6C/12T Ryzen confirmed having 3.3ghz stock - Bad overclocking

File: 1483628859502.jpg (156KB, 633x758px) Image search: [Google]

156KB, 633x758px

>>58751981
DELETE

>>58751981
hard to believe considering i managed to undervolt my rx480 and OC (over 200mhz on the mem, over 100mhz on the core)
>still thinking ghz speed matters outside of the same line of cpus
go home shill.

File: 1482866251494 (1).jpg (51KB, 600x515px) Image search: [Google]

51KB, 600x515px

>>58751981
>canard
Fucking leafs

File: AYYMD-on-suicide-watch.jpg (434KB, 1920x1080px) Image search: [Google]

434KB, 1920x1080px

>>58752018
>undervolt my rx480 and OC (over 200mhz on the mem, over 100mhz on the core)

>amd fags literally downgrade their gpus
>ignore that's info from one of most reputable pc magazine who got zen early
>is so delusional that think 3.3ghz zen will defeat 5ghz intel lol
>it really makes me think

AYYYMD fanboys literally on suicide watch

File: 1475804622608.jpg (80KB, 1000x769px) Image search: [Google]

80KB, 1000x769px

>>58752052
You guys are worse than CTR
go back to /v/ with your company wars

File: 1483974609716.jpg (540KB, 1100x1002px) Image search: [Google]

540KB, 1100x1002px

>>58752078
>being this MAD

>>58752052
undervolted and overclocked retaaaaard

>>58751981
DELET

>>58752052
>dat jenhsun face in the background
kek

>>58751981
>Rumor = Confirmed
We /v/ now.>>58751981

>>58751981
But what about the price?
t. poorfag

File: 1483628705179.jpg (954KB, 1432x1707px) Image search: [Google]

954KB, 1432x1707px

>>58752283
it's not rumor you retard it's confirmed fact

>>58751981
I dont speak leaf, what does its stock speed effect its overclocking potential?

>3.3Ghz base clock
>bad
Try harder, Intelshills

>>58752459
$60 Pentium is better clocked

File: 1483193560378.jpg (66KB, 568x612px) Image search: [Google]

66KB, 568x612px

>>58752459
>i5 7600k - 3.80 GHz STOCK
>i7 7700k - 4.20 GHz STOCK
>i7-6700K - 4.00 GHz STOCK
>I5-6600k - 3.50 GHz STOCK

>3.3ghz base clock for 8c/16ht
>NOT BEING FUCKING BAD

you amd fags are so delusional

>>58751981
they are validating 6 cores.
what's the problem?

>>58751981
So first they confirm overclocking then they go back on that claim?

https://www.techpowerup.com/229090/amds-upcoming-ryzen-chips-to-reportedly-overclock-5-ghz-on-air

jesus christ how desperate to have to be for (You)s to samefag this hard?

>>58752633
No. They're basically saying that there are chips which are stock clocked at 3.3 GHz with 6C/12T, and they're assuming that there's no 8C/8T CPUs, since the binned CPUs will be more probably used for 6C chips that for 8C chips without SMT.

>>58752689
they're not assuming, they're claiming

File: 1485199838756.jpg (49KB, 500x500px) Image search: [Google]

49KB, 500x500px

>>58752556
>comparing different dies clock for clock

is it 2001?

>>58751981
There were rumors there wasn't going to be a hexacore in the first place.

Step up your trolling.

>>58752714
They're claiming that no 8C/8T are being validated, you're right.

>>58752689
>>58752714
All rumors....

http://www.forbes.com/sites/antonyleather/2017/01/29/no-6-core-amd-ryzen-cpus-4-core-and-8-core-only/#40cec6077f9c

File: 1483413065316.jpg (3MB, 8624x3896px) Image search: [Google]

3MB, 8624x3896px

>>58752647
>>58752721
AMD claim they have same TDP as Intel now so it's obvious that you can compare 3.3ghz of zen to intels i5/i7 3.5-4.2ghz

>>58752721

No because that was back when AMD was competitive.

>>58752744
i meant IPC

>>58752721
That's what I've been thinking too. Seems like people care a lot about clock rates again.

i'll be waiting for benchmarks.

>>58752744
Having the same TDP does not mean they're the same architecture.

>>58751981
why 12 thread thought

Literally what is the problem if this is confirmed? It's still a 6C/12T at 3.3 fucking GHz for CHEAP

>>58752736
Since he benchmarked an engineering sample of a 8C/16T, it's not much of a stretch to assume he's seeing some 6C/12T going around, and no 8C/8T at all.

