http://wccftech.com/amd-confirms-zen-debut-q4-2016-apus-2017/
>Q4 2016
>Q4
AMD IS FINISHED & BANKRUPT
it better be good. if it's another 2010 tier effort amd is finished.
>>52311149
i'd rather have it late than unfinished.
Fine by me I'll wait. My 4690k is serving me perfect other than long transcode times. I'll gladly wait if Zen delivers as promised.
>8c/16t $400 cpu with 90% haswell single core performance
Thats if they price it right. If amd makes something good for once on the CPU front, gets a big head, and jacks the orices sky high, they can suck my dick. I'll buy x99.
>wccftech
>confirmed
i shyggy family
>>52311562
The source is Lisa Su, everyone is covering it.
Just publicly stating what we've known from their leaked roadmap since mid last year.
Dont know nothin about zen - someone redpill me. Why is it good? I'm sitting on i7 950 3ghz now...will zen be an upgrade or sidegrade?
>>52311149
Q4 was confirmed a while ago. You are late with the news. Go suck intels minority loving cocks.
Q4 really means 2017 because they won't be able to ship until then
AMD is truly finished and headed to bankruptcy
>>52311604
Considering the i7 900s are similar to FX 8350 performance, i would say zen would be a decent upgrade
>>52311614
Their internal roadmap stated in channel October 2016.
Don't pull things out of your ass.
>>52311646
I just looked at benchmark comparison results comparing a fx 8350 ($170) and a g3258 ($50) and the g3258 won. How do you explain this then.
>>52311604
AMD are promising a significant increase in performance relative to their old architectures, which people are speculating will put them on par with Intels last generation of processors in single threaded applications.
>>52311901
Not him but hard to explain it when you don't mention which benchmarks you're looking at. You're being so vague that you're bordering on being an obvious troll.
>>52311947
>Not him but hard to explain it when you don't mention which benchmarks you're looking at.
http://cpuboss.com/cpus/Intel-Pentium-G3258-vs-AMD-FX-8350
the price is not $70 but $50 in store at microcenter. That puts it at 1/3 the price of amd.
>>52312005
>cpuboss
Pffffftthahahaaaaaaaaaa
>>52311660
What is fiscal year?
i get amd cpus simply because they are the small guys even though they still make millions of dollars like the competitor
Q4
How can AMD go bankrupt, aren't all the current Gen consoles currently using AMD OEM cpu and gpus? If anything they'll shrink and stick to OEM contracted shit.
>>52312005
CPUboss isn't a benchmark, it's a comparison site and a bad one at that.
Out of the six metrics they have, only two are benchmarks, one of which is Passmark. Passmark is so terrible that it's widely known as pissmark. It is not to be used, at all, ever.
>>52312029
I know he's wrong but you're not helping.
>>52312149
>If anything they'll shrink and stick to OEM contracted shit.
That's not a good thing either. We don't need another via, releasing one new architecture every three years and producing chips only good for set top boxes and dirt cheap Chinese netbooks.
>>52312198
Didn't say it was good but it's probably the only thing that will keep them from going completely under.
So now the only option is Intel and their fucking racist bullshit...great.
>>52312304
What?
>>52312304
if you actually cared about anything you'd do something about it
>>52311320
>If amd makes something good for once on the CPU front
Not important, but amd had better processors than intel back then.
Sadly, thanks to the heavy marketing of intel and a couple of dirty tricks AMD never got substantial money.
>tfw waiting for Zen to build a comfy AMD PC
Zen CPU
8GB of DDR4 RAM
Radeon 400 series GPU
A comfy SSD and 1TB hard drive
A Seasonic PSU
Fractal Design Define R5 case
>>52312464
Oh I know the history trust me. I was ysing FX since gen 1 bulldozer. Regardless, as a consumer, I'll not buy an inferior product just because it has a name slapped on it.
Even if the company with good products is a Jewish Monopoly who pulls dirty underhanded tricks to get by.
Why does everyone believe they will price zen so low if it's at haswell performance? All the times before AMD has been competitive they priced their products thereafter or even higher. Stop being so gullible AMD fanboys.
>>52312519
what
Bulldozer was always shit
This is not news
>>52312465
The Define R6 will be out by then, if not the R7.
>>52311320
>>810c/20t $400 cpu with 125% haswell single core performance and 150% multi thread
see, anyone can pick random numbers from their asses.
>>52313135
>improving on perfection
Nah m8.
>>52313264
They've done it before with the R5, since everyone thought the R4 was perfect.
>>52313290
Nah, the R5 hasn't been out nearly long enough for them to replace it. The R4 was around for forever.
>>52312465
My build this November/December:
8 core zen
16-32GB DDR4
R9 490x
240GB nvme m.2 PCIe ssd or 3d xpoint
500GB - 1TB TLC ssd for mass storage
A nice silent mid tower case
>>52311149
>AMD IS FINISHED & BANKRUPT
I thought they already were . . .
Will my 8350 @ 4.7ghz keep me alive until Zen? I planned on getting an oculus, too, but I have a Fury X, so maybe it'll be enough...
>>52313473
no, if you keep running that, you will die
>>52313490
Goddamn it, anon...
>>52313473
Your house will burn down before Zen drops.
>>52313344
You know they're f'd when they only thing they can brag about with their new GPU is it's perf/watt efficiency.
