nvidia and vice versa.

I want it to be true, I really do.
But we both know this is just AMD fantasy.

>>56022644
D..D..DELETE THIS RIGHT NOW, ILL CALL THIS S.SHILLS!!!

>>56022694
You're a retard.
Intel and AMD processors have the same fucking instruction sets so you physically cannot optimise for one other the other. Unless you're suggesting that AMD designed their processors JUST for AOTS.

>>56022694
4/8 and 6/12 core processors are the future, AMD is just going to get stronger as time goes on.

>>56022644
>15% faster than 4790
>losing to 4790

Phew lad.
It sure is a big jump, but it's competitor isnt haswell or devils canyon. Its competitor is kaby lake and broadwell E

>>56022722
More like they optimized their GPUs to run better on games like AOTS or i believe the new Deus Ex or Tomb Raider

>>56022722
>cannot optimise for one other the other
Unless you're Intel and make your compiler fuck with competitor cpus.

>>56022644
>AMD just now getting around to competing with intel's 2014 processors
damn this industry is fucked.

https://forums.anandtech.com/threads/first-summit-ridge-zen-benchmarks.2482739/page-4#post-38414448

The shill who started spreading FUD about the Summit Ridge ES being behind Sandy Bridge got called out for cherry picking questionable benches, and got his shit slapped by a mod.
Ultimately the AoTS benches mean nothing, but on face value it showed a 3ghz~ Summit Ridge performing something like 80% better than the FX 8350.

>>56022722
Instruction set support is not the same as instruction set implementation. Learn how compilers and cflags work.

>>56022763
All that matters is the price.

Besides even with the Neo and Scorpio upgrades, consoles (which dictate gaming graphics) have CPUs way worse than Zen.

>>56022772
Well yeah if you control the fucking industry you can cripple competitors with 'Genuine_intel', but that's a different issue.

>>56022644
Go back to your consolewar threads, /v/baby

>>56022644
Wait but that pic shows Zen being behind Intel in most AMD shilled game in existence.

>>56022722
You're right, but there was a time when benchmarks did a check for intel and if it wasn't found performed worse for no other reason.

Removed that code and performance increased magically.

>>56022763
Summit Ridge is competing against Haswell-E and the mid range 6 core Broadwell-E.

>>56022813
>most amd shilled game
You mean a multi core game?

That's an i5 it beat, which is outdated for future.

>>56022801
I care because emulators. My sister works for intel, but i will take the highest possible IPC available, even if i can get a 6950x for 750 dollars.

>>56022798
AMD is the king of benchmarks that mean nothing, nvidia is merely a prince of meaningless benchmarks.

>>56022644
>faster than 4790
thats not what u have in ur picture famlam o.0

>>56022644
Wow that's like 40% faster than an 8350 and is only clocked at 2.8GHz

>>56022836
I don't get the hype. Worse than 4790 which you can buy for less than $250. From what I've heard summit ridge is gonna be expensive as fuck.

File: 2848030.png (584KB, 1195x577px) Image search: [Google]

584KB, 1195x577px

did anyone notice these two?
http://www.overclock.net/content/type/61/id/2848033/
http://www.overclock.net/content/type/61/id/2848030/

they're both ran in standard, both at the same resolution, and both 8 core 16 threads. but they're different ES samples.

one is ES 1D2801A2M88E4 and scores 5,300.
the other is ES 2D2801A2M88E4 and scores 3,100.

wccftech does touch on this but they decided to completely ignore the 202801A2M88E4 results.
>There doesn’t seem to be any difference at first sight between the 1D and 2D variants. They are both purportedly clocked at the same rates and the only difference that we can see is the fact that the 2D variant fares much worse in terms of performance. ..................
>In any case, because of these reasons, I decided to focus on the 1D variant for our comparison.

clearly these are indeed ES samples and being ES samples we cannot truly gauge zen's performance since they have TWO ES samples floating around with drastic performance differences between the two. these are not finished zen products, only test, ENGINEERING samples.

also to note the 1D in the leaked order part number of the ES sample showed the 1D versions with a base of 2.8 and a boost of 3.2. but in ashes we don't exactly know what speeds they're running at and if its turbo is activated and if it is, is it being maintained. we also don't know how many cores are activated. its all guess. we don't know much at all. ashes only reports the default cpuid infor.

with the 2D sample we don't know anything about it outside of ashes showing it too as a 8 core 16 thread zen processor.

and that's the point, we don't exactly know anything about these ES samples. its all guess. we don't know much at all.

>>56022675

that means it beats skylake as well you dumbass. clock for clock its about 8% faster than skylel

>>56022867
Who knows, but it does have the Wraith cooler on it.

>>56022644
I hope for the best for AMD and ZEN I really do and if it's great then I'll look into making me a new pure AMD rig but till then I'm sitting snugly over here with my i7 so I'm good either way.

File: meme6863422076.jpg (18KB, 198x225px) Image search: [Google]

18KB, 198x225px

>>56022644
>AOTS
why do amemede shills only use this game for comparing against NVIDIA or Intel??

>>56022936
Don't feel bad, you i5 fags have your old DX11 games to play.

>4790

don't tell me zen can only beat stock clock 4790, 4790 is like 2 years old cpu.

>>56022867
You make it sound like Intel E cpu are cheap.

>>56022644
Please don't do this to yourself again AMD fans, its the same story like the RX480 don't let Rajesh trick you with the hype benchmarks again.

AotS is a game that can fully utilize an 8 core CPU. The fact that it's not 2x as fast as the 4790 is in this benchmark is the disappointment. Zen is most likely only 20-30% faster than current FX 83xx chips clock for clock and seems to have lost most of the frequency scaling that came with those housefire architectures.