That said yeah, a twitter quote isn't much of a source, as much as I like Canard PC Hardware.

File: AMD-logo.jpg (13KB, 336x352px) Image search: [Google]

13KB, 336x352px

>>58752744
>AMDREDTEAM

>>58752556
Parts with more cores tend to be lower clocked for power reasons. That's why those cheap E5 2670 are just 2.7 GHz.

>>58752556
>>58752744
>comparing quadcores to 6 and 8 cores
Is Intel this desperate?

The 5960X is 3.0GHz. Am I missing something here or is it blatant shilling?

>>58753040
it's a twelve year old from /v/ trying to troll

>>58753040
>POOZen barely clocked higher than a years old midrange chip from based Intel
AYYYYMDlmao

>>58752052
>is so delusional that think 3.3ghz zen will defeat 5ghz intel lol

No-one thinks that. The 5GHz Intel is going to cost 4-5 times as much and nobody will bother except professionals that need that kind of power.

>>58751981
>bad overclocking potential

aka optimised hardware

File: 1483560094820.png (93KB, 717x699px) Image search: [Google]

93KB, 717x699px

>>58753068
can't pic related be implemented?

3.4 ghz was confirmed to be the minimum, this news is fake

>>58751981
>Bad overclocking potential confirmed just like rx 480
Where do you get this claim from?

>>58753217
in tel aviv

>>58752556
>5280k 6c/12t 3.3ghz

>>58753170
10% is barely higher? Enjoy your 1% gains faggot.

>>58753213
3.4+ was meant for the 8 core.
I see no reason why the 6 core should clock lower, other than a budget SKU.

>just like rx 480
My 480's base clock is at 1.1GHz.
I run it overclocked on 1.45GHz @ 1.18v

File: amd1.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>58751981
NOOOOO

>>58753295
Clock binning and defect binning go hand in hand. These chips would also likely be 65w SKUs rather than 95w.
Su's statement about clock speed was that at launch Ryzen CPUs would have clocks of 3.4ghz or higher. The launch window is the qualifier there.
Lower clocked chips being available later doesn't conflict with that.

File: 9fa.jpg (13KB, 275x183px) Image search: [Google]

13KB, 275x183px

>>58753295
Yeah AMD downvolts really well for some reason. I remember some cheapo laptop I got in 2009 that had an absurd stock voltage, 1.425v, wasn't even clocked that high. Took it down to 1.050v and it ran fine.

File: 2017-02-01_20-55-26.png (19KB, 539x355px) Image search: [Google]

19KB, 539x355px

>>58751981
>Intel released a desktop CPU for $1800 with only 3GHz
AHAHAHAHAHAAHAHAHAAHAHHAH

>>58751981
>5820k 3.3 Ghz
>6800k 3.4 Ghz
>6850k 3.6 Ghz

Am I missing something? Also

>implying clocks matter when comparing different architectures

>>58752550
>a chip with fewer cores has a higher base clock

OMG the g4560 has a higher base clock than the 6950x 1!1!"1!!1!1!1"

>>58753385
This is why I'm interested in the new line of server chips. I've had nothing but trouble with voltage scaling on Intel. The lower you can go the more you can save (which matters when it's running 24/7).

>>58752800
they're red since they acquired ATi

>>58753840
That never happened.

>>58753887
exactly, they killed ATI

ATI made some of the greatest GPUs until AMD bought them out

>>58751981
Time to bankrupt.

>>58753840

AMD's cpu branding is still often green. Radeon branding is mostly red.

>>58751981
nothing mentioned about overclocking

>>58752347

Because this?

http://www.guru3d.com/news_story/amd_might_not_release_a_hexacore_ryzen.html

Anything more than 4 core will be parts that are going to be considered high end as they all will not have a gpu on chip. 4 core will be considered low tier with apu occupying the mid tier.

I fucking hate you all. Making meme wars out of anything. I will buy ZEN. You can keep your skylake refresh retards.

when do you guys think they'll stop manufacturing skylake? i'm sitting on parts right now and was waiting for ryzen to hopefully cause intel to adjust prices.

>>58751981
please amd for the love of satan dont fuck this up

File: Screenshot from 2016-06-08 18-24-48.png (308KB, 557x792px) Image search: [Google]

308KB, 557x792px

If you want cheap chips then you'll have to settle for the low clocked SKUs, this is inevitable, you can't expect the Zen F4 stepping 4.0 boost 8 cores to cost the same as the 3.7 boost ones.
That's why overclocking exists, but don't think for a second you'll get good silicon by buying the cheaper ones, they're also more likely to overclock less unless you're real lucky.

