It's Over

it's absolutely nothing new OP

>>57970839
If AMD sells good Intel will get royalties, same in reverse.
There won't be a winner as long as either does x86 cpus anyway.

Isnt this kind of their thing though? Intel releases a processor that's better than their last one, then AMD releases one that's better than intel's new one like 6 months to a year later, then a few months after that intel announces something else thats a little better than that, rinse/repeat.


They've literally had these processors made, ready to ship like 10 years ago, but then they strip them down and release them piece by piece (think: dlc for games, but way more expensive and wasteful) when their main competitor releases their little bits and pieces of their max potential.

>>57970916
no what the fuck are you talking about

no

>>57970916
Wew lad, nice conspiracy.

>>57970916
AMD hasn't released anything that comes even close to Intel since 2012, following it up with slightly overclocked or TDP dropped revisions of piledozer. What you described perfectly fits AMD vs Nvidia though.

Typical back and forth. Neither will go bankrupt since they command large assets.

File: IMG_0008.jpg (57KB, 356x448px) Image search: [Google]

57KB, 356x448px

>>57970839
>Shitty gaymer name
>Giant poo splash in background
holy fucking shit, this is too good to be true

It's OVER only 3.4GHz meanwhile Kabylake is 4.5Ghz AMD stock dropping like pajet poo hhahaha

>>57970956
Ya pretty much this. Only good AMD CPU in a while has been the 8350. Since that it's been intel all the way really.

However I am hoping they have a new 8350-like processor in the new bundle. One that can compete but for a lower cost. I'm really looking into a 6800k but if AMD has anything marginally close (probably wont but here's to hoping) I'll end up going with that.

>Ryzen
>ThreadRipper
>Grok
>Vara
>Joro
>Jitzu

>>57971140
Someone from marketing went to Japan I see.

>>57971114
>Kabylake is 4.5Ghz
Those are quad cores, this is going to compete with Intel's 8-core chips which run about the same speed (3.2 - 3.7ghz) as AMD

>>57971140
>shitty Mortal Kombat fanfic character names

AMD cant sink this low. Shit like FX or Phenom sounded perfectly fine and professional, it seems AMD thinks all they can sell to is dumb gamers with low budgets

>>57971219
>implying Sandy Vagina, Fiery Bridge, Baby Lake, and Caribou Coffee are good names for processors

Ryzen makes sense. Playing off of Zen and being risen from the ashes. Better than the nonsense intel comes out with.

File: trashman7.gif (2MB, 580x433px) Image search: [Google]

2MB, 580x433px

>>57970839
>RYZEN LMAO XTREM
>3.4GHz

>>57970916
The last AMD processor that could come close to intel was the phenom 2

Even if Zen is as good as they claim anyone using a 2014 broadwell and up wont care.

>>57971140
Ryzen
>Family was murdered by the Intel clan long ago
>rose from the dead using the powers of the great Pajeet, now seeks to get revenge
>Uses his poo bending powers to make people meet their shitty demise

ThreadRipper
>member of the Enterpise clan
>has the power of over 32 cores and 64 threads which he uses to shred his enemies
>his only weakness is that he cannot take on single threaded foes

>>57971361
ThreadRipper is probably the name of their SMT implementation or something like that.

>>57971278
But none of those are actually Intel's names? Stop pretending to be retarded.

>>57970839
>posts literally nothing of value
>claims competition is fucked
/g/ in a nutshell

>>57971304
>muh gigahurts
is it 2001?

>>57971849
>too much of a low IQ shit dick to understand simple hyperbole

>>57971989
Amd throws cores at everything but cores are useless for most people

>>57971278
i-ill just use the same tactic children use to to make threatening things less scary and name them poopy crap names haha got you intel
stop having autism

>>57971278
What is Intel's naming scheme anyways? I don't see a pattern other than being named after lakes or something

>>57972193
Not an argument. Fag.

>>57972403
Its all over the place, but they're just sticking with names of random lakes for now with their mainstream consumer chips.

>>57970839
ry ZEN?
why not just ZEN?

>>57972466
because the core architecture is zen, but the architecturial adjustment to each market bracket also need their own trading name

>>57972466
Zen is too generic and already widely used to copyright and use for a product line.
Zen is the codename of the core architecture, and that doesn't need a copyright.

The real question would be why they didn't stick with Summit Ridge.

>>57971304
Broadwell-E 8 core only has 3.2GHz

>>57972514
maybe Ryzen is their new marketing name like FX with bulldozer architecture

>>57970839
3.4ghz lol

File: 001.png (142KB, 415x367px) Image search: [Google]

142KB, 415x367px

>>57970839
>tfw was about to pull trigger on 6800k

thank fuck I waited

>>57972514
>calling trademarks copyright

>>57973082
Same, probably still going that way but waiting a bit just in case wont hurt.

the 6900k has a base of 3.2 and a boost of 3.7, so this part will be extremely close in performance. if the rumor about the specially binned overclockable SKU is true, AMD might finally have something that's objectively better than the competition.

I hope the rumor about the $500 top price tag is true but somehow I really, really doubt it. If it does turn out to be a 6900k killer AMD will probably price it to match.

>>57971361
RYZEN FROM THE ASHES TO RIP YOUR THREADS APART, BITCH!!

the fuck is SenseMI?

>>57973437
If that 6800k Competitor is offered at a better price I'll be interested if it's even in the ballpark.

File: trash.jpg (31KB, 300x400px) Image search: [Google]

31KB, 300x400px

>3.4 ghz

>>57972967
>>57974039
As opposed to 3.2GHz of the i7-6900k?
You really must try harder to make Zen look bad.

>>57970839
>>57971140
Is this b8 or those are the actual names?
Why does this miserable company can't get _anything_ right?

>>57971361
>rose from the dead using the powers of the great Pajeet
Daily reminder that the CPU division is practically pajeet free

Single. Thread. Performance.

I am looking to replace my i3 6100.

>>57974324
Ask your parents for a 7350k for xmas.

>>57972060
Can we add more pipelines as well?

>>57970879
lol are you retarded.
That's not how the x86 license works

>>57970839
Enjoy getting buttfucked by PSP. All of the Intel users are getting ready for the great uncucking. We can disable Intel ME on most of our CPUs. Stay cucked, AMD nigger. You are literally a poo in loo.

>>57975328
Even really new ones?