>>56022870
this

2D is worse than the bulldozer benchmarks and the other is slightly faster than a i5 haswell.

clearly these are not finished products. engineering samples be engineering samples.

>>56023028
see
>>56022870
>>56023043

>>56022644
looks like I was right about it beating Haswell i5
but this is only a engineering sample its not the final product
but it gives us a clear indicators on where it (Zen) sits roughly speaking.

>>56023053

the fact that it's an ES doesn't matter, it's A0 stepping and AMD doesn't have time to respin before launching (because it takes 6+ months to manfacture and retest and validate etc).

best case is that there was some showstopping errata that degrades performance like the TLB bug, worst case zen turns out to be another bulldozer.

>>56022644
>believing "leaked" benchmarks

Did you all forget about how bad AMD fucked Bulldozer up?

AMDrone always overhype everything

>>56023081
Wow. who could have foreseen AMD dissappointing again?

Like I've been saying for years, AMD is only able to deliver disappointment.

>>56023081
The Summit Ridge ES CPU scores were right behind the scores of a stock i7 5960X.
Being an ES does in fact matter, because things like the NB might not be running at full speed. Any part of the uncore could be barely functional.

AoTS is a useless bench
and you're a shitty shill

>>56023127
>The Summit Ridge ES CPU scores were right behind the scores of a stock i7 5960X.

no they weren't, they were on par with a stock clocked i7-980.

>Being an ES does in fact matter, because things like the NB might not be running at full speed. Any part of the uncore could be barely functional.

it's the launch stepping. AMD can't respin before launch without pushing it back or putting out defective chips on day 1 (if they're actually defective and not just bad by design).

File: 2848027.png (604KB, 1194x582px) Image search: [Google]

604KB, 1194x582px

>>56023081
they have two 8 core / 16 thread ES samples floating around with drastic performance differences.

one is WORSE than bulldozer and the other with 8 cores and 16 threads is sightly faster than a 4 core 4 thread haswell i5.

if these are final samples then this makes zen an absolute failure.

its actually WORSE than bulldozer in performance with either sample because it takes a 8 core and 16 threaded zen processor to outperform bulldozers 8 core 8 threads only.

clearly, these are not final samples nor give a valid representation of zen at all.

>>56023094
>Bulldozer fucked up
in syntetic benchmarks and maybe 2 games not by its self
it bet the i5's of that era for 2 Gen's

AMD was only i5 tier anyway lol

And i'm a ex-intel shill so embrassing to tell another shill off.

>>56022842

look at the clockspeed you dumbass

>>56023157
I knew you were the exact same shitposting child from earlier, you're regurgitating the exact same bullshit word for word.
The results linked by that faggot Sweeper were overclocked, not at their stock clocks.
The Summit Ridge ES was only slightly behind a stock i7 5960X.

Feel free to keep regurgitating his debunked bullshit though. Its likely that you are him anyway.

File: r_600x450.png (124KB, 600x450px) Image search: [Google]

124KB, 600x450px

Zen was supposed to be HEDT, yet it can't even beat 2 generations old mainstream Intel products in a parallelized game and it's ~1/2 the perf of actual HEDT products.

>>56023187
>The Summit Ridge ES was only slightly behind a stock i7 5960X.

[citation needed]

OP pic shows Zen as ~50% slower than a 5960x despite having higher clock speed.

>>56022644
AMD must be stuck in some warped dimension. They just keep releasing useless GPUs and CPUs thinking it's 2012-2014.

>>56023168
>it bet the i5's of that era for 2 Gen's

Wat

It was on par with the 2500K for the most part but not beating it.

Also if you go anywhere else other than /g/ or read basically any review of release Bulldozer is was WIDELY regarded a dissapointment

http://www.overclock.net/t/1139571/official-amd-bulldozer-reviews-thread

>And i'm a ex-intel shill so embrassing to tell another shill off.

Not a shill, I care about performance

>>56023210
the i5 of last gen i5-750 and i5 760

the 2500k was only beating in single thread and alittle on multi thread so its meh teir..

Who cars anyway bulldozer should of been aborted on release.
I owned a bulldozer I don't need to read shit.

>>56023165
pretty much this

after all the whole reason why amd ditched cmt and went back to a smt design was to focus on, and increase single threaded performance.

if these benchmarks are legit, it clearly shows zen performing worse in single threaded performance over bulldozer with both ES samples.

it makes zero sense for amd to release zen at all if its worse than bulldozer in single threaded performance.

yeah i agree with >>56023165 this does NOT at all represents a finished zen. its just a testing chip. kinda does indeed show this is NOT a launch sample that this op >>56023081 wants to proclaim. two ES samples with such varying results alone shows this.

>>56023294
>it bet the i5's of that era for 2 Gen's

That Era was the 2500k nigga what

>Who cars anyway bulldozer should of been aborted on release.

You should be aborted

>>56022644

Oh god the false-flagging has already begun.

Yes that's it. Build up unrealistic expectations and crank up the Hype. Then when it's released and it's pretty much what they said it was going to be you can all cry about AMD lies.

File: AMD_FX-8150-18.jpg (113KB, 650x563px) Image search: [Google]

113KB, 650x563px

>>56023300
>it makes zero sense for amd to release zen at all if its worse than bulldozer in single threaded performance.


you say that but Bulldozer was actually worse than Phenom

Someone tl;dr this shit for me, please. Wasn't zen pushed back to Q1 2017? How do we have benchmarks?

>>56023363
They should have just kept updating the Athlons instead of wasting a ton of money on Bulldozer

>>56023364
Nothing was delayed, this too is pure FUD.

A guy with no connections to AMD got a hold of one of the test systems AMD had at Computex Taipei, and someone eventually found his AoTS profile where his benchmark runs were recorded on it.