>>58754867
Intel doesn't cut prices for shit, faglord. Go price a 4790k.

>>58755518
thanks for help. now tell me when they stop manufacturing skylake fagmax.

>>58753178
8/8 or 8/16 will beat out an i7 at 5ghz on parallel tasks.

>>58751981
no idea what this is in context google spits out

"AMD currently validates Ryzen 6C / 12T @ 3.3 GHz base but NOT 8C / 8T. The goal is clearly to recycle the failed 8C dies."

This makes it sound like they sent out a 6/12 for validation and it was 3.3ghz, amd has said the lowest clocked ryzen will be 3.4 so i'm also guessing engineering sample.

Also does this imply they do NOT have 8/8 cpus?

>>58752786
6 cores, 2 threads a core, 12 threads.

I might have to leave /g/ due to embarrassment after shilling this so hard. :(

>>58753902
and then they made some of the best gpus they ever put out

>>58753295
i'm taking mother sue at her word, no desktop ryzen will be base below 3.4

>>58756177
"at launch"

>>58754767
if there is a 4 core at launch it is a heavily fucked 8 core with no gpu, they won't have apus till h2

>>58756193
4 core doesn't sound feasible unless yields are real bad, you kinda need real shit yields to have either 2 cores per CCX not working or a full CCX not working on such small dies on a relatively mature process.
Like, 1.7% bad

>>58756227
Well, they did do that for Phenom, Phenom II, and the Construction cores. Pretty sure on release all of those architectures had SKUs utilizing the main CPU die but with the bad cores turned off, including tri and dual cores.

>>58751981
I don't sprache frog, what's he saying?

File: bigguy.png (112KB, 250x250px) Image search: [Google]

112KB, 250x250px

>>58751981
>stock up 15% today

yeah, really disappointed here

>>58756251
Not the same thing, Zen isn't designed like K* and Intel Core with a monolithic core complex
So we actually don't know how these CCX work, do they need to be symmetrical to even function in a 6C config (3+3 per CCX), or can they just work with 2 cores disabled in a single CCX.

Bulldozer's design made it simple, but Zen's design makes it not so simple at the moment.

>>58756297
True true, but at the same time AMD would be insane to not have the ability to go hex-core or otherwise turn off single cores with this, because the hex-cores and the quads if they're any decent will be the real moneymakers.

>>58756323
And we come full circle again, quad cores aren't gonna be in this launch window and they won't be using summit ridge dies, if they do there's something REALLY wrong with GloFo, they'll be from raven ridge, which I assume is a single 4 core CCX with a vega GPU if there's any sanity left in AMD

Six cores were all but confirmed, we just don't know if they'll be 3+3 or 4+2 cores
Disabling a single core is pointless, there's a reason nobody ever sells 3 cores or 5 cores.

>>58756426
AMD might create a 4 core Summit ridge by fusing off a ccx on purpose just to put something in that market spot.
Which sounds kinda silly and not very profitable to me, this would just screw over their stock of 6 core SKUs.

>>58752052
>Undervolt card
>frequency goes up

Try harder retard.

>>58756551
While you're responding to an idiot it is in fact quite possible for the frequency to go up by undervolting if the card is thermally or TDP throttled, in most cases the latter.

>>58756426
Selling a salvageable die is better than throwing it away.
SKUs don't have to be super high volume. Viable 4 core chips that pass clock binning may be less than 1% of working chips, but that still works out to be a shitload of potential profit. Fab 8 can put out 60k 14nm wafers per month.
If the Zeppelin die is 200mm2 a 300mm 14nm wafer has about 280 candidates per wafer. If yields are 40% then we have 112 good dies per wafer.

If only 1 of them per wafer turns out to be a working 4 core chip they can make 60,000 quad cores per month.

Though this is assuming that Zen based parts are using 100% of Fab 8 14nm capacity which isn't the case, but its serves to demonstrate the point. Even if they were getting 2,500 per month, a $100 chip is potentially a quarter million dollars in revenue.

>>58756591
A quad core summit ridge needs to fail spectacularly binning for both 8 and 6 core to exist, which is really kinda pushing it.
Here's what it needs to fail.

