WHO HYPED FOR ZEN!?!?!
>amd cpu
>>54235879
>supporting jewtel
kek
>>54235672
I'll buy one just because it's Keller's brainchild
>>54235672
I'm excited, I personally couldn't wait for Zen, I already have bought jewtel and have no real regrets, Zen is just taking longer than i'd hoped. But, I am hopfeul Zen can breathe some new life back into AMD.
Performance, if the rumors are right is solid and should at least make it somewhat competitive again.
I just want something to upgrade.
6 years with no real performance boost is retarded.
CPU market is so fucked.
>>54236292
>Performance, if the rumors are right is solid and should at least make it somewhat competitive again.
Honestly if they reach haswell IPC and 4GHz clock and if they price it right AMD is gonna do really well with an 8 core zen.
Considering im still satisfied with my FX 8320, i may upgrade to Zen next year, maybe prices will have dropped a but by then. AMD is definitely going to milk the hype, probably offering just barely competitive pricing to intels offerings for the first 6 months at least.
those 512kb l2 cache are shared between code and data or are just data? x-way?
i kinda wish that was split into larger chunks shared between modules and get rid of the l3 one
>2017
>Zen-based working machine
>Zen-based HTPC
>Zen-based server
Life will be good.
>>54236920
Why? That's more L2 than a 5820k
5820k has 256KB per core of L2.
>>54235672
I'm excited for AMD to finally have something competitive again
Got 2-3 builds planned in the next few months that I'd use it in if it's good enough.
>>54236935
to me it would be worth treading in that size for one less level, less cache coherence stuff needed, less die space needed
still, with ddr4 larger cl, having a larger cache and more levels might make a bit more sense, but it makes the chip way more complex too
>that bogus intel crash with the avx instructions from last october
>>54235672
My 5 year old sandy bridge will shit on that.
>>54235672
I just hope Zen gives me Haswell performance for half the cost so I can finally replace this 2009 Phenom II.
>AMDneets wait for their poverty chips that will get dominated by Intel chips that are years old
Loving every laugh.
>>54237177
Intel is not that great either, still using Sandy here.
Intel a shit.
>>54237177
have fun financing intel's 80%-margin.
>>54235672
>32KB L1 per core
Seems normal
>512KB L2 per core
Okay I guess
>8MB L3
why.jpg
I'm not expecting that much disappointment because of Keller, but yes, I'm hyped. I'd love an 8 core for less than $500. If it can go faster than my Ivy i5 at 4.3GHz, then I'm all for it. I'll build a server around it too, you can basically have my money, AMD.
>>54237335
>Okay I guess
double an i7-5820k one of intels top chips, is only okay? Jesus christ, AMD would have to suck your dick dry for a month just to get you to say something nice about their hair.
>>8MB L3
>why.jpg
well ignoring the fact it's an engineering sample and could have some of the cache disabled (manufacturing defect or similar), but there is nothing wrong with 8MB L3 cache, that's what's on the I7-6700k. 8MB L3 cache. Also the i5-6600k only has 6MB L3 cache, how terrible. Yet it still outperforms every current AMD CPU.
Going off just cache sizes is retarded and wont tell you anything about final performance.
>>54237393
only thing I'm worried about is TDP.
I hope it won't be the usual AMD-housefire-starter.
Is Zen going to be Duke Nukem Forever?
It seems like AMD has been hyping this thing for ages.
>>54237400
Reason I am saying "wah wah" about the L3 cache part is that AMD doesn't utilize cache properly, this is why I am worried, their current CPUs have to have a high amount just to work "good".
>>54237335
>8MB L3
>why.jpg
well it is an ES. Some could be disabled.
>>54237422
Well considering the extra L2 cache, i'd say they are obviously using a different Cache scheme than previous CPUs, until we see the cache perform, it's useless speculating about how it will perform based on the SIZE of the cache.
>>54237430
>>54237440
I hope so, currently AMDs CPU tend to depend heavily on the L3 cache. So if they have improved the cycle cost thus eliminating the need to have a bigger L3, then I might be a bit happier but in any case I will still be worried.
>>54237403
It can't be worse than the disaster that is Bulldozer.
>>54237457
It's AMD, you should be worried until the day it is released, and then you should still be worried in case there is a latent manufacturing defect or similar.
