Quantcast
[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology


View post   

[ Toggle deleted replies ]
File: 277 KB, 556x727, 5I2.png [View same] [iqdb] [saucenao] [google] [report]
69490864 No.69490864 [Reply] [Original] [archived.moe] [rbt]

It's up to the jury to enlighten amd drones on what actually constitutes a core (and a good processor). Pretty embarrassing for ayymd owners, but they have probably got used to it.

https://www.theregister.co.uk/2019/01/22/judge_green_lights_amd_core_lawsuit/

>> No.69490873

>>69490864
>>69490864
.....................

>> No.69490895

>>69490864
>embarrassing for ayymd owners
Why?

>> No.69490913

>>69490895
Because it unveils AMD's marketing strategy of misleading bigger numbers because it rightfully takes their customers for gullible tech illiterate retards.

>> No.69490930 [DELETED] 

How long until ayymd gets sued over fake """7nm""" hoax?

>> No.69490965 [DELETED] 

>>69490864

i actually own an FX, is there any way to profit from this?

>> No.69490982

>>69490965
Yes. Throw it into the trash and buy yourself a decent cpu. Profit comes from enhanced user experience.

>> No.69490992

>>69490965
No because it will be tossed out. The judge was just incompetent.

>> No.69491029 [DELETED] 

>>69490864
>California judge
Protecting their jewish investors I presume

>> No.69491032

>>69490965
I have 3 FX-8320e CPUs. If I can get money that would be great.

>> No.69491049

>>69491032
>I have 3 FX-8320e CPUs
Are you perchance mentally deficient? Usually it takes quicker for people to learn from mistakes.

>> No.69491056
File: 1.55 MB, 640x785, 8.png [View same] [iqdb] [saucenao] [google] [report]
69491056

>>69490913
are you unable to count to 8?

>> No.69491071
File: 17 KB, 836x768, UPGRADE2010.png [View same] [iqdb] [saucenao] [google] [report]
69491071

a dont undestud about what this tread, but AMD FX super good CPU even now

>> No.69491094

>>69491056
he must be the judge.

>> No.69491113

>>69491056
I do
Bulldozer core however cannot because it lacks its own FPU. Just like amd niggers lack a brain to make informed purchasing decisions.

>> No.69491183
File: 13 KB, 209x241, 386.jpg [View same] [iqdb] [saucenao] [google] [report]
69491183

>>69491113
And the 386 had no FPU. What is your point?
FPU count ≠ core count.

>> No.69491192
File: 61 KB, 500x499, 1492027643450.jpg [View same] [iqdb] [saucenao] [google] [report]
69491192

>>69490864
>Pretty embarrassing for ayymd owners

In what kind of upside-down, topsy-turvy world are you living in, where people feel embarrassed because a company they bought a product from did something shady? Not everyone is a butthole-snorkeling cretin who supports multi-billion dollar companies like they're a football team, like you do, you sub-90 IQ waste of flesh.

Fags like you are the problem with this board, and in my opinion, society as a whole. Neck yourself.

>> No.69491228

>>69491192
People on /g/ here were pretty upset when Nvidia lost the 3.5Gb lawsuit.

>> No.69491257

>>69490864
Late to the party, no?

>> No.69491258

>>69490864
Who gives a shit, they still perform well for what they are.

>> No.69491263

>>69491113
>>69491183
shut down

>> No.69491314

>>69491258
Lmao no they dont. They were big and hot and expensive at launch for the shit tier performance you got. The best one is that 9390 or whatever that had a tray price of like $800 to jew the oems and as soon as it came out to retail they dropped it to like $300 which was still overpriced

>> No.69491329

>>69491228
Thats because the cult of Nvidia had to admit they were jewed.

>> No.69491352

>>69491071
>tfw still on 1156

>> No.69491356

>>69490864
They tried that "not a true core" once, and will fail again.

>> No.69491384
File: 2.95 MB, 320x240, Shekelshuffle.webm [View same] [iqdb] [saucenao] [google] [report]
69491384

>>69490864

>> No.69491474

>>69491314
I agree, but i’m mainly talking about now.They perform just fine if you want mid/high performance for dirt cheap.