8 core:
fail voltage target, fail clock target, fail SMT(?), fail power target.
If one of these fails and there's a bunch of these failed chips, you have another 8 core SKU, if more than one of these parameters fail, you got your 6 core.
Now that six core has to fail a bunch of targets targets to get binned to a 4 core.

This is a really small chance, especially since these dies aren't really big.
1% chance? I doubt it's even that big, there's more of a chance AMD artificially creates those 4 core SKUs if they even exist.

>>58752283
>Validates
>Rumor

>>58756669
Voltage and power are effectively the same thing. Its clock binning, hitting a target clock within a certain voltage envelope, and that determines how much power it uses, and how hot it runs. Very few chips are thrown away for failing to hit nominal clocks with acceptable voltage. If a working die cant hit 4ghz at 1.2v for example, it might be held in inventory. When inventory builds up they'll revalidate the chips as either higher power draw high clocking SKUs, or lower clocked low power SKUs.
Either way they're not simply cast aside. These low end parts are typically about recouping losses more so than making a profit. Low end chips soften the blow of poor yields on high end parts.

6 core chips will be dies with either 1 or 2 defective cores.
4 core chips will be dies with either 2, 3, or 4 defective cores.

>>58756227
We don't know how the architecture can be disabled yet, we may not be able to disable single cores, only groups of two, if we could disable single cores, where are the rumors of 7 or 5 core skus?

really, a fucked core could be one spot on one core that takes out a second, and another on another core that takes out another with it, 2 fuck ups on a single die, not exactly statically uncommon even on high yield processes.

from a calculator and figures from samsung on yield they are at the point where 75%+ of the dies is perfect if not more.

>>58756669
we would have rumors of 5 or 7 core units if they were disableable as one core, but we dont, there would be more then 3 skus at launch, but there aren't.

This leads us to believe something interconnects between cores and if one is fucked with fucks 2 cores.

this means 1 defect = guaranteed 6 core unit
2 may give you a quad core unit. and i'm guessing units defective enough to be dual cores just don't happen on this mature of a process.

>>58756551
higher voltage means more chips are viable as many were not low voltage contenders, those were binned away for embedded skus.

you undervolt, you now allow for a higher clock at a lower tdp or potentially you were thermal throttled and the lower voltage allows you to go higher that way.

many things can be undervolted and still run perfectly fine.

>>58757591
Intel never felt the need to release odd numbered core chips like that. Their top server SKU has 24.

Well except that weird 15 core one, but that was the entire die actually.

http://ark.intel.com/products/75242

Seriously what the fuck were they thinking while designing that die.

File: Intel-Xeon-E7-v2-Die.jpg (2MB, 2724x2290px) Image search: [Google]

2MB, 2724x2290px

>>58757711
Honestly what in the fuck.

>>58757711
they took the 10c die and slapped another row of 5 cores onto it, extended the memory controller for some proprietary bullshit that I dont think ever panned out, then sold it.

>It has bad overclocking potential because of low stock speeds

Yeah, these 2500K's sure never even pushed 4GHz as well.

>>58754867
On about 4 (3.5) years.

>>58757735
I've always wanted to get my hands on one of these, but not for $1200.

>>58751981
>Bad overclocking potential confirmed just like rx 480
Bullshit, the top model has 3,65 GHz base/4,0 GHz boost. See >>58758645

So you'll get at least 4 GHz out of that sixcore, more with nasty voltage.

>>58752106
>undervolted
>overclocked.
pick one.

>>58758750
or you can pick both

>>58758750
Nope.
Stock voltages on Ellesmere are so generous that most RX 480s can sustain peak turbos for longer by lowering the voltage in the top pstates. The top pstate is thermally limited.
Most people can hit 1300mhz slightly below 1.1v.

File: overclocked stats.jpg (737KB, 3840x2160px) Image search: [Google]

737KB, 3840x2160px

>>58753295
>1.45GHz @ 1.18v
Neat. Which brand?

>>58756520
if polaris is any indication there's going to be a lot of 'serviceable' bins with overall inferior power characteristics. fusing off the shittiest cores and selling the four core skus with kaby-lake rivaling clocks would be a good strategy imo, and you preserve the market segmentation for the 6 and 8 cores with the best binning.

We have literally no confirmation on anything except 8/16T with at least 3.4GHz. Stop this inane shitposting.