>>57975409
Yes. Check the catalog. There's a thread up.

>>57975429
I have no idea what Intel's current processors are called since they've kept the names i3, i5 and so on for years just adding extra numbers.

>>57975555
Use the catalog. Find the link. Click it. It's not that hard.

>>57971199
But who needs 8 cores? 4 cores with an entire Fucking 1.0Ghz advantage will blow this thin away 90% of the time.

>>57975638
Who needs more than 256K.
Fucking moron

>>57975638
Not if you use an operating system with scheduling instead of templeos

>>57975638
Are you this dense? Having high clocks on a 8 core part are pretty much a guarantee of quite higher clocks on lower core parts
>4 cores with an entire Fucking 1.0Ghz advantage will blow this thin away 90% of the time
You should go back to 2011

>>57974307
haha... white man cant make cpu... this is why amd is dying...

File: trustme.jpg (18KB, 256x300px) Image search: [Google]

18KB, 256x300px

don't forget to buy AMD guise and watch Clit TV

>>57970839
>8 core so its going to be a slightly better FX8350
>3.4 GHz
>2017

>>57975741
It's going to be more than slightly better, not to mention that these are the first AMD processors to have SMT so that's 16 threads compared to the 8350's 8

>>57975741
>slightly better
>40% higher ipc

shitposters are really clutching at straws to make this thing sound bad. just stick to the edgy pajeet memes.

>>57971199

5960x and 6900k can both clock to 4.5ghz with ease.

if the rumors about zen being sandy vagina ipc and not clocking above 4ghz are true, then it's a DOA product, especially if it's $500 MSRP.

>tfw RY'ZEN sounds so cool I might just get it

I'm kinda scared that AMD isn't fucking up for once.

Maybe release a 10 core version named Rikka and let the Internet get to work.

Intel better not underestimate meme faggotry either. If everything goes right 2017 will be the year of Trump and hardware with weeb names.

>>57975774

fx9590 was clocked 47% higher than the 3.4ghz zen sku, so if it's 'only' 40% higher ipc it will be worse performance than vishera.

Just bought a mini ITX case, should I expect an 8 core APU or should I just go with the Atom I had picked out?

>>57975853
40% over excavator, not bulldozer or piledriver. that puts it in the range of haswell.

>>57975876
>40% over excavator, not bulldozer or piledriver. that puts it in the range of haswell.

no, it doesn't. it's a little slower than sandy and ivy bridge clock for clock. https://www.pcper.com/news/Processors/Leaked-Geekbench-Results-AMD-Zen-Performance-Extrapolated

>>57975922
>geekbench results from some random and unconfirmed test platform
Tech. Illiterate. Retard.

>>57975796
>if the rumors about zen being sandy vagina ipc
This was debunked when the demo was done months ago, it has pretty much the same IPC as Skylake, and thus at parity with Kaby Lake
>and not clocking above 4ghz are true
This was the last rumor still holding, but this leaks shown Zen clocking higher than comparable Intel parts, while keeping a lower TDP
>then it's a DOA product, especially if it's $500 MSRP.
Even before the last rumor was debunked, Intel was already making moves never seen on the market, like the leaked 32 core Xeon that will go against the 32 core Naples, or the rumor of unlocked i3's
>>57975922
>posting the old ass Geekbench
The Blender demo is out there, stop shilling with ridiculously stale crap

>>57975934

not an argument

>>57975949
>not an argument
>redditspacing
You have to go back
The first line is an argument, now are you mad that you aren't getting paid? Fucking pajeets shitting up /g/

>>57975947

>believing AMD's marketing department, who has a history of exaggeration and lies, over a real world benchmark

nothing was debunked, stop trying to wiggle out of this and stop hyping yourself up. nobody else except you fanboys are going to buy this DOA product so if you temper your expectations now it will make the asspain a lot easier to deal with in a few weeks.

>>57975961
>someone who posted facts that dispute my worldview is a redditor

last i checked it was redditors who live in their own world and can't handle truth, stop trying to project

>>57970839
Fuck that shit and its muh more coars my 3570k still stronk for years

They have anything cool in the actual budget (under 100 bucks full build like am1) going on?

Its the only reason zen would worth shit

>>57975949

Remember when a geekbench leak showed kabylake with a 30%+ lead over skylake despite Intel themselves saying that isn't the case?

So, what kind of performance increase can I expect to get over my Q6600?

>>57970879
Intel and AMD have to share core libraries and IP's. Not revenue.

would this even be considered an upgrade over a 3930k @ 4.5ghz

>>57975991
about tree fiddy

>>57975963
Blender is a real world benchmark, Geekbench has always been shit, also it's from an unconfirmed chip, on an unconfirmed platform, and even then you can't really even know if it's Zen since that can easily be faked
And that's not even counting that Geekbench puts Apple's A9 chips faster than most Intel chips at single thread workloads
Also, the Blender run was done under the scrutiny of the media, and AMD made public everything they used, including their Blender builds
>who has a history of exaggeration and lies
Straight up lying? Great anon, now you will surely get paid
>>57975975
You're the one who literally dismissed an argument with >not an argument
heh
>>57975977
>They have anything cool in the actual budget (under 100 bucks full build like am1) going on?
Not until 2017Q2, the first wave of Zen products it's going to be server and HEDT parts
>Its the only reason zen would worth shit
Being at parity and even beating Intel is hueg
>Fuck that shit and its muh more coars my 3570k still stronk for years
After Zen Moore's law will go back into effect
>>57976001
Not really unless you need more coars

>>57975977
yea if it's not 2x c2d t75xx im not interested

>>57975838
It's an awesome EXTREME 90s style name

AMD is going to make computing great again

>>57975741
for fuck's sake nigger have you even read anything before you posted

>>57976014
>Blender is a real world benchmark,

depends on the workload, i'd expect anything AMD's ministry of truth puts out would use a workload specifically crafted to give them the best performance possible for their uarch, they would never show what the real world performance is like because even in their best case they were only on par with haswell, which is almost 4 years old.