>>52313617
Hasn't in the past year, anon. In fact, temp are nice n' low
>>52313679
So they're as fucked as Nvidia was two years ago?
single socket for Zen cpu's and apu's.
I think that'll be good for MoBo choice and variety.
I work for a FAB that manufactures these chips and our order is for the end of q1 start of q2. What is AMD doing with the extra time?
>>52313793
masturbating.
>>52312304
Anti-white is not racist. It's racial correction.
>>52313209
An i7-5930X will run rings around the top-tier 8-core Zen. Intel versus AMD is not even a contest at this point. It's a slaughter.
>>52313793
>>52313793
Sure those aren't GPUs?
>Zen on track
>Polaris on track
>Good console sales
>Nothing but good news
>Stock tanked for no reason
wtf is going on?
>>52313679
they can't even do that...................
>>52314071
China go boom
>>52314487
>Market down 10%
>NVDA down 10%
>INTC down 10%
>AMD down 30%
There is more than China going on here
Don't get your hopes up.
Su already said in a german interview that zen will be released Q4 - to OEMs and server market. It is very much possible that it'll be Jan 2017 till zen can be bought in shops.
>>52314567
short sellers
>>52314567
weakening foreign currencies, strength of dollar, oil prices (mostly denotes oversupply but also indicative of slowing global growth)
AMD is toast. I considered shorting at $3 but was a pussy. They need a gamechanger product and they best they will have is cheaper last gen intel-equivalent products and muh performance per watt GPU's
>>52314918
>muh performance per watt
worked for Nvidia
>>52314969
And it'll work for them again. AMD is just playing catch-up.
>>52312465
>Seasonic PSU
My nigger, Platinum series is the fucking best PSU series on the market.
Don´t understand how all the retards buy corsair or something like that and think they are the best.
Gonna buy somewhat of the same PC end of the year, just with a Phantek Evolv ATX.
>>52315069
>tfw no Seasonic master race PSU
Oh well, got a decent price on this CM Silent Pro M700. Thing's dead quiet, even when I'm pulling 500W.
>>52315069
Seasonic actually just announced their first Titanium line, called PRIME. Will have titanium-rated units at 650W to 1000W and platinum-rated units at 550W to 1200W
AMD<33333333
>>52315069
Because Seasonic PSUs start in the low €200s.
Who can afford that shit?
>>52315069
>tfw poorfag
>tfw budget corsair
>tfw 5 years without a hitch
I'll deal with it.
>>52315767
A good power supply is really the only thing you can "future-proof" on a PC. It's also arguably the most important, as it is in part responsible for the lifespan of the other components in your system. If you aren't willing to put in money for a high quality power supply you might as well not build your own PC in the first place.
Intel gon suck dick if Zen turns out to be competitive and cheaper
>>52311149
Hello /v/
>>52312005
>cpuboss
Hello /v/ again
>>52315069
Crosair has a couple of PSUs made by Seasonic. Buying their 250$ Platinum units is also more like placebo unless you go for some hardcore OC + triple SLI builds. The G or even their SIIs are way enough for most (single gpu) systems. Overspending on what you need is almost as autistic.
>>52313777
I agree, we may finally get some ITX choices for AMD builds.
>>52315799
>test results show that seasonic psu's are not better quality than other well-regarded brands
>often louder than the competition in the same efficiency range, for double the price
You're effectively paying €100 more for the brand name 'seasonic'.
Sure, PSUs are important, but how many pcs run on shitty chinese psus for years without trouble? Both my brother and my dad's computer have been running on a chinese knockoff brand psu for over 7 years and no problems.
I spent a bit more money on a quality psu, but there is no way I'm paying double for a seasonic psu.
Anyone who does just got tricked by marketing.
>>52315820
>posting benchmarks means you're from /v/ and your opinion isn't to be taken seriously
you got the logic of a woman bud
>>52314071
Market took a nosedive. In China, it's gotten so bad that the government put in new laws preventing stock owners from selling their shares until they submit the sale for approval to the government, which has a mandatory 15 day limit between the date the application was filed to the date which the sale can take place upon approval. Basically, the Chinese government is blocking people from selling stocks to keep the market afloat for just a bit longer, as futile as it is.
>>52315886
I hope your computer dies
>>52315931
Why?
Are you that butt-blasted that you got tricked by marketing? It's not my fault.
>>52315966
his self-esteem was crushed upon finding that he paid 100 euros more to feel superior
>>52316000
Poor guy.
No pun intended :^)
>>52316000
I live in clapistan, a decent Seasonic PSU is under $100.
Maybe if they release a 760k/860k like they did for Richland/Kaveri.
My 860k kicks some ass for a $70 chip.
>>52315798
That's because most Corsair supplies (Starting from the CX) are actually okay (Fine, not good) and OP is seems like some autistic psu etilist (Think of an audiophile, just for PSUs)
>>52315799
There is absolutely no reason to buy a 200$ platinum Seasonic PSU unless you're going to stress your components to a higher degree 24/7, want to be really secure with a server, or keep your PSU forever and miss future technological advancements.
For your average budget system, some mid-tier units like the Seasonic G or FSP Raider are enough - even some entry-level units such as the EVGA B(HEC) Corsair CX (CWT) just work fine, even to spans over 5 years. Their voltage regulation is within spec, the ripple is usually <75mV and they can supply the wattage they're rated for with their capacitors not being the best, but doing the job just fine for at least a hand full of years under general room conditions.