>>58758853
The Polaris dies were the first 14nm production parts ran at GloFo. They were a special case because AMD was targeting a low price point for their cards, they didn't want to run into the limited supply issues that Nvidia had with their Pascal GPUs on launch, and they doubly had to deliver ample volume to Apple. In order to have plenty of dies available for consumer cards, so supply wouldn't be limited, they gave a really broad window for tolerable required voltage. That of course meant TDP was slightly higher than necessary in call cases.

That doesn't reflect on other products in any way.
If that was the case then 95w 3.6ghz/4ghz Ryzen CPUs would have plenty of chips being able to undervolt and pull way less than 95w.

>>58757711
not talking about intel, with ryzen someone speculated one of the caches is shared between two cores rather than separate, or two are linked in some way, it makes the cpu faster, but when shit goes wrong with one the other is unuseable as a result.

It sounds retarded to me, but a possibility, because if amd could fill out a lineup with a 4/8 5/10 6/12 7/14 and 8/16 core they would have every price point they want covered without resorting to ghz binning outside of the 8 core.

>>58758703
that 6 core is an es, and potentially fucked like the 3.15 8 core was.

really wish amd would talk, would be nice to know whats going on.

>>58758925
>That doesn't reflect on other products in any way.

I think it reflects very well on AMD's overall business strategy and it's absolutely no secret that the entire point of Ryzen is essentially to undercut Intel's modern offerings rather than smash them out of the ballpark in raw performance, although I expect they will perform better than broadwell at the very least. The 14nm process has matured a lot but there's still going to be good bins and bad bins. Will every ryzen chip without serious defects be able to hit 4ghz boost on 8 cores in a 95w envelope AND save room for overclocking? I seriously doubt it.

>>58759019
I mean the choices made to shit out boat loads of RX 480s in the shortest time span possible doesn't mean that Ryzen CPUs would be treated in the same way. CPUs have a lot more steps of clock binning, there are higher margins here, and as you mentioned the process is much more mature now.

Samsung's 14nm node was online before the start of 2015. Granted it was LPE at firs, but all that dramatically changed was the light source. LPP has a more refined higher output light source, better definition leads to more consistent structures, and etc. Samsung and GloFo's Fab 8 has been running test wafers between facilities to ensure they maintain consistent yields. A couple solid years of refinement, Polaris parts serving as the pipe cleaner for Fab 8, Ryzen yields are going to be pretty high.
I'd bet on a refresh later in the year hitting closer to 4ghz base clock.

>>58759139
I guess we'll find out in a few weeks. I'm very confident there will be at least a couple four core skus at launch, if only to cockblock intel's four core gaymer segment and ensure total market domination by the middle of the year.

>I'd bet on a refresh later in the year hitting closer to 4ghz base clock.

Trading more speculation obviously but I think the safer money's on the next zen iteration being the full 7nm die shrink.

>>58758823
xfx gtr or someone got lucky on a chip slipping through the binning process.

>>58759380
Why? at least consumer side there is a good chance that amd puts the 8 core at i7 price, and if they dont they have the 6 core at that point. and it skews down depending on if amd goes 8 core i7 or not. so i7 - 8/16 the i5 - 6/12 and the i3 - 4/8 with a binned sku going for 500~$
remember, the entirety of the consumer lineup is failed to be enterprise chips.

>>58751981
There won't even be a 6 core amd cpu due to how the architecture is done, it will instead be replaced by an 8 core cpu without multithreaded capabilities.

>>58759888
There have been 6 core engineering samples known for weeks. Stop regurgitating clickbait headlines.
Cores within the CCX can be disabled. So long as the two CCXx remain symmetrical so the L3 buffer isn't interrupted there is no issue.

>>58751981
Obviously fake, all Ryzen chips have a 3.4ghz base clock

>>58752744
RIP little fat kid, had his GPU stolen by his brother

>>58759752
>>58758823
Bullzoid has a 1580mhz XFX GTR chip

modified BIOs under water cooling

https://youtu.be/6Lw5Yryb168?t=6944

>>58759888
Of course there's going to be a 6 core part, the FX 6300/6350 was probably the most popular chip on the platform

File: imconfuse.jpg (24KB, 278x299px) Image search: [Google]

24KB, 278x299px

>>58760065
Holy fuck, isn't that approaching Fury X levels?
>mfw the falseflagging from back then is slowly becoming real

>>58760399
the false flagging was real once he hit 1500, the first non retard images showed a binned chip and where its performance was, what you are seeing is a binned chip, and then some.

>>58760399
It would be dangerously close, but lacking in memory bandwidth. Even at stock clocks the card is slightly memory starved, ComputerBase.de showed it in their review.
If you could feed it effectively then you have a near 25% uplift in performance, and that performance scales almost linearly baring being particularly pixel throughput limited.