>>56023378
You mean Phenom right?

cause Phenom was an amazing time to be alive.

>>56023394
Yes, but Phenom was just a minor update to the K8 CPUs, not an entirely different architecture like Bulldozer.

>>56023392
So the benches are legit?

>>56023394
You mean phenom II, right? Since Phenom I was a disaster

>>56023454
Legit engineering samples.

>>56022644
How much faster is it than my dual 2670 v0?

>>56023466
Thanks for explaining.

>>56022644
Wew i never bothered to look at the actual game. Looks like modern supreme commander. Is it good?

AYYMDFAGS ON SUICIDE WATCH

File: 1457920584548.jpg (46KB, 600x549px) Image search: [Google]

46KB, 600x549px

>>56022644
>TFW when you faggots overhype this and it underperforms.

Next thing I know some faggots are going to start hyping OC'ing and getting FX9590 levels of OC'ing.

>>56023481
No it's just another total commander clone

>>56023209
>muh price to performance!
Zen will be half the price of the i7s it is competing against and AMD fanboys will be like
>paying twice the price for 20% more performance

>>56023490
I agree. It's time to stop hyping AMD until they deliver.

>>56023524
It's a fair point though, why should everyone look to use the latest unlocked i7?

>>56022870
I thought 2D was the mobile variant?

>>56023547
To get rid of AMD driver overhead

>>56023547
Because I would rather have that 20% extra performance than pay even more to get a dual socket board to stick two cheaper processors in.

>>56023524
Just like 480 is price to performance king while being as expensive as 15 percent faster 1060 and 50 quid more expensive than 3% slower stock 970 ey?

>>56022644
Look, Zen is shaping up to be a good chip but we have no fucking clue how it scales with clock speed and we have no clue what intel has hiding in the wings as a counter to it.

Relax, take a break and make yourself a bloody cup of tea. Don't get so hyped over a chip that isn't even in final silicon yet.

>>56023308
>>You should be aborted

Little late for that, don'tcha think?

>>56023615
Literally everything in this thread makes me think that Zen will be biggest AMD fuck up yet, how to hell you came to opposite conclusion.

>>56023650
So long as their 8 core 16 thread CPU is close to intel's performance at 50% of the price they'll be fine

>>56023603
Depends on where you live. Some countries overprice the 1060.

>>56023658
He's an i7 6700k fag so he doesn't understand.

>>56023658
Look at the fury x and pro duo

You're expecting amd to not be retarded when it comes to the pricing.

>>56023650
Greater than Haswell IPC, 8 physical cores and 16 threads at AMD prices.

I'm getting real tired of my 2500k and really there isn't any point upgrading to intel's mainstream chips, Zen might finally represent a value for money upgrade that gives me something MORE than my current chip in real terms without requiring me to move to an expensive premium platform like LGA2011.

Plus, Zen based APUs could be a big step up for small form factor media computers and mid-sized laptops in terms, current AMD APUs are hampered by being tied to the god awful bulldozer core, a GCN 4.0 + ZEN based APU would help rebalance that.

>>56023695
Fury X is now $400, Radeon Pro Duo was meant for VR Devs because they let it use workstation or gaming drivers, was retarded at $1500 of course though

Hopefully they don't fuck up with Vega, it should have HBM2 which they should do well with given their experience with the Fury Cards

>>56023735
>Hopefully they don't fuck up with Vega
This is AMD we're talking about

What's it like to believe in fakes? Fucking AMDPOORFAGGOT RETARDS

>>56023747
This can't be fake because it shows AMD's processor sucking

>>56023005
There's no Indians in the CPU division, they had a team of actually competent poos in 2010, they designed the first Jaguar core and soon were poached by Samsung

>>56023764
It's 100% fake because AMD is incompetent and AMD is literally 120% slower in IPC than Intel NOW and anyone expecting that to go back to 10% is LITERALLY retarded

>>56023742
I believe! If nothing else my Fury X should last me a good while for my 1080p 144hz display. Really need to replace this FX chip though, hopefully AMD doesn't fuck up Zen either, otherwise I'll have to go X99

>>56023603
>muh yurop pricing

>>56023735
Is Vega RX 490 or 580?

AMDRONES SERIOUSLY BELIEVE ZEN IS GOING TO BE BETTER THAN ANY CURRENT INTEL CPU

LMAO

>>56023695
Pro Duo is priced the way it is because, much like the Titan X, it's a workstation card. Not meant for consumers. It's geared towards content creation, while being capable of gayming, and supports shit from both the Radeon and Firepro series.

FUCK AMD!!
AMD IS GARBAGE

>>56023490
>FX9590 levels of OC'ing
But anon, that's an OC'd FX8320.

Wtf i hate amd now

>>56023796
Vega is probably going to be called Rage following Fury

RX 490 is probably just a dual 470 or 480 card

and there's probably a $140 3 TFLOP 465 coming eventually

File: 1466901897245.gif (244KB, 150x113px) Image search: [Google]

244KB, 150x113px

gj mr amd

>>56023809
Titan X is squarely a gaming card, only the first Titan was used as a workstation card since you could get TOTL GPGPU performance for a fraction of the price of a equivalent Quadro card
The Radeon Pro Duo is almost the same, though it is aimed at VR development too

>>56022722

then there's differences in cpu architecture but lets just completely ignore that...

>>56023827
ATi RAGE was a terrible product, are they really gonna go with the same name?

Then again most people don't even know what ATi was nowadays, much less about the RAGE series

>>56023695
Fury x was 600€ and titan x performance the time it was released.

>>56022763

This was using the slower 1D engineering sample. There is a faster 2D engineering sample floating round there, too.