>> No.69491883

>>69491113
>>69491183
One of the most definitive BTFOs i've seen on /g/ all this month.

>> No.69491969

>>69491183
The 386 is a 0 core CPU

>> No.69492029

>>69491314
>FX 9590 $1000.00 with liquid cooler
It performed worse than a 3.2ghz Sandy Bridge, and released during haswell.

>> No.69492064

Is everyone ready to claim their $5 from amd?

>> No.69492147
File: 224 KB, 600x495, AMD_Bulldozer_block_diagram_(8_core_CPU).png [View same] [iqdb] [saucenao] [google] [report]
69492147

>>69490864
Thankfully it's easy, for most of history, a CPU has been an integer core, with the floating point core being a special, separate upgrade. An FX module is two integer cores with one floating point core, giving two addressable cores.
Meaning in all intent purposes, a FX core is a real CPU core, like it has been for most of CPU history, a CPU by definition is still defined as something doing integer calculations. This will be very easy to explain to a jury with actual examples and proof.

Damn it feels know not being a dumbass.

>> No.69492152

>>69492064
will this be for mutts only?

>> No.69492184

>>69491352
>still on 1156
suck hard VS amd FX

>> No.69492212

>>69492152
>California
Most likely.

>> No.69492218
File: 71 KB, 512x512, 0.jpg [View same] [iqdb] [saucenao] [google] [report]
69492218

>>69492147

>Damn it feels know not being a dumbass.

>> No.69492239

>>69492147
>4 Core IF
>4 decode
AMD BTFO

>> No.69492256

>>69492147
What if the jury is filled with StarCraft players? AMD would probably be completely fucked especially since California has a lot of ethnic Koreans.

>> No.69492257

>>69492147
>Damn it feels know not being a dumbass.

embarrassing

>> No.69492413
File: 31 KB, 331x437, 1547516912898.jpg [View same] [iqdb] [saucenao] [google] [report]
69492413

>>69490864
Someone somewhere seems to be really trying to manipulate AMD stocks, first with the fake Ryzenfall and now with this.

>> No.69492420

HAHAHAHHAHAHA remember when amy faggots used to justify buying these piece of shit over Intel master race processors by saying "muh future proof multicore"

>> No.69492485

>>69491113
Kek

>> No.69492540

>>69492256
Jury members are weeded out by the attorneys from both sides if they're too familiar with the subject matter of the case, they don't want informed juries, just disinterested seat warmers who sway easily and vote unanimously.
By the way, lying and playing dumb to stay in the jury is the technically perjury.

>> No.69492593

>>69492413
JIDF has been on damage control for the last 2 years and so far their efforts haven't amounted to much, now its time to REALLY get down to business.

>> No.69492597

>>69492540
>By the way, lying and playing dumb to stay in the jury is the technically perjury.
So is claiming that 4 cores is actually 8.
AMD is going to get rekt.

>> No.69492633

Intel is the only cpu with cores. That's why they are called "core i 7" and "core i 5".

>> No.69492639

>>69491113
>>69491183

oof

>> No.69492656

>>69492633
Seems legit.

>> No.69492672

>>69492633
Xeons don't have cores. Intel btfo.

>> No.69492684

>>69492147
>Damn it feels know not being a dumbassOof

>> No.69492697

>>69492672
Xenon is for servers only and SERVERs are not for GAMING.

>> No.69492728
File: 36 KB, 480x678, 1547091205045.jpg [View same] [iqdb] [saucenao] [google] [report]
69492728

>>69491183

An FPU does not a CPU make.

>> No.69492779

>>69492147
Who's side are you on?