>>58760399
>>58760411
>>58760423
Naw the RX 480 needs to hit 1700mhz to hit Fury X/1070 levels

But once the process matures and AMD has an average base clock of 1500mhz, Nvidia is fucked, they're hitting a hard 2100mhz barrier

https://www.youtube.com/watch?v=HYg123osc04

>>58760029
It's a 3.3ghz 6c/12t engineering sample it is talking about. The idiots that try to wrap it up to mean anything other than that however are obvious fake bullshit and clickbait

>>58760423
Polaris 10 1500mhz base clock 8GB GDDR5X RX 490 fucking when?

>>58758750
actually my 2600K runs 4.5GHz at a lower voltage than the it uses with stock/automatic settings.

I also downvolt it all the way to 0.9V when it's downclocked to 1.6GHz at idle.

>Dat feel when I already ordered 6700k

Still I somehow just can't believe that 6c/12t will be magically cheaper than 4c/8t of Intel.

File: 1485083307083.png (4MB, 1543x1615px) Image search: [Google]

4MB, 1543x1615px

>>58751981
son of a. 6 core should have HIGHER base clock than 8 core. WTF AMD get you shit together

>>58751981
lol even adoredtv said these guys were extremely reliable. i guess this means ryzen really is a flop.

>>58760620
That smaller dice surely hurts Pascal. Still following that retarded logic 980 Ti oced to 2Ghz+ is a better card than 1080. Still even then you are wrong cuz 1080s were hitting 2500Mhz on 1.35V.

>>58761778
So AMD will end up killing nVidia but Intel will counter it, forcing AMD to be cut in half, only for it to come back later in some animated series with it's upper half intact but it's lower half nothing but a robotic shell of it's former self?

Huh....

>>58761765
the die size is what determines the cost, not number of cores

>>58761869
DELET THIS NOW. That's IT I'M GOING INTEL FUCK YOU AMD!!!

>>58761765
Summit Ridge has no iGPU. That alone makes it half the price of Intel chips.

>>58756081
The trouble with 8/8 is that the performance lands pretty much at the same spot as 6/12 assuming around 30% performance from smt. Leaves them with marketing problems I think, I hope there's nothing without smt, i5s are in a shit position these days because of it.

File: IMG_8878.jpg (85KB, 690x690px) Image search: [Google]

85KB, 690x690px

All you Intel faggots with your gigahertz. Intel is an SJW company, and you're all SJWs.

Let's wait until the fucking products are out and on the shelf, and have been benchmarked with a proper benchmark suite.

>>58764300
What the fuck is a SJW?

>>58764300
>Intel is an SJW company, and you're all SJWs.


>>Implying AMD isn't SJW Company

https://www.youtube.com/watch?v=qyD_Q8RZh_M

>>58764703
DELET

File: gender tranny surgery giveaway twitch.jpg (274KB, 1179x757px) Image search: [Google]

274KB, 1179x757px

>>58764742
>t a

File: YO DAWG.png (196KB, 321x236px) Image search: [Google]

196KB, 321x236px

>>58764703
>caring about that, in both cases

>>58764703
>The company that literally gives money to Feminist Frequency is less SJW than a company that made a single video about gays.
End yourself.

File: 389e18fc3ea09d8bf87bf26d6910_w827_h1255_g4ab3df36b7e211e6af6e002590604f2e.jpg (366KB, 827x1255px) Image search: [Google]

366KB, 827x1255px

>>58765414
feminism is good
you are a faggot
have a nice day

>>58765442
>Marxist cultural division with ideology not even remotely based in reality
>good

Nah.
All vestiges of Communism need to be violently ripped from western society.

>>58765473
feminism isn't maxist. That's like saying it's anarchistic or punk because some people decided to follow both ideas at once.

>>58765509
>feminism isn't maxist.
This is just gross ignorance. All of gender studies is Marxist.

>>58765523
bullshit, what silliness the people studying that choose to subscribe to is orthogonal.

It's likely to be something more or less left, but that is mainly because rightwings are hostile to it from the start so anybody with those interests in driven away from them.

File: trans-solidarity-461x405.jpg (60KB, 461x405px) Image search: [Google]

60KB, 461x405px

>>58765509
"Feminism" is a generic umbrella term for several different kinds of activism based around women's interests. It includes everything from terrorism in the early 1900s to racists demanding that women be given the vote instead of negroes and modern day slutwalks. Literally anything that involves women protesting for any given reason at all, something that's been around as long as societies have.