>>56023873
Pretty much, no where to go from Fury anyways

>>56023873
A lot of Ti cards were awful too, yet we got the Ti moniker back

>>56022644
Hope they didn't make the pipeline TOO short and this thing can't clock over 4GHz

The only product you should be judging is the one that is on store shelves.

>>56023878
Problem is though the Titan X and 980ti could actually OC

This is a ES, it can be anything from disabled microcode to lower clocks, even higher clocks.

Just wait until we get actual leaks.

>>56023892
They're probably being fairly conservative assuming it's real, it'll probably be like most recent intel chips were the average OC maxes out at 4.5ghz

>>56022772
THIS
Holy fuck
Seems everyone forgets that Intel did this

>>56023781
>I have no idea about CPU design but it's impossible that they have done that
>jim keller architecture

It sure smeels autism here

Is AMD gonna fuck this up again and run the stock units with too high voltage(like I want to undervolt to get what the chip is meant to perform!) and ruin their perf/watt ?

>>56023885
The first Ti (560) was a great budget card.

>>56023948
That's just to get as many units out as possible.
Opterons, you know the actual market that requires perf/watt is gonna have a much smaller voltage range.

File: 1462275452514.jpg (166KB, 1920x1200px) Image search: [Google]

166KB, 1920x1200px

>>56022765
>GPUs
How is that relevant to Zen?
Is Radeon Technologies now developing CPUs?
Are you retaded?

>they optimized their GPUs to run better in games like [...]
Who would've thought that technology utilising hardware async-compute would do better in games that use low-level APIs that take advantage of async?

>inb4 b-but only AMD does async therefore it gimps nVidia!!!
nVidia themselves admit async is the future of GPUs, they implemented software async in Pascal, Volta is destined to be "fully capable" of it in 2018 (6 years behind AMD)
nVidia also "optimises" nVidia sponsored games. "Optimising" them by adding a metric shitton of useless gimmicks like HairWorks™, tesselation and other GimpWorks™ bullshit which historically hindered competitor's performance as well as their own, but not in such a large scale.


Basically, you're a fanboy who cannot observe the truth no matter how obvious it is.

>>56023955
>doesn't remember the Ti500
Go to sleep kiddie it's past your bedtime.

>>56023965
Nvidia has hardware schedulers back in Fermi, they ditched them to reduce power consumption

>>56023983
Has anyone actually don't A-sync testing for fermi? Or is it too old for it to matter/

File: 94f.png (121KB, 680x497px) Image search: [Google]

121KB, 680x497px

>>56023983
>tfw after nvidia adds a hardware scheduler back in their performance per watt will tank hard

File: 1377757349286.png (165KB, 417x360px) Image search: [Google]

165KB, 417x360px

>video game benchmark
>engineering sample
They just pulled random results out of the benchmark database for comparison and don't know the actual clock rates of the CPU or GPU.

>>56023903
You can oc the fury x to 1250mhz..

>>56023923
While that was really shitty this is probably what did AMD in

http://money.cnn.com/blogs/legalpad/2007/02/suit-intel-paid-dell-up-to-1-billion_15.html

>>56024166
The article you posted was completely unrelated. That's just (yet) another example of Intel being scummy.
http://www.agner.org/optimize/blog/read.php?i=49
Many games use the Intel compiler, notably anything from Blizzard. Starcraft 2 shows significant gains when spoofing the genuine Intel CPUID.

>>56024271
Most people dont check benchmarks imo those exclusivity deals hurt AMD more by giving Intel more mind share to the layperson

>>56024306
Perhaps but I think it's more of a case of the creationism vs evolution "debate". AMD gets more publicity by being presented as a competitor to Intel. A slightly more legitimate use of the strategy in this case though.

>>56023965
https://rbt.asia/g/thread/S54129897#p54130673

>>56022722

What? That instruction set is literally a higher level programming language. Instructions are interpreted by microcode. You can optimize the shit out of it.

Also, WHAT IS BRANCH PREDICTION??

>>56022644

CPU's don't matter. Every thing from Sandy bridge is fast enough. Only if you fuck up by buying an AMD CPU you should worry about new stuff.

>>56022761
When will AMD have 22/44 cpus like Intel?

>>56022981
Its still the fastest CPU in single core performance available

>>56023923
>did
Still does. Difference is they're open about it.

File: 1470611346961.png (577KB, 480x476px) Image search: [Google]

577KB, 480x476px

>CPUs made by gf
>GPUs made by gf

No thanks, I'll pass.
It's an abstract kind of hell.

File: 1467252211583.jpg (15KB, 480x354px) Image search: [Google]

15KB, 480x354px

>>56022722

>>56025093
>Wanting chink made CPUs and GPUs
Wew lad

>>>/v/

File: kill_niggers3.jpg (59KB, 616x834px)

59KB, 616x834px

That's pretty impressive assuming it's ture.

File: MY FUCKING ZEN IS GONE FOR GOOD NOW.png (57KB, 1174x471px) Image search: [Google]

57KB, 1174x471px

>using CPUs made in a fab in which "qual"comm SoCs are made

>>56023081
ES does matter, actually. I have an ES Xeon that was never released to the market. There was a variant with lower clock speed and TDP released, and a variant with a higher clock speed and TDP. So just because an ES performs at a certain level doesn't mean that particular processor will ever make it to the market.

The release processors might have higher clocks/TDP with the same chip.

File: CoVe-2IUIAAzQ9P_1.jpg (29KB, 400x286px) Image search: [Google]

29KB, 400x286px

>>56025168
Hallo Zuckerschnütchen

>2016
>still believing in Zen vaporware

LOL

File: 3046688-8311357204-AMD-Z.png (104KB, 1094x745px) Image search: [Google]

104KB, 1094x745px

INTELFAGS BTFO!!!
#AMDMASTERRACE

>>56022798
Is any one aware that the bulldozer and piledriver are sold on the old notion that overclockers will buy them and oc them 30%+ over stock. just like the k7 line.