>> No.69493046

>>69490864
> And here's the sticking point: these two CPU cores, within a single Bulldozer module, share caches, frontend circuitry, and a single floating point unit (FPU). These shared resources cause bottlenecks that can slow the processor, it is claimed.
> Within each module, alongside the two x86 cores, is a single branch prediction engine, shared instruction fetch and decode circuitry, a single floating-point math unit, a single cache controller, a single 64KB L1 instruction cache, a single microcode memory area, and a single 2MB L2 cache.
If people seriously imply that sharing resources (where the sharing might impede performance) means that its only a single core, then congratulations: both AMD and Intel have only ever produced single core CPUs -- after all, they share L3, PCIe lanes, memory lanes (and memory itself), etc. along all 'cores'.
Where would you draw the line of which resources are allowed to be shared by 'cores'?
The only sensible answer is that any individual hardware instruction stream (nowadays called a thread) is a LOGICAL core. Which is what most people already agree on anyways. Making a distinction/definition of 'physical cores' is already almost impossible when you look at other kinds of processors (like e.g. GPUs, where vendors are already struggling with calling things 'cores' and rather call them e.g. 'wavefront'). The term 'physical core' has almost no meaning at all in modern (2000+) HW designs and the court decision should and probably will reflect this, whatever that means for AMD.

>> No.69493064

>>69490864
Faildoozer keeps delivering meanwhile everybody is still rocking their 2500k

>> No.69493069

>>69492147
Based and corepilled

>> No.69493090

>>69491384
do they have to do this every time they find a shekel on the street

>> No.69493100

>Or it can go to a jury trial in which members of the public (albeit likely tech-savvy ones since the case is taking place in Silicon Valley)
This is The Register, so I'm not sure if he's being serious. I really hope he's not.

>> No.69493113

>>69492597
It's a silly lawsuit, by all intent and purposes, it's actually 8 cores with 4 FPUs.

>> No.69493126

>8 cores
No no I said
>quaternATE reflectORS
That's what we call quad cores where I come from

>> No.69493131
File: 69 KB, 646x687, 1510593104325.jpg [View same] [iqdb] [saucenao] [google] [report]
69493131

HAHAHAHA AMDRONES LITERALLY BTFO

Can't wait to see the delusion when the court rules against AMD.

>> No.69493143
File: 90 KB, 1080x1350, 1547960653697.jpg [View same] [iqdb] [saucenao] [google] [report]
69493143

>>69493046
It's a lawsuit for the matter of being a lawsuit. I wouldn't be surprised if it's some Intel shadow puppet behind it, like with Ryzen Fall.
It's like people suing over a hot coffee not having the words HOT on it and they burn themselves.

>> No.69493151

>>69490864
>It's up to the jury to enlighten amd drones on what actually constitutes a core
Pretty ridiculous that it's not engineers.

>> No.69493154

>>69493113
Nobody from AMD ever claimed that they were eight cores outside of their marketing department.

>> No.69493162

>>69493151
Welcome to US of A law. People least qualified must be the jury, otherwise it would be unfair!

>> No.69493181

>>69493154
AMDs own documentation around it calls "modules" as 2 integrer cores with 1 floating point core.

>> No.69493205

>>69490864
Intel shills out in full force I see.
Can't wait for you fucks to have a mental breakdown later this year.

>> No.69493213

>>69492413
>intel cause everything bad that happened to AMD
wtf with this loser mindset

>> No.69493224

>>69493181
Well then they're fucked.

>> No.69493225

>>69493213
losers adore AMD because they see it as a fellow underdog in the world of big boys

>> No.69493251

>>69491113
God
DAMN

>> No.69493257

>>69493224
Not really, since a each core has its own integrer core that shared resources. They can literally show 8 8088's and say that this is no different and they would be right.

>> No.69493270
File: 49 KB, 1280x720, (((henry))).jpg [View same] [iqdb] [saucenao] [google] [report]
69493270

A core is a core, you can't say it's only 6.8 cores just because it lacks all the features of a normal CPU.

>> No.69493279

>>69493225
I just don't like kikes

>> No.69493284

What's up with Intel fanboys thinking this is breaking news though? When I was browsing FX CPUs as an option for my parents computer, I saw >>69492147 picture several times, stating clearly how the Module/Core system works. People buying those CPUs weren't lied to, the modules are just normal cores sharing FPUs and everyone buying it could see it for themselves.

