If zen is coming out at the end of this year or early next year, will it be able to compete with Intel's single thread performance?
Will AMD have to resort to more cores to compete?
Dual core Pentium level processors?
Tri core cheaper than i3?
Quad core slightly more expensive than i3?
Hex core i5 competitor?
Octa core i7 competitor?
What are you expecting from zen /g/, will you buy one?
How about the APUs? What sort of proving can we expect?
Maybe the Pentium competitors will be budget APUs?
Maybe a tri core APU at i3 prices? Or skip the tri core and go for a quad core APU at i5 prices.
How will they decide which processors get hyperthreading? How much extra will they charge for it?
>>52180454
This confuses me
>>52180463
Shut the fuck up, retarded mac fag.
>>>/fa/
Zen is supposedly coming at the end of this year and Zen APUs come sometime in early 2017.
The current generation bulldozer chips are actually pretty competitive with Intel's lower end chips, i5 and above there's nothing to compete there. Zen is supposed to be a 40% improvement over excavator so we'll hopefully see something more in line with sandy bridge. At that level Intel's high end would still be a little better but the situation would be much better overall. They could at least compete on price with performance like that.
>>52180578
If zen does use more cores than Intel then even the high end won't be able to compete. If zen is 10% faster, just adding one extra core would be enough to give Intel a run for its money.
Especially if they're priced more competitively than Intel.
>>52181222
If zen is 10% slower*
If Zen doesn't deliver then x86 is dead. The 8350 came out a couple years ago and Intel is only giving us 1% performance increases. Moore's law is still in effect for ARM so it's just a matter of time.
>>52180578
>something more in line with sandy bridge
It's over, intel is finished
>>52181263
You sir are a faggot. If Zen fails it isn't gonna be the end of x86.
I do think that Zen is gonna fail if it is priced nicely I might pick it up
Zen will be like Fury: a disappointment
>>52179967
I thought zen would be available mid year, I'm mad now.
>waiting until 2017 AMD goofed
>i5 will be disgustingly cheap
>>52181505
How is fury disappointing? It's exactly where it belongs in price/performance and some can be unlocked to full fury x.
>>52179967
If they roll out a CPU of above i3 level that has more cores, can be overclocked, AND CROSSFIRES WITH GPU that will be a success.
>>52181376
because of lack of competition in market CPU's weren't advancing one bit for last several years.
I've been nursing along this Phenom II for a while now, I'm thinking I won't be able to wait until Zen. I'm thinking I'll build a new PC with an Intel CPU in the next month or so and see how Zen pans out. If it's decent I guess I can sell my CPU and motherboard to hop back on the AMD train.
I really want AMD to stay in business, it's pretty autistic but an Intel that only has to think about competing with ARM is no good.
>>52181696
>AND CROSSFIRES WITH GPU
What are you on about? Are you talking about APUs?
I don't see that happening. I hope they keep APUs and CPUs separate unlike Intel. I have no interest in paying for an integrated GPU I'm not going to use.
Skylake has almost double the IPC of Piledriver.
Zen is expected to be a 60-70% improvement over Piledriver.
This would put Zen at 80% of Skylake's IPC, but with double the cores and potentially higher clock rates.
Basically Intel is bankrupt and finished.
>>52181674
>NVIDIA HAS NO HBM, ITS OVER
>LOW POWER CONSUPTION
>THE OVERCLOKER DREAM
>FURY WILL BLOW TITAN X
>FURY WILL EXPLOSE THE 980TI
>FURY NANO WILL BE THE BEST PRICE/PERFORMACE RATIO
![snapshot_06.50_[2015.12.24_02.15.19].png](https://i.imgur.com/zUe7ifT.jpg "snapshot_06.50_[2015.12.24_02.15.19].png")
>>52180454
what's up with these newfaggots
>>52181801
The enthusiast CPU coming out using Zen cores is Summit Ridge. Its an 8 core/16 thread chip competing against the i7E line. It is not competing against higher clocked mainstream i7s and i5s.
That 8 core chip absolutely is not going to have high clocks, it'll probably be pinned around 3ghz.
Dipshits like you who continually post this kind of nonsense do more damage to AMD than their own incompetent marketing department ever could.
>>52181823
No one said fury will blow the Titan X.
Nano was hyped by the community, not AMD.
>overclocker dream
Never recall AMD saying that.
Are you okay? You seem a little delusional.