>>57976157
>depends on the workload, i'd expect anything AMD's ministry of truth puts out would use a workload specifically crafted to give them the best performance possible for their uarch
Good thing the workload was made public and no one found anything wrong with it :^)
And even then, different workloads wouldn't make Blender rendering substantially different
Even if it did, then it would show the AMD chips beating Intel, how does it feel to disagree with yourself?
> they would never show what the real world performance is like because even in their best case they were only on par with haswell, which is almost 4 years old.
Literally wrong, it's at parity with Skylake and thus Kaby Lake, Intel preparing a 32 core Xeon proves this, they wouldn't need to do anything to keep their lead in the server market if Zen had the same IPC as Haswell
Also, 4 years old Haswell is still within 10% IPC wise with Kaby Lake

>>57976157
>i'd expect anything AMD's ministry of truth puts out would use a workload specifically crafted to give them the best performance possible for their uarch

it's exactly the software intel recommends to bench their own processors.

>they would never show what the real world performance is like because even in their best case they were only on par with haswell, which is almost 4 years old.

because Broadwell-E is still current; intel haven't updated the architecture for their enthusiast line in all that time. If the 6 and 8 core zen SKUs are comparable to haswell in IPC and clockspeeds then AMD have a competitive product in their hands, depending on retail price. It's that simple.

>>57973500
SMT

File: strawman.jpg (26KB, 400x400px) Image search: [Google]

26KB, 400x400px

>>57976199
>it's at parity with Skylake and thus Kaby Lake, Intel preparing a 32 core Xeon proves this, they wouldn't need to do anything to keep their lead in the server market if Zen had the same IPC as Haswell
>Also, 4 years old Haswell is still within 10% IPC wise with Kaby Lake

If I was a crow I would be fucking terrified.

>>57976228
>pissphone being retarded
Go back at believing AMD joined Khronos in 2013

>>57975838
who the heck cares? it will have some kind of number that everyone will remember , 80% of i5 buyers have no idea what lake it is

Should I buy 6800k or wait for this.

>>57976281
Wait, prices are definitively dropping after Zen launches

>>57976199
>Good thing the workload was made public and no one found anything wrong with it :^)

people found that it fits entirely into L3 cache at all times and produces similar results on any CPU released since nehalem. there's nothing conclusive to draw about performance from it at all except for l3 cache bandwidth and latency.

>And even then, different workloads wouldn't make Blender rendering substantially different

workload is everything, cycles has tons of features and things you can tweak that can radically change how much time your render takes.

>Even if it did, then it would show the AMD chips beating Intel, how does it feel to disagree with yourself?

point is that it wouldn't be that fast in real world usage and therefore isn't a real world benchmark. we've already seen multiple real world benchmarks that put 8 core zen on par with nehalem and sandy bridge chips performance wise.

>Literally wrong, it's at parity with Skylake and thus Kaby Lake, Intel preparing a 32 core Xeon proves this,

the 32 core xeon is an unfounded rumor and the chip from that leak wasn't even on the correct platform (the chip shown was for intel's HPC platform). it's much more likely that naples will be competing with a 26 or 28 core chip on whatever platform skylake-e/ep comes with.

>>57976296
Thanks senpai

>>57976296
probably not, intel is too comfortable they won't do anything intill they see volume drop

>>57976210
>If the 6 and 8 core zen SKUs are comparable to haswell in IPC and clockspeeds then AMD have a competitive product in their hands, depending on retail price. It's that simple.

of course they would be competitive if they could achieve comparable IPC and clockspeed. too bad we already know they are 5 years behind IPC wise and aren't able to clock as high as Intel can.

maybe that will improve in the future, but as it stands now Zen will disappoint AMD's fanboys just like how phenom, bulldozer, fiji and poolaris did.

why do so manys poster on /g/ want amd to fail?

even if you never buy their products having a competitor for intel/nvidia will save you money

>>57976370
Because fucking morons post on /g/.

>>57976321
>people found that it fits entirely into L3 cache at all times and produces similar results on any CPU released since nehalem.

source

> we've already seen multiple real world benchmarks that put 8 core zen on par with nehalem and sandy bridge chips performance wise.

source.

>>57976321
[citation needed]
>we've already seen multiple real world benchmarks
We literally haven't, leaked Zen benchmarks have been scarce at best, quite less than a dozen have surfaced
>the 32 core xeon is an unfounded rumor and the chip from that leak wasn't even on the correct platform (the chip shown was for intel's HPC platform)
Zen is targeting HPC as well senpai, this was literally confirmed today with the new compute Radeon slides

>>57976342
Repeating something won't make it true, IPC wise Zen has been confirmed to be at near parity, and clockspeed wise they are faster on their 8 core variants
Polaris literally brought back around 20% of marketshare to AMD, along with the hueg gains their stock has been experiencing
Phenom was a great chip all around, you don't remember anything at all, do you?
Or are you a mad shill that's mad that he didn't buy into Su's rocket to the moon? :^)

>>57976321
>trying to regurgitate things you don't understand
>being a tech illiterate retard

Blender loops through L1, not L3, and even if AMD tried to alter their build to exploit some nuance of their cache system, Blender has a default feature that analyzes the cache performance of a CPU and optimizes its codepath for it when it starts rendering.

You're a fucking moron.
You're probably the same tech illiterate retard troll who tried to claim the Blender demo was only competing against a 6 core Broadwell-E instead of an 8 core.

>>57976406
>Or are you a mad shill that's mad that he didn't buy into Su's rocket to the moon? :^)

>tfw was going to buy when they were hovering around 2.50

could have been rich ;___;

>>57976406
>Phenom was a great chip all around, you don't remember anything at all, do you?

it was great 2 years after it was released once AMD fixed all the problems with it, phenom 1 was complete garbage.

>>57976411

applications have no control over what data goes into cache nor the level of cache that data goes into, they can only organize data in a way that can best utilize cache if the data ends up in cache. if you have less data being used at any given time than space in cache, then obviously all your data is going to be in the cpu's cache.

> Blender has a default feature that analyzes the cache performance of a CPU and optimizes its codepath for it when it starts rendering.

>hurr codepath!

there would be no seperate codepath as there are no instructions related to cpu cache in x86 and x86-64, there's nothing mandated by it that says you need l1/l2/l3 cache

>>57975872
Even Raven Ridge won't have 8 cores beside the iGPU at the end of 2017. It's not even sure Bristol comes to the desktop market, so go with something thats available now.

>>57976406
>Polaris literally brought back around 20% of marketshare to AMD
By litterally you mean helped AMD bounce back from about 20% to 30%

What I feel helped AMD bounce back was their focus on value, meanwhile nvidia said fuck it and chased profits instead.
Its good for AMD to have a decent hit with Polaris, but I still stand firmly that the company is not in the clear and has a shaky foundation with very high expectations to meet.