>> No.69493297

>>69493225
>multi billion dollar company
>underdog
The only reason I buy AMD is because I'd never buy Intel or Nvidia, for obvious reasons. If there were even other alternatives, I'd go by price/performance.

>> No.69493322

>>69493297
> for obvious reasons

because you detest good hardware?

>> No.69493335

>>69493297
>poorfag nigger who concern about $100 different in price

>> No.69493418

>>69492147
Lmao. How can you fuck up this bad. You were fine up until that last sentence, then you went full brainlet

>> No.69493442

>>69493335
Explains why I bought a GPU with the exact same performance and price as another companies GPU.
Yeah, no, nice try though.

>> No.69493471
File: 12 KB, 225x225, 1548256102224.jpg [View same] [iqdb] [saucenao] [google] [report]
69493471

>call up actual Intel engineer
>make him swear so he can't lie
>reveal a table with a 8 core FX and 8 single 8086 chips
>ask him if in the core design of the cores they are different in the definition of a "core"
profit.

>> No.69493775

>>69493143
>It's like people suing over a hot coffee not having the words HOT on it and they burn themselves.
Ah, I can see you are an idiot. You've never actually looked into the particulars of that case have you?

>> No.69493953

>>69493257
What if someone asks about the 4, and only 4 "core ifs"?
>>69492147

>> No.69493995

>8cores
Well you see the 4 cores you see in the diagram are actually 8, and here's why. We split a core into two, and now 8.

>> No.69494054

>>69490913
Its only wrong when intel does sir. Pls delet.

>> No.69494514

>>69491113
how will amd pajeets ever recover?

>> No.69494531

even when they're winning ayymd is still getting cucked

>> No.69494565
File: 129 KB, 745x500, lelddit.png [View same] [iqdb] [saucenao] [google] [report]
69494565

>a fucking redditor is more knowledgeable than all /g/entootards in this thread
/g/ BTFO

>> No.69494689

>>69491071
FX sucks for single threaded applications tho. Also dat powerbill.

>> No.69494753

Sounds like they needed to debut that there are registers running the machine for the processors and then when the cache is revealed the arm core is significant and thus the processor may not suggest activity the way it should. Sounds like a jew read over debut and thought a new thing instead of meaning an old thing with a new exception like clauses for engineering practices and not video gaming theory cores like we get from intel and not so ayyyy md. i think they can defend if they switch lawyers.

>> No.69494857

>>69491329
the only people that cared about the 3.5gb meme were seething amd cucks who were jealous about missing out on the best gpu deal of all time that was the 970, the 2500k of graphics cards

>> No.69495035

>>69491192
based and redpilled

>> No.69495321

>>69491183
nice

>> No.69495336
File: 99 KB, 1280x960, 19-113-284-04.jpg [View same] [iqdb] [saucenao] [google] [report]
69495336

>>69490864
>when you still create controversy over 7 years after you launch
FX is a strange thing, the integer cores are real cores, but the 8350 behaves like a doubled up, hyperthreaded quad core, to truly use an FX module, which is really what the FX core consists of, you'd have to use the 2 threads, and to use the full four modules, you need to be leveraging the 8 threads.

>> No.69495426

>>69491049
The 8350 chips were strong as fuck if you didn't game.

>> No.69495541

>>69490864
tfw my fake chip at 1/4th the cost works better than anything intel provides.

>> No.69495558

>>69491183
HAAAAAAAAAAAAAAAAHAHAHAHAHAHAHAH

>> No.69495569

am*rican cpu manufacturers in a lawfare knife fight while chinese moon people go down to 5nm

>> No.69495628

>>69495569
>while chinese moon people go down to 5nm
Highly doubt.

>> No.69495648

>>69494565
>all these amddrones btfo'd
Glad I got a 2500k back then, can you imagine seriously buying a bulldozer?