>>52181970
>Trying to predict clockrates 10 months in the future
>>52181376
>You sir
lel +3 reddit gold my good sir! Kappa Keepo Biblethumb
[spoiler]Only le reddit friends will understand this post, lel[/spoiler]
>>52182005
>No one said fury will blow the Titan X
>>52182032
>being a tech illiterate shit eating retard
Power per core, dipshit. Look at intel's 14nm parts and draw a comparison.
Summit Ridge has a 95w TDP, they aren't going to have high clocks like the Bulldozer family.
>>52179967
if Zen wont be able to compete with i7 it will fail, this is just the fact of the situation.
>>52182005
Calm down m8, the guy is right. Before it was released, I was reading almost daily that it'll blow the titan x out, HBM being the big deal and apparent high overclocking possibilities duo to it being a hybrid card.
Call those people retards but do not try to deny that things like that weren't posted here.
>>52182491
So you're saying the fury was shilled to btfo the Titan x?
Do you even read what you're writing?
>>52179967
current amd spcs put it 40% over their current laptop ipc, which each iteration sense piledriver has gotten about 20% better, so that places it within 2-5% higher or lower than haswell, and the average ipc jump from haswell to skylake was about 2% with some applications reaching 5% and one VERY specific one hitting 10%
so there you have it, amd could be within spitting distance or matching of intels ipc, but thats not whats interesting.
what's interesting is the 8 core 16 thread at mainstream prices, as they are releasing it to replace the 300-350$ sku.
intel at that price may have single core, but amd would have at least double their multicore.
this isn't going to be a "more cores makes out chip kindof the same or better" its going to be "you don't notice the single core difference without in depth benchmarking tools, but you do notice the multicore performance without"
>>52181648
still not confirmed i believe.
>>52182116
zen should beat out non E i7's and even than, beat out non 8core E's
if the new rumor is true, 12 core amd will beat out intels 10 core just by virtue of amd is close enough single core to not matter, and multicore amd will have 2core/4thread more.
it will all come down to price, and amd is putting the 8 core 16 thread out at 300-350$ if the sku replacement is correct (which is a amd supplies slide)
>>52181245
even if it was just 10% faster than the excavator it would give intel a run for their money muilticore and be good enough single core, current piledrivers can compete multicore and win against quad core intels.
>>52182005
>>52181823
nvidia fucked amd by getting wind of the fury power and price and they cut down the titan x to compete with it, and they still lose out when it comes to 1440 and 4k
which one is jim keller working on? that's the one im gonna buy.
>>52182870
kaby lake
AMD cannot pull "HURR DURR CORES" like they did in the Bulldozer generation and win. First of all, few things use more than 2 or especiallly 4 cores. Secondly, at least up until pre-Zen architecture, AMD "cores" are not the same thing as Intel "cores" - the latter being a lot more powerful/useful plus have things like Hyperthreading and the like. Single (or dual, or quad) threaded performance matters and so does the amount of power/heat it puts out. An Intel 5960X has 8 full on boner physical cores that can run as 16 logical, can be overclocked to each of those running at over 4 ghz easily. Yeah, that's an extreme example but Zen has to at least come close to Intel Mainstream Skylake 6600K and 6700K performance (at a lot cheaper) t be worthwhile. Likewise, KabyLake or whatever else is mainstream when they launch. They can't afford to have some 12+ core monstrosity that yet can't game as well as Intel's mainstream high end 4 core platform.
>>52182093
something you may not know, intel measures their process differently than everyone else from my understanding, they are on the 16nm process but instead of measuring the outside they measure the inside.
intel's 91w draw cpus are 4ghz
wonder how far you could easily push an overclock with amd, intel can usually hit at least 4.6
>>52182830
> which each iteration sense piledriver has gotten about 20% better
Not even remotely close.
>what's interesting is the 8 core 16 thread at mainstream prices, as they are releasing it to replace the 300-350$ sku.
This isn't the least bit factual, so you shouldn't be spreading such wanton misinformation. Su publicly stated that they didn't want AMD to be considered the budget brand, and Summit Ridge is competing against intel's much more expensive i7 Extreme line.
Stop talking out of your ass.
>>52182917
DX12
>hypetrain.jpeg
>>52182830
This sounds like they're about to repeat the same "single thread doesn't matter, it will be good enough" mistake they made already. Not promising. Each one of those cores isn't close enough to Intel's to be worthwhile, and even if they have 2 extra or whatever it won't mean they '"beat" anything as history has showed us. Thats what they promised with Bulldozer/Excavator etc... "Look at this 8 core thing that will crush all the 4 and 6 core stuff because MOAR CORES ". Didn't work then, won't work now.