Modern academy feminism and social justice on the other hand is almost entirely descended from and inspired by Marxist schools of thought and actual Marxist academics. Everything is there, a group of people who dindu nuffin who are all the same, they're all being oppressed by the bourgeoisie who have to be overthrown because they deserve what's coming to them, etc. I mean you have large numbers of actual, unironic communists overlapping with the same groups and protests and people chanting the fucking Communist Manifesto. It's obviously not a KGB conspiracy or something ridiculous, but you can't deny that there are a LOT of similarities and overlaps.

>>58751981
if you think clock speed directly translates to performance you need to leave this board

>>58752744
> R9 380 for a 1440p monitor

>>58765848
Maybe in the west where marxism never lost teh armchair ideology appeal but that can be very different in europe. I went to courses at central european university out of curiosity and the staff weren't really into marxism. Stuff like structuralism, psychoanalysis which are kind rooted in people who were also leftist/marxist, but that's not the important thing IMHO.

But naturally you will find people who are banana in the field, too. The academics (almost all women naturally) I met there were all fairly sane, reasonable and down-to-earth.

>>58759888
This guy knows what his talking about, architecturally its impossible to have a 6c/12t RYZEN CPU...FFS /g is going to shit.

>>58765848
>"Feminism" is a generic umbrella term for several different kinds of autism
ftfy
Social studies are only brain masturbation and are useless, don't waste you time trying to rank and understand how shitheads behave when confronting reality. They sholdn't even be alive in the first place.

>>58752556
>i7-6900K - 3.20 GHz base clock
>8C/16T SR - 3.30 GHz base clock
Now fuck off you retarded shill.

>>58765944
You mean he has read so on a thai website?
The sources that claim otherwise have more base. Also it makes more sense of yield purposes.

>>58765932
Well, obviously. Needless to say it's all rich, college educated "intellectuals" doing most of this shit, not refugees from Yemen or charity workers in Africa who know what actual oppression and suffering looks like. It's starting to gradually make its way to the more well-off countries in Europe as well, though. France and Germany already have several prominent Youtube vlogger type feminists spouting the same patriarchy and rape culture shit, and you're starting to see it pop up here and there in Central/Eastern Europe too, although in a much lesser volume.

>>58765984
From a yield point of view for sure, it would make sense. Lets see what happens, architecturally it doesn't seem to be possible. Also, one Tweet by some BS trading company is not going to get me to think that there is a 6c/12t version.

>>58765944
This is some pathetic disinfo.

>>58766094
Explain...

>>58766049
Well youtube should just not be taken seriously as a rule, IMHO. It has "youtubers", not experts or intelectuals, mostly. And if weren't just shallow and dumb attentionwhore/calculating egomaniac, you will never get the attention there anyway. So there probably are actual intellectuals there, but they have no influence/watchers.

>>58766128
Whats to explain? The fact that 6 core engineering samples exist, and have been known about for weeks?
Dipshits regurgitating clickbait headlines don't deserve to be spoonfed.

The rumor about there being no 6 core chips came from a single guy who admittedly was just guessing. Every echo chamber tech news site on the internet republished the exact same nonsense. They make money off of your stupidity, so they don't care if they post factually incorrect articles.
At least TPU and Fudzilla published later articles admitting their fuck up.

>>58751981
>6C/12T Ryzen confirmed having 3.3ghz stock
It confirms there will be a 6C/12T part with 3.3GHz, but that doesn't mean there won't be a variant with higher clocks, similar to the FX-6300 and FX-6350.

>Bad overclocking potential confirmed just like rx 480
Now this is just pulled from your ass.

>>58766730
It confirms there is a 6C/12T F4 ES clocked at 3.3GHz

>>58766964
And there's a confirmed F4 with 3.6/4.0, clearly clockspeed isn't a issue with Zen.

So just overclock that 6 core at 3.3, it's gonna be cheap anyway at those clocks.
I mean why would anyone buy a cheap OCable consumer CPU with a lot of cores and not overclock it?

>>58751981
actual non-shill translation:
>AMD validates 6C/12T Ryzen chips at 3.3GHz base but no 8C/8T chips. The goal is clearly to recycle failed 8C dies.
it says nothing about OC potential.