>> No.69495674

>>69495628
7 foot 12 inch genetically modified chinese supermen are going to be banging your girl while intel funds "moderate" luddite groups to oppose ayymd interests

>> No.69495728
File: 2.04 MB, 2576x1932, 1546052593322[1].jpg [View same] [iqdb] [saucenao] [google] [report]
69495728

>>69495426
This, i7 4770 non K scores about this much.

>> No.69495744

>>69491113
>>69491183
Gekko in Gamecube has 2 FPUs.
Is it dual core?

>> No.69495745

>>69495569
why would they show their ass to such small people?

>> No.69495756
File: 11 KB, 384x384, 1518938266539.jpg [View same] [iqdb] [saucenao] [google] [report]
69495756

>>69495674
I've got a secret weapon for those chinese abominations.

>> No.69495758
File: 23 KB, 228x221, jewdicialsystem.jpg [View same] [iqdb] [saucenao] [google] [report]
69495758

>>69491032
>If I can get money that would be great.
Here's how it works. Assuming they win, you get a 3 $5 checks. The attorney gets $50 million.

>> No.69495986
File: 349 KB, 2560x1440, Untitled.jpg [View same] [iqdb] [saucenao] [google] [report]
69495986

>>69495728
dumb question, but why is your core voltage so high?

Also how did you get a 25mhz FSB OC stable?

>> No.69496038
File: 886 KB, 858x1000, AB50837_1.jpg [View same] [iqdb] [saucenao] [google] [report]
69496038

>>69495986
The 1.4v is needed to get the overclock truly prime 95 AVX stable, my chip can do rendering at 4.4GHz stock voltage but it crashes on prime, I'm not sure how my northbridge went +25MHz without a voltage increase, maybe it's the really good load line calibration on my board, I'm using the Gigabyte 990FXA UD3.

>> No.69496047

>>69496038
Also FX chips can handle 1.5375v on the FX9000 chips that are the same silicon, so it's well with in the VID range of 0.936v to 1.4250v for turbo.

>> No.69496164

>>69496038
Prime 95 fails on stock settings, even way underclocked for me on the FX chips. a random worker/thread will fail, it does not seem to be tied to any one core or compute package. it could be related to a Low EMI emissions setting i have turned on to eliminate hiss out the audio jack.

>IBT AVX Is stable for me @ 4.4ghz 1.288Volts, NB i can mess with, but as soon as any tweaks happen to the HT frequency it fails to post, OR starts throwing random memory errors.

My ram tested fine to 1866 @ 1.5v, but i backed it to 1600, just because i saw no benefit.

Boards a AsRock 970a-g/3.1 - not the best, but it was hard getting a am3+ board that didnt suck ass in 2018.

Also for what its worth, i run a massive BeQuiet top flow cooler, with a silent fan curve. IF i up to 4.5ghz on the cores and transcode video or something, i have to speed up my cooling fan.

I valued silence over out right performance, though sometimes it would be nice to have a little more grunt.

Running games like doom or whatever CPU stays below 50c (though i know the FX chips measure implied temperature deltas and other weird crap, and lack a real thermister) VRM stays around 45c, and GPU (rx580) sits around 50-54c.

>> No.69496212

Shared FPU doesn't matter. If they shared a 1024-bit FPU that could have been faster than dedicated 256-bit FPUs. FPU count isn't core count. And furthermore, core count isn't indicative of performance.

All AMD needs to show is that it could execute 8 instruction streams simultaneously and they should be OK.

>> No.69496238
File: 266 KB, 1694x1200, RAM benchmarks.png [View same] [iqdb] [saucenao] [google] [report]
69496238

>>69496164
>Prime 95 fails on stock settings
That's really strange, I've mine locked to 4GHz at dead stock and it never fails prime.I too paid about msrp for my new 990FX UD3 in early 2018 because I didn't want a 760G board.

FX really doesn't give a shit about how fast your memory is, I don't know what it's deal is but I'll take it.

>>69496212
AMD will be fine, even back in 2014 they were pretty much like "well it has 8 processing thingies so it's 8 cores" and that is technically good enough

>> No.69496245

>>69496212
You can execute 8 instruction streams on a 2600k dummy, does it make it an 8 core cpu?