>>52182930
yea, amd slides showing they are replacing the 8350 with "8 core zen" isn't reason to hope for 300-350, that skus initial price.
and after they said they didn't want to be the budget brand, the fury comes in at 650, what every non retarded person thought it would be at least, this is fueling hope for the 300-350 range.
where you get that icp from, from what i remember last time was each iteration of the piledriver was damn near 20% each, but 20% of shitty is still shitty.
>>52182984
from what i remember, its haswell performance at least and skylake barely moved the ball ahead, so those extra cores are what separate a they are the same performance wise to why would you chose intel
>>52183010
from what i remember dx12 scales better but its not a fucking miracle worker you will get full 4 cores but after that its depreciating returns.
>>52183033
Its the halo product of their CPU line, that has literally nothing to do with pricing. Jesus christ you should be killed for being so stupid.
Piledriver was only roughly 10% higher IPC over Bulldozer, the rest of its performance uplift came from higher clocks. Even the Anandtech writeup covered this.
Stop posting on this board. Permanently.
>>52182870
His team worked on Zen
Ehh, I'll prolly pick up an 8350. Keep/get it to 4 ghz, and just have a really nice Plex server in my basement.
>>52183070
There wasn't really a "his team" he was head of the entire department. The actual lead designer on Zen was Suzanne Plummer.
Keller was the head boss, the "big boss" if you will, of everybody working on both Zen and K12, though according to Su he personally worked on Excavator first thing when he rejoined the company.
>>52183070
gonna keep my eye out for zen benchmarks then
>>52183164
hopefully by having him in charge lead to some good decisions
>>52182662
But it's true, Fury X was a very disappointment. Seriously, fuck the AMD pre launch marketing
>>52183271
>Fury X offers Titan X performance for half price
>Nobody remembers because Nvidia paperlaunched 980ti 2 weeks earlier
>>52183271
>But it's true, Fury X was a very disappointment. Seriously, fuck the AMD pre launch marketing
how was it a disappointment? remember with out it nvidia would never have bothered with a 980ti
>>52183414
everyone remember, they are just shilling nvidia hard.
>>52183414
>Fury X offers Titan X performance for half price
What is the 980ti? Also they release the car before
>>52181222
Intel already has 16 core processors. 32 if you count hyperthreads.
>>52183501
Those are Xeons, server chips.
Summit Ridge is not a server chip, there will be Zen based Opterons to compete against Xeons.
>>52183414
>Fury X offers Titan X performance for half price
Every custom ti are faster than the titan x. Fuxy are good only in CF
>>52183577
how many people do you think returned their titan x after they realized how ripped off they got when nvidia released 980ti ?
>>52182920
Intel's measurements smaller than the standard way of measuring. The 22nm node was something like 25 or 28nm.
Even then, 99.9% of consumers don't give a shit about how small transistors are on their si.
>>52183643
Dunno, same shit happened with the 1st Titan vs 780ti. Titanfags are cuck
>>52183164
Keller tried to salvage what he could, aiming at the mobile market more less trying to crank out some half decent silicon for their ever dying mobile market.
He did a pretty good job, as seen by the 8320e/70e chips. They get some damn good overclocks on low power.
>>52183732
Keller had nothing to do with any Vishera chip, the later e variants were just better binnings. GloFo's 32nm PD-SOI process had major issues that took years to sort out.
>>52183271
Fury, or fury x?
Cause the first guy definitely wasn't talking about the fury x.
>inb4 he forgot the x
He didn't forget it on the Titan x.
>>52183848
Fiji (x, cut, nano)
>>52183414
This. While people were speculating way too much before the release given that we knew the core amount and its relative performance to the 290x, we all knew it was a beast, but you should really not fall into the hype so easily. Nvidia buttfucked everyone: 980 users, Titan X users, and most of all Fury X users with the 980Ti. Nvidia knew and they had to be swift about it so they did, the absolute madmen.
>>52183769
Whoops, my bad.
What he did with the mobile silicon was still impressive in my opinion.
>>52183887
>>FURY
>>FURY
>>FURY
>>52183530
the zen is a server chip, at least first run will be.
>>52183660
that's not the point, i was just stating that you can compare directly to intel rather than changing the math for "larger" process
>>52183925
considering i don't want the taste of cock in my mouth, i would still have gone with a fury or fury x if i was in the market for a new gpu... nvidia is doing their damndest to kill gaming and i refuse to help them.
>>52183933
You quote the wrong guy, btw It's probably the x and he is right...
>>52184170
>fury x
He was pretty specific about it being the fury though. The only x I saw was for the Titan X
>>52184163
There is no such thing as "the zen."