>>57976488
If I can get a 4 core APU I'd be content as that'd at least give me 8 threads

>>57976478
>applications have no control over what data goes into cache nor the level of cache that data goes into, they can only organize data in a way that can best utilize cache if the data ends up in cache
This is literally what anon said with the codepath reference you 'tard

>>57976508

there would be no seperate codepath, the difference is just in how you lay the data out in memory (and it would be laid out the same as amd and intel have both used 64 byte cache lines for decades)

we can be sure AMD not gonna pull AM5 out of pocket next year?
I'd be happy for once not worry about changing motherboard

File: 08.08.16_AMD_RavenRidge_teaser85.png (207KB, 1006x354px) Image search: [Google]

207KB, 1006x354px

>>57976504
they'll have one but don't expect to see it until the middle of next year. if you want a stopgap get an equivalent bristol ridge apu and upgrade later.

>>57976478
Approximate Minimum Degree
Retard.
You're regurgitating bullshit from one autist on twitter who was immediately corrected on his bullshit. Months ago. Retard.

>>57976528
That works out well, I've got to get a new car before I build a new computer and I already dropped $280 on a case because it was on sale

>>57976535
Just say compression you fucking nerd.

>>57974675
kekd

>>57971332
>Even if Zen is as good as they claim anyone using a 2014 broadwell and up wont care

Speak for yourself. I'm using a 6700K and plan to switch to an octa-core Zen chip if the single core performance is there.

Why the fuck do i have to sign up for the New Horizon livestream?

>>57976714
why bother? just read important bits later

>>57972514
Reminds me of how intel tried to trademark the letter i for ages. I guess the settled for i3, 5, and 7

If you enjoy lots of "oh yeah if you're one of those AMD hipsters we could never quite work out how to stop this random glitch happening, you should've gone Intel like the rest of the world" in your games, aye.

WOLVES HUNTS IN PACKS ARYZEN!

All I need is a zen+vega APU based NUC. A 15watt TDP tiny box for non-AAA games and emulators. It'll give skull canyon a run for its money.

>>57976370
fuck off poorfag

>>57971174
their marketing team is literally japanese.. they came with amada kokoro character afterall..

>>57976370
someone posted yesterday the slides of the new vega line
and it was one guy shilling for nvidia for like 30 mins literally comparing stuff from 10 gazzilions years before to something that is not even out yet..

>>57977597
Amada was created by a local shop.

>>57977638
every single tech anime from the big companies came from the marketing team...

>>57977688
>the marketing team
Do you mean this guy:
>>57977638
? :^)

I don't really care about faster speeds, did they manage to release a good <4watt CPU?

4 watts is where it's at.

>>57977688
Stop trying to talk out of your ass. It was a sketch put on an end cap of one small PC shop in Japan. It had nothing to do with AMD themselves. Its fan art at best. Dipshit.

>>57970839

so realistcly speaking how much of a perf boost will I get compared to my fx8350 oc @4.6ghz?

is it a worthwhile upgrade?

i wish gaymen /v/ermin would stay the fuck away from discussing hardware besides graphics cards.

>>57977739
That will never happen on /g/raphics cards.

>>57977723
you do realise that amada was literally on the gpu coolers right? they were literally being sold with them not stickers not fan art shit

AMD needs to do well this round for the sake of the market. Both Nvidia and Intel have basically gotten less ethical and more expensive (some more so than others) without AMD offering serious competition in terms of performance. That's what we need. AMD has refined their craft bigtime - their drivers are no longer shit, hell they are supporting Linux better than Nvidia at this point, they came out with really solid GPUs in the RX 460 / 470 / 480 when it comes to their range... now its time to take a swipe at the big guns!

On the GPU side we need Vega to be a serious competitor to Nvidia's high end cards, at a much lower price. That's where AMD's high end GPUs shine, being $400-500 cards that are equal, better, or sometimes within 10% less than NV's $600-800 range cards. Ideally it would be nice to have Vega's "big" card be a parallel to this dealing with the 1080 Ti (or whatever the HBM2 next gen card migtht be) and also have a cheaper version to be a "better price/performance than the 1070 Ti" type.

On the CPU side, Zen needs to be a return to form of the Phenom 2 era or Socket 939 glory days. It needs to compete with Intel's latest gen mainstream high-tier proc like the i7 6700K in both single thread and multi threaded performance and be on a chipset that is equal to Z170 or better with latest high end features. Ideally overclocking well, it ought to again be cheap enough as a platform to be worthwhile. I'd love if it lacked the "Intel ME-style" blackbox/NSA backdoor shit too, but we can all dream.

>>57977802
wait wait a minute
you are saying that nvidia was more ethical in the past? WHEN last time i saw them being ethical was back at riva tnt era..and since the 5xxx time nvidia is pretty much dirty up to their head..

>>57977726
About 40% extra IPC but 26% lower clockrates so the total performance should be 3.6% higher.

1.4 * 0.74 = 1.036

>>57977795
How fucking retarded are you?
It was drawn up by an employee of the store. Years ago. It had literally nothing to do with AMD. They didn't create the character.

Get murdered, you retarded faggot.

File: 1456813501773.png (461KB, 730x1080px) Image search: [Google]

461KB, 730x1080px

>>57977832
>he thinks ipc scales linearly with clock rates

POINT AT HIM AND LAUGH

>>57977851
/facepalm...
seriously you need to visit Akihabara and ask them about it cause you are clearly out of the acg league..

>>57977832

ah makes sense, thanks for the info

but doesn't it have double the cores, on't this make a difference, if properly used by the program?

also I hope it becomes highly overclockable, I mean I can reach on my cpu 4.6 ghz on a non water middle-class custom cooler and the temp wouldn't go higher than 62c with the highest tress test

>>57977872
>I'm just going to keep talking out of my ass

>>57970839
Substantiate your claims, faggot.

>>57977863

well elaborate then, how does it scale?

File: ipc uplift.png (10KB, 582x162px) Image search: [Google]

10KB, 582x162px

>>57977832
Its 40% higher IPC than Excavator, not Piledriver.
Your napkin math is still laughably wrong regardless.

having more than 4 cores is still dumb for ordinary user.