>> No.69496261

>>69496245
No the hyperthreaded cores are not physical, AMD FX pretty much is physical hyperthreading.

>> No.69496285

>>69496261
Yes it is indeed AMD's take on HT, so they should've marketed it as a 4 core/module CPU.

>> No.69496297

>>69496245
no, simultaneous instruction streams, not swapping between two front-ends over a shared back end like they do for their FPU

>> No.69496315
File: 138 KB, 930x797, NgBdlEE.jpg [View same] [iqdb] [saucenao] [google] [report]
69496315

>>69496285
Marking it as a quad module, 8 thread CPu would've saved everyone the trouble, because when you understand what the 8350 is and how it works, it was actually an awesome CPU, and because AMd market it as 8 cores, reviewers judged it based on one of a modules cores, which you cannot do because the "Bulldozer core" block diagram consists of a module with it's 2 cores and resources being used all at once.

>> No.69496337

>>69496261
That is a more accurate way to put it.
They are still 8 cores, but they are weak ones which can be a bit stronger when the sibling core isn't in use.

I don't think there is an established and accepted definition of "core" though.
For AMD to be sued over Bulldozer, everyone would also have to be sued over their touting of "graphics cores". Those aren't real cores. GPUs are really more like one massive graphics core as they only have a single pipeline for them all.
It's also not fair to say having shared pipelines and resources makes a core any less a core, as various CPUs over the years have had more or less shared components.

>>69496315
Yeah but that's not entirely accurate either. Nothing you can say in 2-4 words to put on marketing is accurate. It really was reasonable for them to call it an 8 core... just that it was 8 weak cores.

>> No.69496350

>>69496315
>>69496337
Also as your image points out, it was clearly two cores when it came to int operations.
It was just in FPU that it was sometimes faster if it was treated like a quad core.

>> No.69496357

>>69496238
it is, ive never seen something like it before. But its clearly stable if it can pass IBT AVX for an hour + or crunch numbers wide open on the CPU and fold proteins on the GPU, hammering the entire board. (i use it to heat my office when not actively using it)

I doubt its a v droop issue, as the board rated to the 220watt chips. PSU is a 800watt Gold Seasonic.

maybe i just need to up the voltage, but that means more heat and noise for marginal gains at that point. 4.4 on 1.28v is doing pretty good based on what i see online.

>> No.69496413
File: 2.28 MB, 4032x3024, IMG_0084[1].jpg [View same] [iqdb] [saucenao] [google] [report]
69496413

>>69496337
>which can be a bit stronger when the sibling core isn't in use
You are right about that, the ipc of one core with all resources used was that of a high clocked Core 2 Duo, It was weak, and AMD Fx was so dependant on the multithreading that didn't really exist for gaymen in 2011, which is sad because for the $70 you can get a new 8350 for today, it's actually a good budget rendering chip that you can have fun overclocking and it's not the end of the world if you kill it.

>>69496357
Passing IBT is plenty stable, and your board and PSU are not the problem, maybe cranking the VCORE to 1.35v would make it stable because that's what mine is, I've got a crappy CX750 with one 12v rail and huge amounts of vdroop isn't an issue, and my load line calibration works great, it's not spiking the power and killing it.

>> No.69496443

>>69496164
FX series has a massive memory controller bottleneck. The "CPU NB" frequency determines the memory controller speed and the default 2400 (?)mhz bottlenecks dual channel 1600mhz DDR3.

On Intel the speed of the memory controller is the CPU frequency, so a 5ghz i7 2600k has a 5000mhz memory controller.

>> No.69496505

>>69492147
>in all intent purposes
>Damn it feels know not being a dumbass.
Anon, I...

>> No.69496517
File: 2.08 MB, 3264x3678, IMGP4415-meh.jpg [View same] [iqdb] [saucenao] [google] [report]
69496517

>>69496413
looks good considering its sitting on the floor.

terrible distortion from a 18mm lens incoming.