>>52315901
>I post a site which mainly relies on synthetic benchmarks and random statistics with artificial scores
/v/ is an acronym for a tech-illiterate faggot coming over here and pretending to know something, which is like most pc gaymen act.
Not only stating cpuboss, but also as your only source shows that you have no actual clue what you're talking about and thus should fuck off to >>>/v/ where you actually belong.
>>52316037
That would be a good move for them to launch AM4 with a budget Excavator dual module CPU.
>>52313962
>Racial correction
So, racist.
>>52316121
>There is absolutely no reason to buy a 200$ platinum Seasonic PSU unless you're going to stress your components to a higher degree 24/7, want to be really secure with a server, or keep your PSU forever and miss future technological advancements.
I never fucking said to buy a platinum, jesus christ. I just said buy Seasonic. They have plenty of quality PSUs at or under $100. If they really start at €200s then I feel bad for yuropoors.
>>52314918
At least AMD can keep up in the performance division, but is ultimately behind. Nonetheless, Polaris could take things up a notch, they're not as behind to Nvidia as they are to Intel. What they need is some bigger investor, and a better marketing team.
>>52315820
>Hello /v/
/v/ is for VIDEO GAMES
>cpuboss
>Hello /v/ again
/v/ is for VIDEO GAMES
>>52316256
If AMD can deliver with their Zen/Polaris SOC next year and succeeds in getting their driver stack on Linux up to par they'll likely get some big investments from companies wanting to make Steam Machines.
Its just a shame that Nvidia is still denied access to the x86 market. Intel badly needs to be broken up with the x86 licensing group being a separate independent organization.
>>52311149
>wccftech
kek
>>52313264
The Fractal R5 is pretty fucking far from perfect. You can't swap HDDs without needing to open both sides of the case, for one.
Perfect would be, say, Antec P183 with 8 hdd bays, every fan slot being 14cm, and three fan slots on the top.
>>52316322
>You can't swap HDDs without needing to open both sides of the case, for one.
Who cares, it's still an amazing case
>>52315069
>My nigger, Platinum series is the fucking best PSU series on the market.
Enermax Platimax was better, I think they are still in the market.
Their 600W PSU can deliver something like 830W at peak.
>>52316322
Corsair 760T is pretty good too. Only downside are the SSD slots, but they're still far from terrible.
>>52312465
Same boat nigga.
Zen 8 core
32GB DDR4 RAM
Arctic Islands GPU
xPoint SSD/NVMe SSD with a 2TB SSD for storage
Platinum PSU
Fractal R5
I might even go with a tiny case. My wallet has been ready since summer 2015.
>>52316178
As I said, certain budget units from other manufacturers (mostly FSP, HECand CWT) can be fine. It doesn't have to be Seasonic - yes, even their low-end PSUs can match the mid-tier on others, but it's not that you need that tiny little bit of better regulation and ripple when you're just running your i3+r380 system at stock (or mild overclocks/without voltage increases) anyway. As long as they're within spec and not running shit-tier capacitors blasting after two years of usage, you'll be fine.
Super Flower also does some pretty good Seasonic competition recently :^)
>>52316336
>Shilling a generic midi tower that's exactly the same as all other generic midi towers
>>52316376
Too bad AMD motherboards won't/cant' support M.2
>>52314071
>tfw not buying at $1.7 because didn't know shit
>>52316382
I'm not willing to take that risk, as small as it may be. I'm going with the reliable brand, even if it costs me a bit more. They've got a good reputation for a reason.
>>52316400
I was worried about this. Source? I'll post if I can find it.
>>52316322
Who the fuck swaps out HDDs on a daily basis?
Are you autistic?
>>52316305
See >>52316121 at the bottom
>>52316316
They can be nice as a backup source when you have one reliable :^)
>>52316307
Indeed, but Intel would never give their baby to someone even more dangerous than AMD, and I don't think we can hope for authorities to change that anytime soon either.
>>52316036
kek, yuropoors getting cucked once more
>>52316376
Pretty much the same here.
8 core Zen
16GB DDR4 or better
AMDGPU supported graphics card that can outperform my current card
Other than that just pull over all my current hardware
Maybe get a new case, my old Lian Li is kind of outdated and behind the curve in cable management.
>>52316412
There's some good reviews on known budget units with stress testing and interior check, that's secure enough for me.
Of course, I wouldn't put them in anything over 500-600$.
>>52316400
Wrong. In fact, the first motherboard to market with an M.2 slot was an AMD motherboard.
>>52315799
>as it is in part responsible for the lifespan of the other components in your system. If you aren't willing to put in money for a high quality power supply...
The thing is designing a power supply that won't fry components is fucking easy. Voltage regulation is a solved problem. Dropping another $100 to get the ripple down from 10mV to 5mV is not really worth it. It's digital circuitry not an instrument amplifier with poor power supply rejection.
I want a low end apu, guess I'll have to wait till 2018.
>>52316524
How many times do I have to say that Seasonic PSUs aren't $100 more expensive when compared to comparable units? There are several quality units that fall in or at $100 at competitive prices.
>>52311149
>planning to switch from fx-8350 to intel
wew lad
>>52316532
AM4 motherboards are releasing with carrizo chips in spring I think, well before zen.