8 threads is just fine.

>i7 2600 here

>>57977893
Exponentially

>>57977925

nice shitpost

>>57977910

so nearly 25% better than fx8350 oc @4.6ghz

nice

>>57970839
The naming sense is fucking cancerous from AMD. Something like Cannonlake? That sounds badass. What the flying fuck is a Ryzen? Zen was decent, far better than the names for previous generations. This is a step right back

>>57971140
>ThreadRipper
>Jitzu
This cannot be real.

>>57977802
Amd isnt going to mske something that comes within 10% of the competition cost hundreds less. Look at your own example of the 400 series for, well, example.
The cards are neck and neck with nvidia counterparts that released later, but they havent cut any prices to make their options a significantly better value. If they can compete directly and turn a profit they will. Theyre a company whose goal is to make money, they are not your best friends.

>>57977991
ryzen=risen god damn it /g/ is full with spastic monkeys

>>57977989
Still no.

The IPC uplift is per core.
You need to factor in Bulldozer family arch has a CMT penalty when both cores in a module are utilized. You also have to factor in the multicore scaling factor. Bulldozer family arch had major regressions in this area, to the point where Llano still had better scaling than Kaveri.
No consider that Zen is going to be better in all these areas, and still have twice the threads.

The 2.8ghz ES sample that was spotted in AoTS was performing around 80% better than an FX 8350 clocked at 4ghz.

>>57978031
80% better per clock

>>57978031

holy shit, now that's quite a difference, if it overclocks good, I would say it's pretty futureproof for a least 4 years

considering that my current build is running like this stable since 4 years and has been good enough for everything for 60fps

but I have a 144hz monitor now, and I'm not considering switching back

>>57978087
4 years? if we talk about dx12 this shit is future proof for at least 6....

>>57978012

that attitude is toxic. It is possible to be profitable without being a shithead because you figure "maybe I can be just a tiny bit more profitable in the short term if I" and then pull a boneheaded move. AMD has done exactly as mentioned in the past. For instance, their R9 290X competed with highest end Nvidia cards at the time and for a much lower price - even when resellers jacked up the price because of cryptomining it was still less than NV. The same thing happened going all the way back to the 4000 series pretty much. On the CPU side, when it was the Core and Core 2 versus Phenom, the Phenom 2 X4 and X6 Black Edition chips competed bigtime with both dual and quad cores for a lot less money.

The current situation is the anomaly. Because they basically don't have nearly the money Nvidia does and they already priced the RX series well, they can't afford to drop the price and still complete their rollout.

>>57978102

hmm now def. excited, I wonder what the pricepoint will be, I hope not too much

>>57978166
Leaks from earlier suggested $350 for the lesser 8 core and $500 for the greater 8 core. No word yet on what the difference between the two is besides a slight change in clock speed.

I've been saving since I got my Phenom II waiting for it to become irrelevant. I've bought two motorcycles with money put aside for new hardware waiting so dropping that kind of money is a non issue. If you're really poorfag the 4 core models go as cheap as the FX8350.

>>57971134
I didn't care for any of the performance cpus since phenom. A8-7600 was the only good recent sku.a

>>57978031
It is going to be 2x an 8350 in single core.

>>57970839
estimated cost ? estimated release date ?

>>57978108
instead of pretending that enormous multinational corporations like AMD actually care about consumers and not profit, you could instead join reality and realize that AMD is no better than Intel or NVIDIA when it comes to price gouging and ethics. The one and only reason they sell their shit for less is because otherwise no one would buy it.

>>57972514
>The real question would be why they didn't stick with Summit Ridge.

I'm fucking boggled why AMD didn't use Summit Ridge as the actual product name. Summit Ridge sounds cool as fuck and normal. Ryzen sounds like a new Razer product.

>>57978289
As long as they didn't do the sr3/5/7 thing to mimic the i branding then it definitely would have been better.

this shit is meaningless anyway. if the i brands came out today you'd think they sounded dumb and confusing too. everyone's just used to it.

File: 1479302187673.png (196KB, 320x303px) Image search: [Google]

196KB, 320x303px

>>57970916
>They've literally had these processors made, ready to ship like 10 years ago,
Wut

oh yeah because 10 years ago they knew exactly when we''d get to 14nm finfet

fucking retard

>>57976240
Most normies see a nehalem i5 and instantly think it's faster than a haswell i3

It's an interesting phenomenon, I'd argue it's the reason intel mopped floor with this naming scheme, it doesn't get easier for normal fags.

Imagine if AMD had have stuck with FX instead of A-, FX sounds like power while A- sounds like "oh shit intels doing well with i- can we get on this gravy train as well?"

3.4? wow....
my 4690k is 4.4

Real question is, are those poocores? if my suspisions are correct then those are poocores. OC'd K i5s will probably beat it in most multithreading becaue intel has real cores and not poocores. 4fast cores can easily beat 8 poocores
it will probably beat the i5 on certain benchmarks but it will lose the majority

>>57978519
if zen is as good as it is made out to be, even the base 4 cores will btfo any i5 and be on par with i7s because of smt

>>57970839
Price?

>>57978564
about tree fiddy

>>57978564
around 350$

>>57978577
>>57978582
Damn do we know anything about the 200-250$ cpus?

>>57978594
4 core with Multithreading or 6 core with multithreding. They want to build rely on zen as answer to intels extreme platform. The cheaper ones (4cores) will follow in the future (end of 2017) as APUs

>>57975233
Underrated

>>57975741
That the equivalent of a 4.75Ghz 8 core excavator chip with l3 cache.

Quite good IMHO

File: 1465851696142.jpg (6KB, 250x164px) Image search: [Google]

6KB, 250x164px

>Q1 2017

>wait another half year to get this refurbished FX cpu

>>57975949
You're absolutely right, geekbench is not an argument and should not be used to compare to even slightly dissimilar products. It's fine for comparing a core i5 to an i7 though, as long as they're both from the same generation and running the same OS.

>>57980525
you know what quarter means?

>>57980644

I sure do but AMD doesn't

should've called it Rypoo

day one sold out

Glad I went Intel. They tried.