>tfw never fell for the upgrade meme
>tfw still on C2D
>>54237470
True, which is why I have stated my concerns about it. Currently any of the i-series from Intel are amazing, but price-wise they're not anything good for the market.
>>54237514
enjoy your bloated power bill
>>54237514
Ugh, I upgraded from a C2Q Q6600, that piece of shit was great in it's time, lasted far longer than it had any rights to. But when I finally did upgrade this year, holy shit was there a HUGE difference. You dont realize it until after the fact, but good god was that thing slow.
>>54237526
Eh, I think people honestly are forgetting how much intel used to charge, prices these days aren't even that bad, i remember back when $400 motherboards were standard with $400 CPU being the low end.
>>54237552
Well of course they're not "too bad" but they're not anything market changing prices. AMD needs to step it up so we can have a proper competition and advancement from both sides.
>>54235672
Not me, I'm not a poorfag so I don't have to get excited for AMD's outdated on release tech.
>>54237572
agreed, with some real competition we could see the 5820k drop below $300 and it could force intel to start pushing 6 and 8 core CPUs into the consumer series i5 and i7s.
>>54237586
Post your 5960x then. Better yet, post your high end Xeons, please.
>>54237613
>5960x
he said he wasn't a poorfag, not that he likes throwing money away.
Anyone who isn't just tossing money down the drain and gets intel will probably have an i5-6600k if it's just for gaming, an i7-6700k if they have a bit more money and want to future proof.
Or a 5820k/5930k if they are doing encoding or rendering or similar.
Xeons for anyone who needs ECC and MOAR cores.
pic related, my 5820k for encoding.
>>54237586
>I'm not a poorfag
My setup probably costs about 10x as much as yours and I'm still excited.
It doesn't take a poorfag to understand the value of market competition, which alone is enough reason for anybody to be excited.
On top of that, I'm also growing increasingly wary of Intel's stance on e.g. privacy (hardware backdoors), ownership rights (remote kill switches) and bad business practices (overcharging to unlock existing features)
>tfw remembering all the bulldozer hype threads on /g/ and how it was going to btfo sandy bridge
I'm going to laugh so hard when AMD fucks it up again.
>>54237639
>5820k
Neat. I personally think the consumer i7 is retarded since the i5 performs just as well on games. Justifying the the price point of a consumer i7 nowadays doesn't make sense if you really need more threads/cores.
>>54237699
The threads were just heartbreaking. I was going to build a new system with Sandy in mind, but decided to wait for Dozer since the Phenoms were alright. What a shitfest it was. Keller wouldn't put his name on something bad though. He has done pretty great things.
>>54237761
Oh no, you'd be an idiot to get a 5820k JUST for gaming. Or x99 platform in general.
X99 is for workstation type workloads, rendering, encoding, etc.
But that being said, if you're debating between 6700k and 5820k, the 5820k is usually a better value, only $30 more for 2 more cores, twice the L3 cache, and it's clocked slightly lower, but can be OC'd to over 4Ghz with ease.
>>54237831
>you'd be an idiot to get a 5820k JUST for gaming
Hence the mention of the i5.
>>54237847
yes but x99 doesnt support any i5 CPUs which is why i mentioned staying away from x99 in general for gaming.
>>54235672
>mfw i think of that glorious MAKEOPTS="-j9" that i will set when installing gentoo on my new FLAMING ZEN RIG
![[Jingrock] Love Letter (Complete) [English].zip-064_LoveLetter_062.jpg](https://i.imgur.com/48iqnP1.jpg "[Jingrock] Love Letter (Complete) [English].zip-064_LoveLetter_062.jpg")
>>54238177
>mfw they stick eight 4 core modules on one chip
>mfw SMT
>mfw make -j64
Let it be good let it not suck please microchip god
>>54237462
At least that.
>>54237462
I never got into Bulldozer, how bad it was compare to Pentium prescott fiasco?
>>54238177
>>54238540
>AMD getting people excited for multithreaded opertions
>AMD multithreaded performance is historically something to be excited about
Oooh, I can wait to see you all create threads about it.
>>54238979
it wasn't as bad, but intel had a shitton of money to fix the problem, AMD doesn't have that.
>octalcore
>>54240174
AIDA64 refers to 8 cores as octalcore