>>52313962
>anti white isn't racist
good bait, stephanie tumblr
>>52316596
cuck
>>52316596
>Switching to Faillake
for what purpose?
>>52316561
They are still more expensive, at the price of a 350W SII without modularity, there's a 430W Corsair CX with semi-modular cables, or a 500W one without modularity. (I just checked it up.)
The Seasonic will be of higher quality, but the CX will be good enough to do the job as well. It won't fry your system and probably lasts long enough until your next build too - and the SII is not a high-end supply either. You might pay for a minor quality increase you do not need. If you want absolute reliability, you will have to pay more either way.
Of course, you can get the SS/SII and pay the premium, but when we talk about budget systems, you usually just buy the quality you absolutely need.
>>52312005
CPUBoss isn't a benchmark site, it's a billboard for Intel. Try again with real benchmarks.
>>52311149
>>52312465
I don't know what's going to be out when Zen is released, but I'm spending whatever it takes to have a high end GPU, Zen, and 32 gigs of ram in a mini-itx case.
>>52316894
>430W Corsair CX
Those are only good for 430 watts at 25 degrees celsius. This is a pretty normal way of overrating a PSU. It might as well be a 350 watt power supply.
>>52316421
>Who the fuck swaps out HDDs on a daily basis?
I'm recovering data from a faulty HDD right now which hangs up every half hour and needs to be reconnected. This is much more difficult on the R5 than it is on my previous P183, where I could just reach down and reconnect the drive cables with one arm. On the R5, the HDD bays are so cramped that I can't do that.
I could install the drives with all the connectors on the front, but then they cannot be routed efficiently (underneath or behind the bays) due to not being long enough. I thought of routing the sata cables underneath each drive, but that wouldn't help either due to the power cables being locked together.
On the P183 I could take out the entire hdd bay, disconnect a drive, rotate the bay, then unscrew and pull it out from behind.
What the fuck is the point of having 4+ hdd slots if you can't swap them, it's retarded.
And the hdd grommets are so fucking useless on the R5, they are too thin, they don't block fuck all drive vibrations. I had to order spares from Antec and put them in the R5.
>>52313069
people actually believe this
>>52316948
Here is a detailed review of the CX430 that pushes 422W testing under 37 resp. 48°C just fine, with nearly excellent ripple, and okay capacitors. The only issue is a Crossload scenario with stressing the 3.3 and 5V rails only, and for data hoarding/servers where this might apply, you shouldn't buy a CX or anything in its class anyway.
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=239
Is there a site like cpuboss that isn't cpuboss and relevant?
>>52312089
>>52312089
Fiscal means financial
Upgrades at end of year:
AMD Zen CPU,
16GB DDR4 RAM,
AMD R9 Rage,
Noctua NH-D16
>>52317160
Anandtech's CPU comparison with several different benchmarks and performance charts
>>52317160
Google -> Sites with real-world benchmarks with actual applications and games under realistic conditions.
This >>52317358 is also based on real-world scenarios, albeit a little outdated sometimes.
Downgrades for 2017
AMD Zen FX-8550e budget CPU
4GB of DDR4 2133 in single channel
AMD R9 Gemini
AMD stock cooler
Foxconn AM4 motherboard
80GB 5400rpm SATA II HDD
Gotta support AMD at all costs, amirite ;^)
>>52317358
>>52317388
It's a shame because cpuboss has a pretty decent list, only other website that haß a similar list is tons hardware but that's a bit dated too I think.
>>52317409
>4GB DDR4
>80GB HDD
>Sata II
>>52312304
this. Intel supports Feminist Frequency.
>>52317424
The lists on cpuboss are mostly based on synthetical benchmarks, which are almost worthless to determine the actual performance of a CPU. (And GPU, respectively.)
What you want are benchmarks with actually used applications. In which case, you have to separately google reviews of certain chips getting tested under these. They're pretty much always compared to other chips anyway, if you google the 4690K for tests, you'll see it matched to the 8350 and others, for example.
>>52312304
What the fuck have you been smoking?
Someone enlighten me on how the dick-hole a microprocessor company can be racist, please...
>>52317502
Those are just angry white men who are mad because Intel is funding pro-feminist and multicultural groups that they don't like.
Mostly Goobergators and Redditchanners, I bet.
>>52314071
Hi anon. Modern Markets are driven mostly via computers (pick/board related). So big index funds literally just by every-single-stock-on-the-market. This is the safest investment to make. So its all the rage. Furthermore the index funds will not only buy shares in the US, but China, GER, UK, FR, AUS, KPOP, Japan, Hong Kong.
Next you got inter-connectivity. China tanks, so AUS tanks because their entire economy is pinned to mining Iron/Coal for China. Hong Kong tanks because IT IS CHINA. Index funds now *rebalance*, which means sell off US shares. But when you sell of a COUPLE MILLION dollars it has ripples.
EVERYTHING drops. Now if you are smart you buy those US Tech Stocks which are basically nothing but good news for 5-10% lower then they were LAST WEEK. Just in time for Christmas earnings reports.
How is /wage/slave doing?
>>52317495
I can almost never find my CPUs in those lists as they're usually too old. Celeron 100u, i5 2400 and a xeon 1260v2. It's more curiosity, to see how much improvement is to current generations compared to mine. Most benches only include a few samples from last generation and compare it to the current.