>>57980876

this

>>57970916
>>>/pol/

>>57981386

>>>/fa/

>>57981432
>>>/s4s/

>>57981497

>>>/mlp/

>>57970839
>" the dark knight has ryzen"

>AMD ZEN SMT won't work on Windows 7
DOA

>>57981699
source

>>57981699

l m a o

b t f o

t

f

o

>>57981711
>https://www.techpowerup.com/225627/smt-and-power-management-behind-kaby-lake-and-zen-windows-10-restrictions
>AMD "ZEN" processors introduce simultaneous multi-threading, a feature that exposes each physical core as two logical CPUs to the OS, for better utilization of on-die resources. Intel's implementation of SMT is the HyperThreading Technology (HTT), and has been around for over a decade. AMD's SMT implementation isn't identical to that of HyperThreading, with the two threads on a CPU competing for resources in a method unique to AMD. This can't work without the OS kernel and scheduler being aware of the method. You'll remember that Microsoft had to update the kernel and scheduler of Windows 7 in a similar way, to optimize it for "Bulldozer."

>>57976321
>we've already seen multiple real world benchmarks that put 8 core zen on par with nehalem and sandy bridge chips performance wise.

We've seen literally ZERO confirmed real world benchmarks except for the AMD's demo.

Just kill yourself intelshill.

>>57981761
What happened to license agreement for Win8 that says they have to continue supporting it for a couple years yet? Is this legal? I don't think they can just back out of a license agreement like that.

Who Phenom ii x4 here? Feels good being a poorfag

>>57981761
>tech blog quoting another tech blog
>pure speculation
>no links to any amd literature

try again

>>57981761
And I thought skylel was bad

>>57981844
I found this when googling. https://fossbytes.com/intel-kaby-lake-and-amd-zen-processors-wont-support-windows-7-8/

If it's true it just makes me really hate Microsoft. I don't want their telemetry infested forced update bullshit.

>>57981821
Supporting Windows 8.1 doesn't mean they have to add compatibility with new hardware.

>>57981873
Not only is that old as fuck news that nobody cares about cuz
> Windows
But that's fucking backwards. Windows won't support them not the other way around

>>57981901
>Agreeing to support our operating system doesn't mean we need to make it support new CPUs that come out
That's bullshit even if it's true.

>>57981909
>no one cares about windows
It's the largest fucking operating system.

>>57976370
>even if you never buy their products having a competitor for intel/nvidia will save you money

Yeah it hasn't helped in either case for years.

File: __reisen_udongein_inaba_touhou_drawn_by_hisae_hisae_collect__13d6f10c6bd17b46b86fe5f01ad5709e.jpg (141KB, 400x581px) Image search: [Google]

141KB, 400x581px

>>57970839
did someone say Reisen?

>>57981909
>But that's fucking backwards. Windows won't support them not the other way around
It says that in the article. Are you unable to read?

File: pepe00095.png (10KB, 500x500px) Image search: [Google]

10KB, 500x500px

>>57981761

amazing news

now /g/ has to switch to win 10 wether they like it or not

>>57981944
Or just use Linux and run it in a virtual machine as needed. Preferably with it locked the fuck down so it can't call back to microsoft.

When was the last time AMD beat Intel?

>>57981959
too much work to pass through a GPU

>>57981944
or I can keep using my FX chip because it works just fine

>>57981967
2012

>>57981973
>too much work to pass through a GPU
You just turn pass through on it virtual box. What the fuck are you on?

>>57981944
I'll probably just install Server 2016 along with OpenSUSE when I get my new Zen system, since I get it for free via university.

>>57981976
>You just turn pass through on it virtual box. What the fuck are you on?
Are you sure it still doesn't take a 50K word essay for the basic steps to do a PCI-e passthrough?

https://www.youtube.com/watch?v=16dbAUrtMX4&t=626s

https://forum.level1techs.com/t/gta-v-on-linux-skylake-build-hardware-vm-passthrough/87440/2

I am wondering if they have finaly managed to beat Sandy Bridge (which is over 5 years old already) in IPC.

Oh and it better overclock like a motherfucker, because 3,4ghz is shit, my 2600k runs at 4.4ghz

>>57982025
>octocore running at 3.4GHz is shit because a quadcore can run at 4.4GHz
Fun fact: The highest clocking sandy bridge-based octocore only reaches 3.1GHz.

>>57982002
Yes I'm 100% certain. I've set up virtual machines through virtual box before. The official documentation on PCI-e passthrough is literally just a couple paragraphs for a reason.

>>57982068
Octacore

>>57982126
octocore

>>57982068
>only 3.1Ghz
pathetic

>>57982002
>50K word essay
>install propriety graphics card drivers if you haven't already
>enable PCI-e pass through for your graphics card's PCI-e via a single CLI command
>assign PCI-e to desired virtual machine
>50k word essay
Man you're fucking dumb. The commands are in the documentation you only have to run a couple commands and not be a fucking moron.

>>57982160
octopussy

>>57982160
>pathetic
Well, yeah. That's sort of my point. You're not going to get the kind of clocks from an octocore you'd get from an overclocked quadcore, especially if the said octocore is limited to 90W TDP.

>>57982025
You also don't have 8 cores.

>gee anon, how come your indian let's you have 16 VMs?

File: 1478109550871.png (6KB, 390x470px) Image search: [Google]

6KB, 390x470px

>ITT
>MUH MUGAHUTZ
kek, fucking kids

why the hell would i want EIGHT cores?

>>57982160
Octacore

>>57982240
I can get 8 core AMD cpu's right now, and they suck fucking dick, just the amount of cores says nothing

>>57982349
They're 4 core with 8 threads. These are 8 core with 16 threads.

>>57982384
You're an idiot, current AMD cpu's don't have SMT

>>57982384
No, they're 4 module with 8 cores. CMT is not SMT.

>>57982333
>333
traps confirm

octocore

>>57982417
>>57982420
There are only 4 physical cores not 8.

>>57982420
There are not 8 cores. There are 8 threads. There are only 4 cores.

>>57982516
No, there are 8 physical cores organized into 4 pairs that share some resources. Unless you're claiming that early CPUs in fact had 0 cores.

>>57982544
There are not 8 physical cores in AMD's FX series. There are 4.

>>57982557
Except that's wrong, retard. There are 4 MODULES and 8 CORES. A module is not a core.

>>57982579
There are only 4 physical cores with 8 threads.

so what's the Estimated retail price?
200$ jew dollars?

>>57982589
You can keep repeating that all you want, but it doesn't make you any less wrong.