Does the performance scale linearly with clock speed?

>>56025256
I want to be leaf

>>56022644
>implying I cannot appreciate it if a brand I don't use makes a comeback
fanboys need to die

>>56025060
You mean like their 32/64 Naples Zen?

>>56025450
You gotta justify your purchase somehow. If you buy a CPU from Intel, you have no choice but to shit on CPUs made by other brands. It's not possible to be content with what you bought unless you also believe everything else is also shit.

>>56024162
I can get 1150 stable. 1200 CAN run, but not with all games. Bf4 is notorious for not liking bad OC's

>>56025604
>>56025604
Sorry, meant to say 1200 stable and 1250 can run, but not stable. Not at home right now

>>56023637
It's never too late. :^)

>>56025589
If you buy a cpu from intel, it is from that point on, your duty to put any amdrone in their place if they get out of line.

File: wccftech.jpg (52KB, 400x400px) Image search: [Google]

52KB, 400x400px

>>56022644

File: 14629405219.jpg (38KB, 236x272px) Image search: [Google]

38KB, 236x272px

>>56022644
>4790
>5 year old cpu

you know that we currently are at the edge of 7th generation?

File: A.png (93KB, 193x260px) Image search: [Google]

93KB, 193x260px

>15% faster than a 4790
????

>>56022644
I don't care about your shitty game benchmarks. Give me a Dolphin benchmark or fuck off.

>>56025730
>shitty game benchmarks
>wants a different game benchmark
>>>/v/

>>56025730
nobody gives a shit.

>>56022644
>ashes of the singularity
>every single fucking time

Do these idiots realize nobody actualy plays this game?

Yeah we know you guys optimized it for AMD hardware

>>56025229
Guten Arbend

>>56023880
>They are both purportedly clocked at the same rates and the only difference that we can see is the fact that the 2D variant fares much worse in terms of performance.

>>56025774
I didn't even know it was a game until recently. Thought it was just some tech demo or benchmark like 3DMark and AquaMark.

>>56022761
>>56025060

>have over 9000 cores
>90% of workload still done on Core 0.

>>56025842
t. gamer

I just hope they don't lock their CPUs like jewtel. I just want something cheap to OC for fun.

>>56025706
Wtf are you talking about the 4790 is 2-3 years old. Are you a newfag or something?

>>56022644

Is AOTS the only game AMD people care about? Thats like the only game they show the bench for.

I know it is because of the DX12 and other dildoshit, but come on. You gotta show you can handle everything, not just one half-new shit.

File: i5 4690 vs i5 6660k.png (175KB, 815x879px) Image search: [Google]

175KB, 815x879px

Meanwhile in Haswell vs Skylake land.

Intel drones seem to think being "2 gens behind" is somehow important when their processors haven't improved in those 2 years outside of I/O

>>56026160
>no legend

>>56026160
This. I still use inferno bridge, and seeing people upgrade from Haswell to Skylake really makes me cringe.

File: Screenshot from 2016-08-11 14-49-14.png (52KB, 707x415px)

52KB, 707x415px

>>56026178

>>56026160
I'm surprised to see Alien Isolation used as a benchmark. I played that with a GTX 660. Even that card managed 60 FPS.

>>56026160
Speaking of I/O, I hear Summit Ridge will have 36 lanes.

>>56026190
Skylake also has that new scheduler that only works on Windows and gives around 10% more performance.
Guess how that's working out.
Or is that only on low power shit?

>>56022761
Ah the usual "amds x in the future will beat everything"


Yeah sure

>>56026344
Yep, for amd the future can't come quick enough and somehow they still manage to never deliver.

File: 1454864049316.jpg (36KB, 640x360px) Image search: [Google]

36KB, 640x360px

>>56022644
>Ashes Of The Singularity

>>56026193
That might very well be why it's used as a CPU benchmark. You need to remove the GPU bottleneck to adequately benchmark the CPU, which is why you often see games benchmarked at 720p or less.

File: 1437977320825.gif (2MB, 352x217px) Image search: [Google]

2MB, 352x217px

>Benchmarking CPU's on games
>Juging the CPU by it while 90% of the computational workload is rendering
This is fucking retarded, i know that /g/ is retarded buy god fucking damn it.

>>56023873
Whatever man. My Rage 128 was pretty baller back in the day.

>>56022722
are you the same anon that said once

"consoles are better than PC because the CPU's in there are for gaming"

>>56025842
ARMA 3 FUCK YOUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUUU

the RX taught me to postpone all hype until it actually launches, is in stock and i see the EU price

>>56022644
pretty good
D stands for dual core

>>56025288
As someone currently on a 4.9Ghz piledriver I'm well aware of it.

300 fucking watt TDP, the housefire is real

https://forums.anandtech.com/threads/first-summit-ridge-zen-benchmarks.2482739/

Fake benchmarks, Zen is only up to Nehalem levels of performance

AYYMD IS FINISHED & BANKRUPT

AYYMDPOORFAGS ON SUICIDE WATCH

>>56022644
Into the trash it goes... i still use some bad threaded programs and amd was using 8c/16t to beat a 4c/4t processor.

>>56022644
>AMD processors new benchmark is it's not completely shit compared to Intel's 3 year old stuff that they don't even make anymore

How the mighty have fallen.

>>56022839
what does she do?
feminist things?