I mean I know not to take cpuboss serious but its a good concept, shame they used the wrong benchmark methodology.
>>52317565
>being this tublr
post your blog link so we can see your dragon dildo collection.
>>52315069
>Don´t understand how all the retards buy corsair or something like that and think they are the best.
went from a s12g 750w non modular PSU to a RM1000x modular PSU that's rated 10/10 by johnnyguru and has a 7 year warranty. your argument should be against people who buy CX PSU's for high end builds autismo.
So its pretty much confirmed to be substantially weaker than intels 2017 offerings since that's what it'll be competing with (q4 2016 zen release). Who in their right mind would buy this shit when it's probably not even going to be more powerful than my current 4690k even though my cpu is old now.
What's going to happen is everyone will just buy old i7 4790ks since they'll be cheaper and more powerful than any zen offering.
>>52313962
>Anti-white is not racist. It's racial correction.
Not correct. Whites are fast becoming minorities in many countries. In the USA for example, whites already lost majority among children. Once that generation grows up, whites won't be a majority country wide.
Some other countries have it worse. France has 50%+ black birth rates in the Paris area (10-20% in other areas, Bretagne was the lowest with 5% or so). And that was in a very old study; the current African refugee flood is making things worse.
>>52317787
>asian
>not a minority
I don't get that. Is it because they don't really kick up a fuss compared to dindus and Arabs?
>>52317836
Asians could be best summed up as diligent worker ants. Plus they also have higher intelligence, so they know better than to kick up a fuss.
And yeah, they are a minority. Hispanics are expected the become the biggest race in the USA in a generation or so. Whites are on a decline, while blacks are, interestingly, levelling out.
>>52317715
>Who in their right mind would buy this shit
People who don't have a 4690k in their system.
If it's anywhere close to Intel whilst at a lower price, we already have a deal.
Stop being an autist and measure everything to the top-level performance, that's only important in the server/enterprise area and even there it could be an option if the performance/dollar is better. (Without the overall performance being severely behind.)
>>52317715
>Who in their right mind would buy this shit
Not everyone has Haslel. This is more for the *bridge and other AMD guys. I would upgrade to a 5820k from a 3570k, but I'd rather wait for an 8 core since it'll be better than my 3570k and have the extra cores for the rendering and VMs.
>>52317648
>its a good concept, shame they used the wrong benchmark methodology.
They use no usable benchmarks, which is why they have such a massive list of CPUs and GPUs. They just scrape the technical details from other sites and use those to give bullshit numbers.
>>52311149
I fucking told you. We won't see Zen APUs on massive markets until Q2-Q3 2017. Intel will bring something better to counter attack fuck this system
>>52317031
But it was shit
>>52311613
Intel gives millions to appease women's feelings
AMD has a CEO that's a female
Well I guess one talks the talk and then amd walks the walk
>>52312005
lol, neat
>>52312149
/g/ is retarded curry kids and Russian warez scene kids on krokodil nao
>>52313962
Oh wow nice
A 10% performance boost for an extra $300
You're a genius
>>52317787
Who fucking cares? Wahhh people are turning slightly more brown!!!!
>>52316910
http://www.computerbase.de/2016-01/amd-ceo-lisa-su-interview/
>>52316307
They aren't 'denied access' to anything
They don't have a fucking license
>>52318843
Which is fucking stupid.
>>52318798
>Who fucking cares? Wahhh people are turning slightly more brown!!!!
Spend two weeks in Paris or maybe Detroit.
>>52318992
>>>/pol/
Post your delusions elsewhere, stormweenie
>>52316937
This this this.
>>52318843
That's exactly what licences are. Any situation in which you need a "licence" to do something is you being denied something.
>>52318798
>>>/tumblr/
now, you self hating submissive cuck
>>52319020
you forgot to take a sip of your starbucks and toss your "chic" scarf
>>52311149
AM4 is still launching in Q2 though, they have something to hold people over
>>52311604
heterogeneous computing and stacked NAND
>>52319587
Trump will not survive the nominations
Hillary has her sights on him, and her finger's on the trigger
Reminder this is a leaked intel road map for releases..
>10 core cpus
>>52319638
Fuck off with the politics bullshit, pinko.
>>52312005
The Per core performance is going to be faster on the G3258, but the 8350 will wreck it's shit in any workload that requires many threads
>>52319645
AMD IS FINISHED AND BANKRUPT etc etc
>>52319617
Is AM4's socket able to accommodate AM3 and AM3+, or is it going to break compatibility like AM3 did to AM2?
honestly, zen only has to be as fast as sandybridge for me to switch to the green team.
>>52319705
AM3+ requires like a northbridge and a south bridge, most CPUs now have it all intergrated and so will AM4 CPUs
the first chips for AM4 will be APUs that make use of DDR4 and will the desktop variants of Carrizo APUs
So no, of course it's not backwards compatible
>>52319645
>broadwell-e
>$1000 starting for the 8 core
>>52319645
>still only quadcores for the mainstream
>10 core will be gorillion dollars
>>52311265
This
>>52313135
and hopefully by then the R5 will be overstocked and ~80 bucks just like the R4 was
>>52315069
anything Seasonic or SuperFlower made is god tier
this includes some Corsair or EVGA stuff
if you're buying on brand alone you're missing out on possible deals. always check who makes what.