>>57982622
You sound upset.

>>57982557
There are 8 physical cores. They share some resources which hurts performance but they are in fact independent cores.

They are basically shit cores with one less FPU or something, so basically useless as nearly everything gets calculated as floating point, because that's how computers just calculate.

It's basically 8 proper cores, but only when you live in the 90s. Later CPUs all shipped with one FPU unit per core. So when you look at it from todays perspective it's more like 4 cores.

Intel gives 8 cores but tells you its 4. AMD tells you it's 8 cores but it's 4 and 8, depending on how autistic you are.

>>57982684
One fpu per two cores

File: poo.jpg (33KB, 460x460px) Image search: [Google]

33KB, 460x460px

>>57982747
Yeah, exactly. I somehow missed to point that out.

>>57970839
They should make a 12 core version and call it a Dozen.

>>57971107
>>Giant poo splash in background
I'm an AMD fanboy but I still find the poo jokes hilarious. I chuckled.

File: 1462557614921.png (2MB, 1600x1200px) Image search: [Google]

2MB, 1600x1200px

>>57982828
bottom kek

File: lVK3HSW.jpg (596KB, 1846x1923px) Image search: [Google]

596KB, 1846x1923px

The era of the x86 jew is finally coming to an end.

>>57982898
who makes these things they are hilarious

>>57982898

AMD has office in Israel and jews work there.

>>57982898
spotted the gaymur retard

File: 1234213121212.jpg (383KB, 1404x1086px) Image search: [Google]

383KB, 1404x1086px

Is the IPC higher than sandy bridge?

I want to know if the quadcores with SMT will beat my 5 year old 2600k at 4.4 ghz

>>57983076
>Is the IPC higher than sandy bridge?
The IPC will almost certainly fall somewhere around haswell or broadwell for general workloads, but we don't really have much concrete data at the moment, so take anything you hear with a pinch of salt.

>>57983076
The IPC of any intel cpu from the last 5 years isn't even higher than sandy bridge.

>>57983150
Sure it is. Not MUCH higher, but higher nevertheless. Even 5% per generation adds up.

File: 1421749647233.gif (2MB, 400x310px) Image search: [Google]

2MB, 400x310px

>>57983193
Except it isn't in a literal sense. People confuse (tech sites shill and leave this info out to get ahold of new test samples) improvements from SMT/Hyperthreading with increased performance. Sandy bridge Hyperthreading can sometimes even reduce the performance (especially when single threaded performance is relevant), that is not the case anymore with newer intels. As high end CPUs always have HT it makes it look like IPC increases.

Sure, it's faster, but only when you have HT enabled (which I don't because I cannot wrap my head around the load graphs with HT - but that's a different story. Just saying that it's relevant).

Jews being jews.

>>57983076
Stop asking this question and just look for yourself
http://www.anandtech.com/bench/product/1684?vs=288

If Carrizo had an L3 performance per clock would be marginally higher, but this is the best comparison you can get.
Excavator is 20-50%~ behind Sandy Bridge with Excavator having a minor clock advantage.

If Zen is an average 40% IPC uplift over this, and it is, then performance will easily be around Ivy Bridge - Haswell in a number of metrics. The areas where Excavator is furthest behind are super heavy FPU ops, which is expected.
Zen will very likely have higher multicore scaling as well, and anything even remotely memory bound like file archiving will see larger performance uplifts from the new high performance internally designed IMC.

Link is up. Still a couple of hours to go.

https://www.youtube.com/watch?v=G8R7frtLDBk

>>57983193
>http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/9

A 20% improvement over 5 years is not how things are supposed to progress in this industry. It only proves that intel has become too comfortable in their position as a monopoly in the x86 market and shows why Zen is so sorely needed to kickstart some real competition again.

>>57976683
He said Broadwell Cia nigger

>>57983339
Nothing is going to change on intel's end.
Corporate espionage is a real thing. Architectures take years to design, and every company knows what every other company is working on. Its just the way things are.
You know what intel is bringing to the desktop in 2018? Coffee Lake. Their 3rd generation of 14nm chips with barely any arch changes since Haswell. They aren't going to start releasing radically higher perf/clock designs. They don't even have the capacity to make them, because they're struggling for ever 1% they can muster. Pushing on perf/watt is about the only thing you can expect to see from them.

>>57983403
4th generation of 14nm chips*

Broadwell
Skylake
Kaby Lake
Coffee Lake

10nm Cannonlake will be low power and mobile only at first. It'll be 2019 before you see a 10nm desktop chip from intel.

File: 1402835022141.jpg (38KB, 413x395px) Image search: [Google]

38KB, 413x395px

>>57970916
>They've literally had these processors made, ready to ship like 10 years ago, but then they strip them down and release them piece by piece (think: dlc for games

>>57983403
Then AMD will simply overtake them as the market leader in mid-to-high end desktop cpu's. Why would anyone buy a $350 4-core cpu from Intel when you'll be able to get 8 cores with similar IPC from AMD for the same price? If intel wants to stay competitive, they'll have to either cut prices or innovate on performance. That's just a fact of capitalist economics.

>>57983299
http://www.anandtech.com/bench/product/288?vs=1646
Two CPUs, neither of which has hyperthreading and with practically identical clocks(although I think 6600 might boost a little higher). The 6600 wins in every benchmark, even if it isn't always by much.

Like I said, it's not BIG improvement, but it is improvement and it's not just jewtel getting better at SMT.

>>57983322
We will see

>>57983479
Intel doesn't get sales because of their performance or efficiency, though these definitely win them good PR. Intel gets sales because they're deeply entrenched with every OEM on the planet. Intel can even subsidize display costs for laptops which is one of the many reasons why they absolutely dominate in mobile.

Even if AMD was beating intel in perf/watt by 50% and ultimate performance intel would still have the vast majority of all laptops, all in ones, and pre built PCs.

>>57976528
>4w Zen mobile cpus

hot damn

>>57983485
What clockspeed does the RAM have? Are the multipliers correctly set? E.g. 1600 Mhz DDR3 and 3.2GHz (that's 3200 MHz on the CPU to get a 1:2 (1:4 real world) ratio? 3.3 GHz baseclock, what about turboboost? Niggernoobs can't even into that and need not apply. I'm not saying you are, but I'm implying it a little. :^)

There's a lot of things to consider, especially when you rely on a website that needs CPU manufacturers to survive.