Don't try to discuss something when you don't have command over basic facts.
>>52182920
This is a complete embarrassment of a post.
You demonstrably have no idea what a process node's marketing name refers to. It is not an exact measurement of anything. It is following ASML guidelines.
Every single word you just posted was pulled entirely from your ass. You're woefully ignorant, and you need to stop posting.
>>52184214
>>52179967
I will save my money for late this year. I will be building an all new (or at least mobo + CPU) computer anyway since neither Zen or any intel offering will work on my current board. If Zen is better, or at least good price/performance, I will go with it. I have a moderately strong preference for AMD, but if it isn't a good performance value, than I might end up with my first home built Intel machine. I have 3 custom built rigs (one wasn't built by me, though, from late 1999) and all have AMD CPUs. This will be interesting for sure.
>>52181674
>it's exactly where it belongs in price/performance
I'm surprised at how accurate this is.
>>52183577
AMD forced Nvidia to kill the titan x with the 980ti, sounds like a success to me
>>52179967
Will we see any updates to the am1 socket?
>>52184887
We're going to see some carrizo am4 processors soon, before zen.
Hopefully they're going to be cheap and upgradable to zen.
it's probably going to be targeted for server use
>>52182830
>laptop
Will AMD try to go full force for laptop CPUs as well. I am disappointed that I can't have a laptop with a good AMD GPU and CPU anymore. The A10 is the best shit they can push out on laptops now and it is not cutting it. I remember buying a laptop with an A8-4500M a few years ago and it was a fucking joke. I hate that I was forced to go Intel + Nvidia for a decent laptop with dedicated graphics since I am a hardcore AMD fanboy. If Zen turns out to be not a good value (and AMD is usually a good value, if nothing else compared to Intel) then I might buy an Intel rig. Either way I am saving up my shekels for the Zen release.
>>52185246
Imagine a zen laptop with hybrid crossfire and ddr4.
>>52185281
You don't understand my obscene brand loyalty. Ever since my first home built PC I have preferred AyyyMD. All three of my home built computers (even the Windows 98 one) have AMD processors. And I can't get upset and turn into a Macfag either since all of those have been Intel based for almost 10 years. There are no other options. The second great processor war is upon us.
>>52185376
You miss my point buddy.
A zen hybrid crossfire laptop should offer amazing value for the performance.
Especially since by then dx12 and hopefully vulcan will have taken off, reviving sfr and using multiadapter to improve multigpu performance.
Windows 10 might be a botnet, but it does offer better battery life and performance on portable devices.
>>52181801
RIP INTEL
I
P
I
N
T
E
L
>>52182917
>First of all, few things use more than 2 or especiallly 4 cores.
5 years ago called they want you back.
>>52181801
>Zen at 80% of Skylake's IPC
I don't know where you are getting these numbers, but I know that Skylake is about 20% faster than Sandy Bridge meaning that SB is about 83% of Skylake. They better not release something slower than Sandy.
>>52186528
http://www.anandtech.com/bench/product/1291?vs=288
Thats Kaveri vs Sandy Bridge, the Kaveri chip has a slightly higher case clock and turbo.
Excavator averages 5% higher IPC than the Steamroller cores used in Kaveri, though in someworkloads its 9%-13% faster. Zen is a 40% uplift over that. Multi core scaling will improve, and over all multi core performance will dramatically improve. A single Zen core has higher throughput than a module.
Zen at least will compete against Ivy Bridge and take some shots at Haswell per clock.
My next desktop will have Zen. Please don't go bankrupt AMD.
>>52186434
Other than cad, video/photo/audio editing etc there isn't much that is well promised for 4+ cores senpai.
>>52188060
>promised
Optimised
![wpjx9[1].jpg](https://i.imgur.com/lDCETEB.jpg "wpjx9[1].jpg")
>>52179967
if they cant get into laptops this time around they are gonna hurt for another decade
http://wccftech.com/xbox-one-may-be-getting-a-new-apu-based-on-amds-polaris-architecture/
So an APU with HBM? I don't really follow AMD news, but how much is an AMD APU? This looks really nice if they can keep the price inline with the i5-6600.
>>52181263
You should read what Moores law is about again
>>52188241
The general trend of AMD's apus is that they fucking MASSACRE 99% of options Intel has when doing anything gpu related. Since people buy SoC to get a bit of everything there is zero reason to buy Intel at the low end.
If Intel bribery wasn't so entrenched current apus would pull the carpet right out from under Intel in the laptop market.
>>52181696
I don't think you know what Crossfire is/does.