>>56028023
She makes........ more money than you

>AMD
>New CPUs run as fast as 3 year old Intel CPUs
>INTEL BTFO
>New GPUs run as fast as 2 year old Nvidia GPUs
>NVIDIA BTFO!

Daily reminder that you only buy AMD products to keep up with 3 years ago tech.

>>56023863
>Titan X is squarely a gaming card
This is wrong. Titan X is a workstation card in the same manner that the original Titan was one, and the same goes now for the Pascal based Titan X.
Sure, people loved to use Titan X in their gayming/benchmarking rigs, but this doesn't make it any less of a poor man's Quadro.

File: Untitled.png (138KB, 1423x1138px) Image search: [Google]

138KB, 1423x1138px

>>56025256
That's hilarious. I want to see the AES scores.

Or basically any function that performs better on a GPU. Intel couldn't build a GPU, or for that matter a CPU that's capable of doing GPU like work if there lives depended on it.

In before muh integrated graphics.

>>56028141
>Cinebench
Reported

>>56028081
>>AMD
>>New CPUs run as fast as 3 year old Intel CPUs
>Intel
>New CPUs run as fast as 3 year old Intel CPUs
>PROGRESS! NEW SOCKET!
Yeah, awesome.

>>56028049
kek rekt
is she a qt? I will apply for job at Intel if women qt

>>56028294
>job at intel
>women qt
top kek fagot

>>56023191
DELETE THIS

>>56028362
>wanting qts
>fagot
srsly neck urslf famm

>>56024162
>>56025604
>>56025633
Maybe with LN2

>>56023948
just undervolt it yourself.

Let's be honest here, if Zen doesn't come in October it's dead.
I'm getting a new rig for Christmas, like most people.

>>56023392
It was pushed back. There aren't even any AM4 mobos yes, and the mobos usually appear 4-6 months before the CPU release. Which makes Q2 2017 a more plausible release date.

>>56029074
>and the mobos usually appear 4-6 months before the CPU release
What? Where the ungodly fuck did you get this idea from?

>>56029103
It was always like this. AM3+ mobos appeared almost a year before the FX. LGA 1151 appeared 4 months before Skylake. etc etc.
There should be prototypes at computex this year if Zen was meant to come in 2016. There still aren't any. Not even engineering samples.

Zen comes in Q2 2017, mark my words.

>>56022644
what is a standard aots

>>56028466
Nice meme stay mad cuck

>>56022722
summer /g/, everybody

>>56029146
LGA 1151 was launched august 5th 2015, the same day as skylake. I can't be arsed look up the release dates of AM3+ but something tells me you pulled that out of your arse too.
>Zen comes in Q2 2017, mark my words.
I'm calling Q1. Bristol Ridge, which will use the same motherboard, will be the end of this year.

>>56022644
If it's able to overclock well, then I might buy it

>>56022952
I mean literally anything worth playing is dx11, there's what 14 current dx12 games? None of which are even relevant.

>>56029252
>LGA 1151 was launched august 5th 2015
LGA 1151 was at the 2015 computex.

I know for a fact taht AM3+ appeared a year before FX, because I bought one for a Phenom II and upgraded to an FX when it came out.

Stay mad, Zen isn't coming any time soon and AMD is finished.

>>56022644
but anon that benchmark shows zen is 15% slower than a 4790.

>>56029280
>LGA 1151 was at the 2015 computex.
It was launch the exact same day as Skylake.
https://en.wikipedia.org/wiki/Skylake_(microarchitecture)
https://en.wikipedia.org/wiki/LGA_1151

I don't care what you think you know, you're pulling shit out of your arse with the whole "Motherboards come before CPUs!" bullshit.

>>56029357
I think the Intel CPU was OC'ed. That's why he said that

>>56029360
There is a difference between announcement and consumer release.
It was announced an computex and we knew the exact chipset lineup long before release.

I'm getting a 7xxx haswell-e this year and you may keep waiting for your outdated architecture to come out in a year or so.

>>56029280
yeah am3+ boards were around in january of 2011

>>56023394
Phenom was a fucking joke.

Phenom 2 was good, on par with the last gen Core 2s but clocked higher and had more cores. Unfortunately, it was up against Nehalem and Westmere.

>>56029373
they don't generally overclock a processor in benchmarks without overclocking the competitors as well, it wasn't overclocked.

File: it wasn't though.png (63KB, 752x810px) Image search: [Google]

63KB, 752x810px

>>56029373
think again.

>>56022675
Good. Kabby and Coffee Lake a shit

>>56029401
From the wiki article
>An initial batch of Skylake CPUs (6600K and 6700K) was announced for immediate availability during the Gamescom on August 5, 2015
Now go check when the LGA 1151 motherboards were launched. Same date. You claimed that the motherboards came four months before the chip. Will you acknowledge that you are wrong or will you keep being a muppet?

>>56022883
>the newest possible cpu beats an older cpu

Colored me shocked.

>>56026251
>Skylake also has that new scheduler
sauce?

File: PState.jpg (730KB, 3000x1537px) Image search: [Google]

730KB, 3000x1537px

>>56029524
Not a scheduler, a new P-state

>>56022722
>What is microcode
>What is cache
>What is bus speed

File: 1462117971891.jpg (84KB, 371x371px) Image search: [Google]

84KB, 371x371px

I don't know what to say. I was expecting AMD to bring something somewhat okay to the table (sandry bridge) but I never expected this.

Might have to ditch my i5 4430 for AMD now. If that 8-core Zen chip is reasonably priced then I'm definitely getting that.

>>56022644
>AMD processors just caught up to years ago for Intel

Neat, and in 2018 when they catch up to Skylake I'll have a Coffee Lake, mainstream hexacore Intel processor and the only "advantage" AMD had will be gone in it's CPU division, unless they actually bother to do something groundbreaking or game-changing instead of playing catch up.