>>52316322
>HDDs
>in a desktop machine
>in 2016
hahahahaha
>>52311149
I want dual socket APU motherboards.
Just imagine having twice of cores with HT and crossfire from integrated videocards.
>>52319645
>Left edge of box represents early edge of production windows.
>Kaby Lake-S only in Q4
How bad is Cannonlake delayed?
>>52320274
>16 cores
>32 threads
>GPUs equivalent to dual 270Xs
>in a low-profile case
I want this.
>>52320298
This will kill all competition of will be not that pricey.
>>52320285
Pretty badly, since Kaby Lake might last until 2017. Cannonlake is expected to launch around Q3 2017 at the earliest.
>>52320298
>270Xs
Allegedly, Zen APUs might fit over 20 GCN units, which is more than the 270X, but less than the 380.
>>52320349
Nice. If they made a budget quad-core Zen APU with that iGPU they'd make the entire low-end GPU market obsolete.
>>52320349
And for power efficiency reasons they will run at lower clocks than a stand alone GPU so the performance will likely be closer to 270X levels.
>>52320364
>they'd make the entire low-end GPU market obsolete.
That's why you pack more compute units into your GPUs. The physical size of the GPU shrinks with each die shrink, so it's only common sense to pack in more transistors in a similar amount of silicon for almost the same price.
>>52320383
Slap a massive heatsink on and overclock the shit out of them. I'd like to see the results of that, to be honest.
>>52320383
>they will run at lower clocks
Don't forget that it will use DDR4 memory rather than a GDDR5 or higher memory. HBM will not make it over to Summit/Raven Ridge.
>>52320349
Zen APUs will use GCN 1.3, so you can't directly compare it to old APUs.
>>52320414
I thought they called it GCN 4.0
It might pack more shaders and TMUs into each GCN unit, or more directly tie the ROPs to each GCN unit rather than having them serve the entire GCN Core. Perhaps more L1 cache as well?
>>52318539
>was
But not anymore!
>>52315069
>Phanteks
BASED
>>52311149
Nvidia will be fucked when DX 12 comes out
>>52320737
>this will never happen outside of this one, modified benchmark
Shame, I really want to know how two 980 Tis would play with a Fury X
>>52320414
Raven Ridge is not third generation GCN/Volcanic Islands. Carrizo/Bristol Ridge is the only third gen GCN chip.
There is no such thing as GCN 1.X, stop repeating this garbage.
>poointhelootech
>>52320737
>talk about cpus
>the 980ti is still faster
>nvidia will be fucked when DX 12 comes out
>>52320811
>1920 X 1080
Discarded
>>52320482
You wouldn't add more ALUs to the CU, you'd be decreasing throughput, and decreasing the number would send heat through the roof.
The primary change to the CU itself is the scalar processor.
>>52321145
http://www.anandtech.com/show/9659/fable-legends-directx-12-benchmark-analysis/2
>Maybe tech forums can now stop with the "AMD will be vastly superior to Nvidia in DX12" nonsense.
>>52321567
AMD did mention in an interview to Anandtech that the GCN architecture is relatively unchanged for Polaris, so it might actually be the same number of shaders and TMUs per GCN unit, but more units per CU core.
>>52321613
>, so it might actually be the same number of shaders and TMUs per GCN unit, but more units per CU core.
Holy fucking shit this is painful to read.
Stop trying to use terms you don't understand.
It is 64 ALUs per CU. This is not ever going to change for the GCN architecture.
>>52321665
>CU core
Fuck, I meant GCN. Didn't realize it until I reread my post. I even had this open on my other tab, so I have no excuses.
https://www.amd.com/Documents/GCN_Architecture_whitepaper.pdf
Where are the flux capacitance impeders in the computational vertex rendering units?
what is latest i5 XXXXk compared to 2500k?
Did Intel actually advance in these years without competition?
They seem to compare CPU's to anything but those golden Sandy Bridge ones, can't see them on any benchmarks.
Is it actually because no inprovement?
>>52321809
Skylake ranges from 5% to 30% faster than Sandy Bridge on average, and in a few more niche workloads its over 70% faster, though no consumer would ever benefit from obscure AVX utilization.
>>52321850
>5 years
>up to 30%
>higher prices
You suckers better get Zen even if it sucks.
without AMD entire PC thing is finished.
>>52321886
Lack of competition has nothing to do with intel's marginal performance uplifts generation to generation. This is simply diminishing returns in full effect, and they're spending money where it matters for the consumer market.
>>52321886
x86 is a sinking ship anyway.
ARM will be our lord and savior from shit efficiency and outdated extensions.
Why doesn't amd just make a cpu with 12 or 16 cores inside of it??????
Why is there no dual cpu motherboard, arn't 2 cpu's suppose to be better than one???????
>>52320364
>If they made a budget quad-core Zen APU with that iGPU they'd make the entire low-end GPU market obsolete.
They are supposed to, and with one stack of HBM to boot (512mb). But you have to wait till 2017 for those.
If they could put 4 stacks of HBM2 in an APU, then they could make a smaller form factor that does away with RAM slots too.
>>52322011
They do make 16 core chips, they're called Opterons.
They do make dual socket mobos, they're for Opterons.
>>52322014
>, and with one stack of HBM to boot (512mb)
Buddy, no one is producing 512MB die stacks. It would be 1GB or more.