Don't get me wrong, I wish it would be quicker. I don't care about brands, I want quick CPUs, no matter what's printed on it. But I really went back from newer intels to an unlocked sandy bridge because it just wasn't worth it.

Cut the crap
Where is the graphene?

>>57983636
Reddit, with the rest of the memes.
Get ready for high performance SOI FinFETs and GAA, faggot.

>>57983584
Go ahead and find a better comparison. You're right in that higher turboboost and other factors(like DDR4 RAM) might explain some of the difference, but I think you'll find that 5% higher clocks for 30% higher benchmark results still looks like higher IPC to me.

I'm NOT saying that the improvements are worthwhile, or that jewtel is doing a good job. I'm just pointing out that evidence points to the statement "Sure, it's faster, but only when you have HT enabled (which I don't because I cannot wrap my head around the load graphs with HT - but that's a different story. Just saying that it's relevant)." being factually wrong.

>>57971114
>le GHz maymay xDD
5/10

>join chat from stream
>filled with 3rd wordlers

someone is going to be upset when they see the price of this new zen cpu

>>57983672
but muh hybridated graphene trilayer transistors

File: heh.png (145KB, 382x277px) Image search: [Google]

145KB, 382x277px

>>57983703
Well, you know, it's actually impossible to know the real world load when HT is enabled. You have to guess more or less (2 cores at 100% (200%) can mean 2 instructions per cycle (theoretically speaking) or just 1, when HT is enabled. If it's not, it means 100% is 100% load. Period.) and also watch temps to figure out what the sweet spot of load is. Ain't nobody got time for that!

New CPUs from Intel are good when they have quicker iGPUs and lower power consumption, I fancy that. But the price tags aren't worth it and I just didn't find the proper notebook to replace my desktop yet.

Also I'd choose 2 TB of SSD over even an IPC increase of 300%.

Holy shit why is AMD so bad?

>>57970916
>intel releases a processor that's better than their last one
1% IPC

>>57971114
that would be hilarious, with how things got threaded all of a sudden in recent years, if tables will turn

year ago 4.5 fx cpus get beaten by 3.4 i5s
next year 4.5 i7 gets beaten by 3.4 zen

>>57983986
>IPC
Literally meaningless

>>57983849
>Also I'd choose 2 TB of SSD over even an IPC increase of 300%.

Wow I'm not sure I could possibly disagree with an opinion more.
You can literally go out and buy the former today, whereas the latter is simply beyond price and will realistically never happen on anything close to what we'd recognize today as a CPU.

>>57981825
athlon ii x3 here
we can be friends.

>>57984039
Big b8

>>57972466
Ryzin=risen(from the grave)

Every AMD product ever hyped on /g/ turned out to be shit, I really hope this one will be good

>>57970839
Is that the actual name AMD has picked? That's cringe-worthy.

Heatsink: Glued or soldered?

Its not over it is oven

>>57981825
Right here, looking into upgrading some shit since it's way too loud and the mobo I have only supports DDR2.

>>57971278
Kek'd

>>57983403
Mate - there is also a limited amout of magic you can do to increase IPC. There is a hard limit on what can be accomplished, and we are asymptotically getting there. At some point in the future there is not going to be much in terms of performance increase except for a radically different programming model or 'moar coars'.

>>57983339
Intel is the the one doing the innovation, a 5% increase in performance is a pretty decent improvent.

>>57983369
>broadwell and up
>"hurr durr he said broadwell retard :^)))"

>>57984522
Meanwhile, AMD is getting 40% increase from their last gen, IBM's Power9 is supposed to be at least 60% better than Power8, and ARM CPUs are also progressing pretty damn fast compared to Intel.

File: 1424715856582.jpg (32KB, 312x342px) Image search: [Google]

32KB, 312x342px

>>57984522

>>57975678
In temple OS you literally only have to change one line per loop to parallelise your programs.

>>57975328
People that care: GNU/libreboot tinfoilhats and 100-150 other autists.

I don't want to buy the Hype, but I'm really looking forward to see the desktop processor market being competitive again.

https://www.youtube.com/watch?v=4DEfj2MRLtA

>>57975429
>>57975589
can't find it
searched for
>intel
>disable
nothing there

>>57984622
>40% increase from their last gen

That was because their last gen was complete and utter shit

>>57981699
Windows 7 is dead, son

>>57984622
>40% higher IPC
>after 5 years
Intel has accomplished the same thing, only they've actually released products in the mean time

>>57985018
40% compared to excavator, not piledriver.

>>57985055
>>57985018
Also, no they haven't. We've literally been through that in this thread, skylake is AT BEST 30% better than sandy bridge.

If this can match Ivy/Sandy, then all those cheap E5-2670's are useless now. Rather unfortunate that it shpuld use DDR4 though - that means I can't buy a bunch of those AliExpress DDR3 RDIMMs on the cheap.

When where you when Intel was kill?

GOD DAMN LOOK AT THEM TITTIES

File: 1478904199921.jpg (48KB, 550x700px) Image search: [Google]

48KB, 550x700px

>>57970839
>LOOK AT OUR BENCHMARKS
>THIS IS BLENDER, BLAH BLAH IT SCALES WITH MANY THREADS
>amd rice *almost* catches up to similarly clocked, four-core-less i7

well
i somehow don't feel the hype

>>57982272
I think you meant to say mega hurts

>>57986818

The 6900K is 8c/16t.

>>57986818
The fuck are you smoking? It was an octocore broadwell vs octocore zen, and the zen was actually slightly faster.

>>57986818
>getting paid to post lies and animu
where do I get this job?

>>57984523
Are you some kind of knuckle dragging MIT nigger?

File: 1475544574239.gif (350KB, 500x281px) Image search: [Google]

350KB, 500x281px

>>57986935
whoops

>>57986938
frankly, i'm volunteering

>>57987021
Shut the fuck up you retarded Snakeshit and get the fuck off this board you cancerous faggot.

>>57971304

5960X was 3GHz, 6900K is 3.2 and the 6950X is 3GHz

I really hope AMD releases something great but even it doesn't live up to the hype I will still probably go with them because I don't want to reward intel for the shady shit they pulled to get this monopoly position.

>>57982332
For the 8k meme ready

>>57971884
Should've watched AMD's stream or read an article.
Were you spoonfed all your life you cuck?