>>56029614
It will be reasonable, compared to Intel's 6 and 8 core chips but I think it will probably be 4c/8t i7 prices or higher.

>>56022644
>currytech
Get the fuck out and stop overhyping zen until we have more reliable sources, you faggots did the same shit with the RX 480.

>>56022644
>2.8
>slower than a stock Haslel 4790
This is either amazing or shit. If it can hit the same clock speeds when OC'd that Intel can (4.5+ GHz) it's going to be amazing, if not it looks quite underwhelming. I'm assuming that's a 4C/8T model, if it's 6C/12T or 8C/16T and it still can't beat a 4790 it's fucking abysmal.

>AotS
Whats so special about this shitty game?

>>56030239
It's heavily optimized for amd hardware. Somehow an fx 8350 matches a 6700k in this game and I doubt it's solely due to dx12.

>>56030239
It's a benchmark made for ayyymd gpus.

>>56022801
No. Scorpio is confirmed to be using Zen based on early leaks and performance metrics. The neo is looking like utter garbage though because it's still using the same jaguar cores that are in the ps4. The ps4 neo is basically using a much faster polaris gpu but with the same shitty jaguar processor that already bottlenecks the shitty gpu in the ps4. The xbox scorpio is an upgrade to both gpu and cpu by utilising new vega and zen hardware.

File: 1467178637980.jpg (118KB, 796x804px) Image search: [Google]

118KB, 796x804px

>>56029277
>there's only 1 dx12 game to play
>there's only 4 dx12 games to play
>there's only 7 dx12 games to play
>there's only 12 dx12 games to play
Gee the number keeps getting higher but Nvidiots just spout the same bullshit
>t-they all suck anyways!!!
Kek
Pic related
You

>>56030415
Not him but are you seriously defending 14 games (10 being shit garbage) over thousands of dx11 games? You'd have to be a retard to buy something right now which has sub par performance in dx11 compared to the competition when literally all the major games and goty titles are all dx11. You're buying into the dx12 shit years too early when it's not even close to being the industry standard.

>>56030542
Triple A titles announced will be dx12 only by m$
Poorfags who want their cards to last years buy cards in this price range anyway. There's still people with 7xxx cards and prior
I think it'll be a good investment for those people

File: 1470506437192.jpg (30KB, 489x426px) Image search: [Google]

30KB, 489x426px

>>56024070
>1060 w/o hardware scheduler @ 120W
>RX 480 w/ hardware scheduler @ 150W
Volta's perf per wat will be VERY interesting

>>56031058
What makes you think Nvidia wasn't making a low power scheduler in the meantime? It happened with x86 decoders a decade ago, when moving from complex to simple decoders.

>>56030542
>thousands of dx11 games
There's not even 250 DX11 games

>>56022828
It gets shit on by a 4790, how does it expect to compete with stronger CPUs?

>gaymen
Why the fuck is this thread in /g/? Take this bullshit to /v/. Hell take it to /vg/ since you niggers seem to love posting this thread over and over.

>>56030333
honestly, one you take away shitty use of threading and multiple cpus, amd's cpus match intel's

not saying amds jaguar are good but the gap when you have low level access is far FAR less and the gpu holds it back more so then the cpu.

>>56022722
That's wrong.

Clock speeds per cycle and the number of cycles required to complete specific instructions differ between architectures.

An example of this can be traced back to the prime days of bitcoin mining with GPUs. At the time, ATI cards used around half as many cycles to perform the same integer operations as Nvidia cards. This made them vastly better for hashing and made them preferred for mining. The nvidia cards in the same price range traded blows for gaming and general tasks, but for this odd thing, ATI was king.

>>56023603
the 480 8gb is close to price performance king, the 480 4gb is price performance king, possibly the 470 now is price performance king.

>>56023481
Its not bad, but its not good either... its middle of the road but thats really all you need to be in a genre that is all but dead.

>>56023735
oh shit, if it used workstation drivers, that was a pretty fucking cheap card considering what it can do.

remember, you only need high memory when it comes to cad software, what is used to make games can get by on far less.

also, the furys in that thing are so heavily binned that it uses the same amount of power as a single fury x.

>>56028127
>Sure, people loved to use Titan X in their gayming/benchmarking rigs, but this doesn't make it any less of a poor man's Quadro.

no, the original titan wasn't a gutted for gaming card, every titan after the titan/z/black has been gutted for gaming, it may be helpful when it has more ram in pro hardware, but it doesn't have the drivers for real pro software and it doesn't have the ability to really step up in other areas either.

>>56025051
for 60fps, amd or intel will get you there
for 120/144 amd stumbles more then intel, and intel only gets 144 is its 4000 or better mainstream or EE for 2000 or 3000

granted, just running the games even my ageing 955 is able to do that.

>>56025774
>>56025832
Its the first real dx 12 game, everything to that point was a shitty conversion or wrapper so this is why its THE new api benchmark, much like doom will be for years THE vulcan benchmark.

>>56023603
>$200 480 same price as $250 1060
>$200 480 more expensive than $350 970

Your inflated prices are almost are retarded as you and they do not reflect the marked.

Oh for crying out loud. This is the 480 all over again.
>NVIDIA IS BANKRUPT AND FINISHED KEK
Now it's
>INTEL IS BANKRUPT AND FINISHED
Meanwhile Nvidia is making bank like a Jewess celebrating barmitzva or whatever the fuck.
Zen will perform around 2500K levels and be around the same price. Anybody who wants performance will still buy Intel, poorfags on a small budget will get Zen or a 2500K.

>>56031132
>what are indie titles

>>56029759
This game doesn't do higher threads well,

game was compared to a quad core i7 to a 16 core xeon with the only difference being a 200mhz clock speed difference, the results were almost fucking identical.