CPC is guessing that bad 8c/16t candidates will have 2 cores disabled to be sold as 6c/12t chips. they base this on the fact that the 8c/16t and 6c/12t chips clock exactly the same.

this means nothing as far as OC potential goes. one would expect clock speeds to go up as core count goes down, but in this case the dies are identical.

I'm not sure why people would want 8c/8t chips to begin with.

https://www.techpowerup.com/230291/amd-readies-ryzen-platform-drivers-for-windows-7

>>58751981
>just like rx 480
>a CPU architecture and a GPU architecture are comparable
I get what you're trying to say (that AMD can't into overclocking), but you know that's not a good comparison, right?

>>58768251
My school has tri core phenoms running at 3.1GHz stock. We managed to overclock one of them to 4GHz and it's been like that for months

We have no idea if it'll blow up if we run prime

>>58763807
>The trouble with 8/8 is that the performance lands pretty much at the same spot as 6/12 assuming around 30% performance from smt. Leaves them with marketing problems I think, I hope there's nothing without smt, i5s are in a shit position these days because of it.

yes and no, anything that doesn't properly use the 12 threads will run better on the 8 core instead, the 6/12 would be a bit cheaper then the 8 core in that case.

If you call the threaded cpus BE and use X(core number here) that solves the marketing problems as it clearly delineates what tiers of performance you should expect.

>>58765868
let me expain something to you.

ipc is how much shit gets done per clock
frequency is how many clocks there are a second

now when intel got 10 things done a clock and amd could at best muster 5, amd relied on proper threading to equal intel, but a fuck ton of us used programs with limited to no threadding support, and threading even when introduced was rarely seen outside of pro applications and even then only on the renderers.

however now amd has caught up to intel in this example, getting 9-10 things done a second.

intel is able to clock to 4.8 for almost every cpu and over 50% break 5ghz

lets use 4.8 as that's almost a guarantee number at this point.

at 4.8 intel gets about 30% more done per core then amd does,

so lets say amds 8 core 16 thread is 100 points a core 25 a thread

you on single core, 125 points
Intel at 4.8 has about 162

across 4 cores, amd at 3.4 will have 500 where as intel will have about 650

now, the interesting part,

amd's base clocks put the cpu at 1000.
Fuck knows where it turbos to
Fuck knows where it caps out
Fuck knows how it cools

But what we do know is canrad pc managed to get one core to 5ghz on its own, on an es that was tossing errors, so there is nothing architecturally that would limit it. knowing that it has about the same ipc as intels 8 core, and that its also soldered, and that it uses less power (2 watts big fucking whoop)

now If i remember right, nearly every intel 8 core can hit 4.2 so lets use that as an example. that would mean amd cpus can get 20% faster,

so single core amd would be 150
compared to intels 162

4 cores would be 600
compared to intels 650

all cores its intels 650 against amds 1200

granted this is all abstract and assumptions for the most part, but still shows you how small a lead intel has in cpu thread limited applications were as amd is going to either match or come close to them in price, and damn near double overall performance,

>>58766352
lets be fair here, 6/12 is as powerful as 8 cores because its in a weird scaleing area when everything works

If amd has no issues with yield, it would be better to disable smt and put an 8/8 over cutting 2 cores off.

>>58767866
THANK FUCKING GOD

I was SO not looking forward to windows 10 as a possibility.

>>58756426
athlon ii x3 exist.

>>58769402
They exist because x4 exists.

x3 wouldn't exist without x4

File: 1483140277087.jpg (2MB, 2480x3508px) Image search: [Google]

2MB, 2480x3508px

>>58766223
>Well youtube should just not be taken seriously as a rule

Tell that to the mass majority. We're talking about people who will eat anything they're spoonfed and net even check it's authenticity or question it.

>>58767866
Nice anon

>>58767866

WTF I LOVE AMD NOW.

>>58751981
>You can't make this shit up.
But you are right now.

>>58767866
The comments, how are these people so impress by this shit?
what the fuck? this shit is blatant marketing come on man.

>>58751981
>French PC enthusiast
>"verified" twitter
everything he says must be true!

>>58765559
Frankfurt School Marxism. Read up on it.

http://wccftech.com/amd-ryzen-6-core-8-core-processor-clock-speeds-confirmed/

>The third 8 core model is the “ZD3601BAM88F4_40/36_Y” which is the highest clocked model with base clock of 3.6 GHz and boost clock of 4.0 GHz. This is the highest clocked model we have seen so far under the F3 revision, which operates at factory shipped speeds up to 4.0 GHz and beyond that with XFR.