HBM isn't going to replace conventional DRAM. Ever.
>>52322069
ah, you are right, I confused the size with the bandwidth (128mb/s).
>>52322101
One die stack can provide 128GB/s bandwidth.
GB/s, not MB/s.
Pretty fucking big difference there.
>>52322124
okay, it's pretty clear that I need to go to sleep then, since I'm making very ridiculous mistakes when quoting numbers (it's 7:42AM mind you and I spent the week reading chunks of data from a dying 3tb had drive).
>>52322069
Why not? The custom jaguar chips used in the PS4 use a similar configuration only with GDDR5.
If AMD were to release an APU with 4/8GB of HBM on die you would have no need for a GPU under $150 since it will not have the same memory bandwith restrictions as their current APUs. Especially when they can improve the amount of CU's.
>>52322157
The PS4 APU uses a couple ring buffers to deal with the fucked up timings of the system memory. This is a work around and a performance compromise. The PS4 has this configuration only to reduce manufacturing cost and complexity.
The final HSA spec does not require a single fully HUMA address space. A chip can have two different address spaces and use them as fit for the given type of workload, and that is the way future chips from AMD will function. DDR4 serves as system memory while HBM is video buffer and flat address space for compute workloads.
>>52322157
AMD is saving HBM APUs for something massive.
I'm thinking 16-core Zen+R9 GPU with Hawaii levels of power. The die and thermal requirement must be massive.
Why aren't there any zen benchmarks yet -- or any real solid information? It's been in development for years, so surely they must have stuff internally; not showing it is not a good sign.
>>52322300
Because they aren't ready to release it yet, probably doing some finishing touches before full production, they also have to wait for fab time as well
>>52322300
Look at all that desperate FUD
>>52322300
Because if they show what they got now Intel will come out with something that appropiatly blows it away by 5percent causing intel fanboys to go like " OHH I TOLD YALL FAGGOTS THAT ZEN WAS SHIT INTEL LATEST CHIP 5% FASTER OHHHH"
Same shit happened with the ps4 and xboxone, remember?
xbox shoed their cards first and everyone was like fuck that shit, then few months later sony shows their ps4 and it becomes the better system prcie wise
>>52322069
PS4 runs on integrated memory using GDDR5 for RAM and it has almost PC architecture and haxors already ran linux on it.
>>52322314
>>52322338
>>52322406
I hope so. Some substantiated hype might help AMD survive while it waits on Zen.
>>52322690
What a stupid thing to say.
The company isn't near bankruptcy, and hype doesn't bring in revenue.
>>52318325
>If it's anywhere close to Intel whilst at a lower price, we already have a deal.
Sure, it'll be cheaper but it'll also be significantly weaker than the latest 2017 intel offerings which is quite a big gap. You can probably compare it to an fx 8350 vs i7 4790k right now. Cheaper yes, but significantly worse performance. And just like now, people would rather spend a little more for that much superior performance. I can't see good things unfortunately. Maybe I'm just a negative guy.
>>52322719
If AMD fails with zen then their CPU division will go under
>>52322069
>tfw quad-socket opteron zen board
>48c/96t
my dick is ready.
>>52322719
more positive brand recognition => more consumer confidence => consumers more willing to buy AMD products other than zen
>>52322300
you are retarded
>>52322300
there was a sisoft benchmark that surfaced i think?
might've been for the carrizo chips though, don't remember.
>>52320811
>>52320773
>cherry picking
With new drivers the fury x is actually faster than the 980 ti.
>>52322726
Yet another horribly stupid thing to say. This sentiment is mindlessly regurgitated all over /g/, and not a single one of you has any idea what the words even mean.
The primary growth segment they're targeting is enterprise where they can leverage their freedom fabric IP. A server chip doesn't need to offer 1:1 performance parity with the competition per socket when its multi socket scaling is higher. The Zen based products will provide a positive ROI, they will bring in steady revenue. Even if they managed just a 20% uplift over Excavator they'd still have a better positioned product lineup than they do now.
Little plebs like you can look at an Anandtech bench and shout "ZOMG GUYS XD LE ZEN IS FAIL XDDDD" but that has no bearing on how a company actually runs.
They aren't near bankruptcy.
The CPU division isn't in any trouble.
As a matter of fact the successor to the Zen core has been well into development since the beginning of 2015, a second team is leap frogging, and this trend will continue onwards for the foreseeable future.
>>52322797
I don't think you realize just how far behind they are even on servers, there's literally no reason to buy an Opteron CPU unless MAYBE you need to run a ton of virtual machines
If zen is a flop they'll have no where to go really unless someone buys them out and gives them the proper funding to compete with intel better
>>52322719
>The company isn't near bankruptcy,
They almost run out of funds. Last time I heard they only had a couple hundred million in free cash, which for a company of such size is chump change.
>and hype doesn't bring in revenue.
It has a positive effect on share prices however, brokers will see the hype and will want to get stocks while the prices are still low, increasing demand, and thus pumping up shares. Strong numbers will also bring in investors; all of this means more cash for AMD to put into either R&D or into increasing inventory for any products they have that do sell fine.
>>52317649
>>52312304
>>52313962
>>52316153
>>52316641
>>52317787
>>52317487
le #whitegenocide XD
>>52322818
>saying even more stupid shit
Oh boy.