>>52188986
>>52181696
they tried hybrid crossfire people laughed at it
>>52188986
I don't think it's called crossfire but what he's referring to is combining performance of IGPU and a dedicated GPU which I think AMD is possible or will be possible.
>>52189020
>>52189014
It sucks for desktop but I think it has great potential for high performance budget laptops.
>>52184214
You need to relax.
>>52188241
I think A10s so far have debuted in the $150-160 range.
>>52189020
AMD is most likely going all-in with HSA when zen based apus reach market.
>>52189139
That's Australian prices.
An i5 4460 is $270 here for comparison.
And an i3 4160 is $160 iirc.
>>52189215
Is it even possible to buy an 8320e in criminal land? I've never seen it mentioned when kangaroos are posting.
>>52189227
To put it bluntly, no.
We don't even get the athlon 860k here and the 5350 is overpriced.
Only AMD worth getting here is an fx 6300 since its $10 cheaper than an i3 4160, but then you realise you paid the same amount as an i3 because the motherboards are $10 more expensive for AMD.
On the other hand, newegg ships here now so we can order it through them.
>>52185246
amd is not focusing on laptops or desktops, they are targeting servers with laptops and desktops being byproducts.
>>52186434
its sad that few things using more than 2 or 4 cores is still fucking true, amd gambled on multicore and lost the bet with bulldozer, and that was 5 years ago, and they were developing it for 5 years...
that said, you run more than one program at once, you will see a better overall performance out of you computer with more cores.
>>52188150
again, not even considering this as a worthwhile market at this point, they are targeting servers.
>>52189530
People who bought vishera must be feeling pretty comfy now that multithreading is taking off.
Good performance for the price back then.
Better now that programs take advantage of more threads.
Aged like fine wine.
>>52189829
>vishera
meh, i have run into more things that require high single core than high multicore... when multicore is concerned its not bad, it can be better than the entire intel like till the E series, but when you need single core you need single core and no amount of multi threading will fix that (unless someone figures out reverse hyper threading)
>>52181263
Moores law has nothing to do with power you stupid faggot.
>>52189895
You need more gHz
>>52189946
great job finding a benchmark that exclusively uses multicore programs.
>>52189960
These days for vidya vishera is highly competitive. Those few programs that are 1) single threaded only and 2) need as much resources as you can throw at it are fairly limited.
>>52189960
It's not though.
The fx only gets close to the stock i5 when it's at 4.5ghz.if it was a multithreaded benchmark the i5 would have been destroyed.
>>52190361
>>52190528
By my standards, that fx 4300 is beating the i3 4330.
Much smaller gap between min and max.
Also
>game benchmarks
Here we go...
>>52190631
The CoD benchmark is in favour of AMD actually. An 8350 costs less than the i5 in my corner of the world - enough of a difference to get a cooler capable of taking the chip to 4.5ghz or so (thus effectively a 9370 on the cheap). I would never suggest anyone purchase a 4300, but the 6300 and 8320(e) are fantastic for the money.
>>52179967
Will AMD even make it to launch? They havent had a competitive chip in the CPU or GPU market for almost a decade, fanbois can't keep them afloat forever.
>>52190739
This is bait.
>>52190697
An 8320 is about 50-60 dollars cheaper than a i5 4460 here.
I'm sure I could easily squeeze another 500mhz out of it on a cheap piece of shit motberboard.
It's the most futureproof budget processor available. They're even better at 4k than Intel when it comes to min fps, which is more important for a smooth experience.
>>52179967
If zen came out today it would still stuck.
>2016
>buying the multicore jew
>>52190817
bf and crysis are on engines that properly scale, no shit they properly scale, winzip is also a no fucking shit it scales properly.
as for black ops, i don't know what to make of it, we will know with one more game if the engine is properly optimized or if it was just that studio that made the shit work.
and with gta5, rockstar makes so few games and their engine isn't licensed out as far as i know so we are looking at the one game in... is it 5 years that they put out on pc.
the overwhelming glut of pc games barely make proper use of quad cores, much less the threading. i'm not saying they don't exist, because they do, but they aren't the norm, that's why its shocking when one that does comes along.
on that note, amd will get damn near every game with few exceptions to at least 60fps, but its alot harder for amd cpus to hit 120-144hrz.
>>52190927
it would murder and face fuck intel (if the specs are to be beleived) till the E series comes out, and even then, price performance would probably skill go to amd.
>>52183164
>lead designer a female
This is going to be hilarious to watch AMD crash and burn at the end of this year. Their console sales will be all that keeps them afloat as Intel completely dominates the CPU market.