What i'm taking away from this is an engineering sample is almost the same ipc if not more then intel's 4790, and in that case its going to sit between it and a skylake

then there is the 4 more cores 8 more threads,
and to top it off, the potential of amd not being shitty with their prices, please note, last time amd has a performance advantage, they put out a cpu that was beating intel's extreme editions dor 1/3 the price, and then priced their 40% better then that cpus at 1000$, given the fury x had a range from 550-1100$ it set itself at 650 (do you really think nvidia set the 980ti at 650 if they had an option to price match a higher costing amd card? the company that did the founder editions?) so there is a great chance this cpu came sout at around i7 prices if not lower.

>>56031199

how is it getting shit on by a 4790? its outperforming it.

read the fucking numbers goofball

>>56031229
No. You have no idea of what you're talking about. It's common knowledge that the current consoles are being held back by their cpu even with low level developer access. Just look at hitman on console which was supposed to be build from the ground up for gcn and runs well on amd desktop cpu. It runs at 20 fps in indoor areas on both consoles due to cpu bottleneck.

At least Microsoft are trying to fix the issue. Sony are trying to sell everyone the same garbage with an overclocked tablet cpu.

>>56031658
No, You have no idea what an outlier is. most games, and my most i mean damn near every game that comes out, especially first part or exclusives, are gpu bound not cpu.

oh, wait, I just looked up someone benching the game on a shittier cpu then the ps4 and with about the same gpu, it seems that at sub max details, the gpu is the bottleneck, as going from high to medium would not result in a higher minimum if that was the case

https://www.youtube.com/watch?v=OJJyo5Q5_9A

a typical 'we want out console screenshots to look the best it possibly can and fuck the frame rates' that you see time and time again, i wrote above the youtube link before i looked for it, i retract everything i said above it, hitman is pushing the gpu hard not the cpu.

>All this arguing over AMD Zen.

Feels good to not have played the waiting game and gone from the fx-9370 @ 5.2ghz to an i7-6700k @ 4.6ghz with stock voltage. Like bight and day. Everyone can keep playing the waiting game. I'm already comfy.

>>56031769
digital foundry did an exact test with the ps4 and xbox one and found hitman to be massively cpu bottlenecked and the xbox one actually takes the lead over the ps4 due to the overclocked cpu in the xbox.

>>56031979
Not him but
>One isolated case out of all games released so far
Very cool but keep in mind that every other game so far is GPU bottlenecked and runs better on the pisspoor than on the exbone.

>Direct x game
>game
>Arguing about game engine thread utilization
>GAMES

I ctrl+F'd the thread and as expected there was not a single mention of compression, cryptographic, video editing or server. Zen is not a gaymen cpu, it's meant to go against intel's midrange xeon line.

File: 1470851655137.jpg (138KB, 653x726px) Image search: [Google]

138KB, 653x726px

>>56022644
Where was I? Installing my new shiny blazing fast 4790k.

>>56030333
>The ps4 neo is basically using a much faster polaris gpu
polaris clocked at like 900 is 4.1 tflops, nothing faster about it.

>>56022936
As of right now, I believe it's the closest to DX12's true capabilities. Especially when compared to games like Tomb Raider that feature heavily cut down DX12. For a while it was the only game with explicit multi-adapter functions.

File: you don't say.jpg (17KB, 600x373px) Image search: [Google]

17KB, 600x373px

>>56022883
>late 2016 cpus beat late 2015 ones

I'd fucking hope so.

>>56031769
>games are designed to not require a good cpu because the cpu in consoles is bad
>a better cpu is not necessary because games don't require a good cpu

>>56032185
Probably no mention because of there's no benchmarks of anything else idiot, thanks for useless comment. Zen is marketed as a desktop part btw.

>>56032316
You know I know that image is a joke but no parents could ever feel the death of their kids as a good thing. Even if you waste your life on chinese cartoons and vydia they'll be forever traumatized if you end yourself.

http://www.liveleak.com/view?i=7c9_1442749005

Even if your parents hate you for wasting your life they still love you deep down. Link related, the father taunting his son probably ended up fucking batshit insane after what he did. He may have hated his son for wasting his life but that he never truly hated him.

Anyway sorry for this blogpost, it's just I can't get over the fact that a cousin of mine fucking hung himself in a forest nearby last month. Not sure what happened with him, he seemed mostly happy and had a job. His parents constantly zone out when I see them. Doesn't help that was their only kid they had and now they're both like 50.

File: hp9x4.jpg (46KB, 460x348px) Image search: [Google]

46KB, 460x348px

>>56032866

>>56025084
What

>>56023300
>it clearly shows zen performing worse in single threaded
What it shows is AMD's new 8 core design being ~45% faster than their old 8 core design, just as they claimed.

>>56029433
Until Zen releases they're still competing against Westmere
4.8Ghz piledrivers are pretty much equal to 4.4Ghz 6 core westies

You are trusting the same website that claimed the 480 will bet the 980 and come close to a 1070

>>56031608

Shitty Greenlight pre-alpha piles of shit.

>>56030415
they all do fucking suck, and even those that don't have legacy support with no negligible difference in performance. defend your shitty purchase all you want but the industry standard is still going to be dx11 with dx12 support. companies want money from the millions who didn't switch to windows 10 just as they want the money from the idiots that did.

>>56030625
triple a titles will never be dx12 only, hell most games still have dx9 support. companies want windows 10 hold out money just as much.

>>56025832
it essentially is, they literally made a game with no plot or story for the soul purpose of, "as many draw calls as we can."

>>56035464
>defend your shitty purchase
But I have dual 980 tis lol