I'm well aware of AMD's position in the enterprise market, they're sitting at about 1% market share.
AMD is a company generating revenue in the range of $5,000,000,000 per year. They have around $2,000,000,000 in long term debt, and they're teetering on breaking even since restructuring. This is not what a company looks like when they're on a death spiral, this is what a company looks when they're about to pay off all their debt and operate cash in hand.
>>52322797
>They aren't near bankruptcy.
They also aren't profitable. If zen doesn't make them a profitable company, it's worrisome.
>>52311149
i want to support amd coz nvidia are jews and intel are sjews, but amd products are always so damn bad.
>>52322797
You are right in that for enterprise market, the most important thing is performance per space, not performance per core.
But, Intel has all the markets bribed up to their eyeballs, and AMD has minimal market presence especially in enterprise. And they don't even have a competitive product to sell, didn't have since years.
I don't think Zen will make or break the company, but it definitely sounds like a big last hurrah that, if it doesn't work out, will leave them in an unimaginably big shit. If not making them bankrupt, then simply marginalizing them further, to where VIA was on the x86 market ten years ago.
>>52322790
>buy a reference 980ti
>>52322924
For the money you could buy 2 mid range GPUs and still outperform it though
>>52322875
If Zen is a flop, AMD's CPU division is dead. AMD will split into half again (Gloflo was the first split), and the CPU division will cut down hard and end up turning into VIA 2.0 or something.
>>52322936
AMD spun off their fab because it was an endless money sink. They weren't competitive, they were constantly late, they ran in to technical hurdles which caused major setbacks routinely. To this day the only reason why Global Foundries is operating is because they've been constantly injected with foreign capital, they require constant stimulus to keep running. Shedding that cancerous weight was one of the best things AMD ever did.
The CPU division of AMD, which is the center of the entire company, is profitable. There are no parallels to draw between this and their fab.
Stop throwing words together just because you think they sound right. The sentences you're posting are painfully stupid.
>>52321953
>ARM
not the great and glorious mips, capable of executing x86 and arm code natively.
>>52322690
If they hype zen they lose a lot of sales on current CPUs.
They'll hype it before release, but not this soon.
>>52322924
Yeah, overclock your 980 ti, good goi.
>>52312149
Barely any profits from the chips they're selling towards the console manufacturers.
>>52323062
>oc fury x
>+1fps
>oc 980ti
>+8fps
Jesus Christ, Fury x is so shitty...
>>52323062
What is the point of buying a $500 graphics card when you can't get 60 FPS on Ultra anyway?
Why not just get a cheap as fuck card and reduce the graphics settings?
>>52324062
I don't think you realise how poor sli scaling is next to crossfire.
>>52322011
>DURRR WHY DNT DEY JUS MAKE BETTAH CPUS HUH???? MOAR COARS!!!
kill yourself
>>52324062
not to mention only 4gb vram and they announced hbm2 before the actual release. i honestly don't know what the fuck amd is doing.
>>52324128
Are you implying at apu can handle 4k gaming?
Those benches are at ultra, if you turn it down to high your fps would probably jump up from 30 to 45.
>>52322724
>You can probably compare it to an fx 8350 vs i7 4790k right now.
An 8350 is not 80% of a 4790k, it barely matches a 4690 and that only with highly threaded shit like x264 - in everything else, it greatly falls off. Count in IPC and we're in the 60s, albeit Bulldozer vs Haswell is a little bit difficult to compare due the first one's inconsistent performance through its modular design. Nonetheless, unless you want some cheap threaded number cruncher, the 83XXs aren't an option right now.
What I meant is something like the Phenom II. It didn't blow away Intel's Nehalem, but it was quite decent at standing up against it, at least for non-enthusiasts, with an IPC around 70-80% and similar clockspeed limits, without burning coal like a steam train. 940s and 955s were all over the place. A lof of people also bought the Thuban chips.
Bulldozer even was a step back from the K10. The first iteration had a lower IPC and higher power consumption WHILST being based on a lower lithography. Even just shrinking their old chips probably would've been a better choice.
Considering that AMD's current FX chips are still based on Piledriver, which is already succeeded by Steamroller and soon Excavator with higher IPCs already, it's very likely that Zen, if it's going to be a major jump, will probably at least reach out to Haswell, which in turn is close to Skymeme and I doubt that Kaby Lake is gonna be more than the usual 10%.
10nm Cannonlake will most likely not approach before summer 2017, if even at this point. Intel even stated so themselves by giving up their ticktock. Zen is going to launch in Q4 of this year, so late 2017/early 2018 will be most likely Zen+.
>Maybe I'm just a negative guy.
You're just acting retarded.
>>52325433
>barely matches a 4690 and only in situations that take advantage of 8 loads
An fx 8350 can beat a haswell i7 when all 8 threads are being utilized.
Not in every application, but the fact that the fx 8350 was meant to compete against lga1155 just shows how powerful it was at the time.
I accept that it has a shit ipc and loses in almost every application that uses fewer threads, but it's still capable of outperforming a fucking haswel i7.
>>52315767
How can anyone not afford one?
It's the one part you'll be able to use for the next 10 years, no matter what else you upgrade.
>>52316121
>, want to be really secure with a server
So you admit they're better than the others?
Then I'm fucking buying them. I have so many peripherals I don't want no fucking wavy shit.