>> No.69496526

>>69490864
Whatever you shills say
The FX is eternal

>> No.69496570
File: 726 KB, 1600x1036, Fine Wine.png [View same] [iqdb] [saucenao] [google] [report]
69496570

>>69496517
Damn bro that's fucking nice, downdraft cooler is technically how CPU cooling is supposed to work.

>>69496526
*sip*

>>69496505
Despite being a little drunk at the end, he made a good point, an FX core is technically the two integer cores in that module

>> No.69496602

>>69496570
i wish i had gotten a larger case that allowed for more drive bays and a internal BR drive... but live and learn.

I just doubled up SSD's on one 3.5" tray, and dual laptop drives on the second 3.5" tray, which is why i have the 3rd intake fan down there to cool that mess of a bad decision.

>> No.69496691

>>69491384
>Rabbi Jacob
>70s France when movies like this were allowed
>nobody dares to make comments because it's a well known classic

>> No.69496794

>>69496413
It was weak for gaming and some common uses, but I would have still recommended it over a 2c/2t i3 that it was often priced to compete with for a couple of years. It at least wasn't a stutterfest like those.

>> No.69496858

>>69496794
>it was often priced to compete with for a couple of years
Yeah even in 2016, the $159.99 an 8350 cost was still way better than a low end intel offering, and you can get a Gigabyte 990FX board for 89.99 on amazon to go with it, overclock to 4.5 easily and be fast as hell.

>> No.69496892

>>69496794
it might be weak for fast refresh rates, but locked to 60hz i have no issues with Dirt Rally, REDOUT, or doom 2016. CPU bound titles like Civ Games run fine. ASSCREED Odyssey runs fine at 1440p Ultra 30FPS, or high 60FPS, without drops or stutters unless you go to the bird for an overhead view in a area with dozens of NPC's.

Maybe i am just exceedingly lucky, but i really cant recall being CPU bound for games in like 15 years or so.

Dooms not the best example but it has a killer overlay that shows my GPU gives me a frame time of ~6ms, and my CPU is around 9ms.

SO yes, if i was trying to do 120hz or 144hz, it would be a limit...but really how many people sit and play non-competitive or single player games up in that region?

>> No.69496938

>>69496570
Judging from the color alone, that wine doesn't look fine.

>> No.69496943

>>69496858
There were times where you could get an 8300 or 8320 WITH a motherboard for $130-$150.
Yeah, definitely didn't compete with a i5k let alone an i7, but it was better (except for power consumption) than intel's offerings at those prices.

But really even then... everyone I know who got bulldozer seems to be happier with it than people I know who got #500k CPUs. 4 threads started to really get stuttery, even if the average fps is much better.
You have people who got a 7600k a year or two ago, and are already looking to replace it because it's stutterfest garbage, where you still have people with Bulldozer that are just happy their frametimes are consistent so it's less noticeable.

>>69496892
It only gets 30 in some games like War Thunder, afaik. Or maybe that's just World of Tanks. I hate using WoT as an example as it's a shit game.
But I guess that's not any worse than a 2500k which averages 70fps, but constantly stutters down to 20fps for a fraction of a second.

>> No.69496963

>>69496794
the other thing to point out is that IF you installed spectre and meltdown patches on a 2nd and 3rd gen i7 package, you were basically totally removing the IPC advantage that intel had. And then you opened up your computer to total meltdown.

without dangerous memory optimizations and preemptive processing, turns out the intel chips were about neck and neck with an AMD chip that was in the same ballpark/class.

who knew?

>> No.69496990

>>69496963
A 4790k with the patches still beats a 8350 but a fair bit. 2600k probably does too, but I'm not certain. It looks like it'd be very close.
But yeah if you had to use either an 8350 or 2500k today, I'd go with the former.

>> No.69497051

>>69496990
I've been happy with it for the 5 months ive owned it. It Has decidedly held up better than even dual CPU xeon workstations i had given out to family members over the years. Not to mention drastically quieter

it my first AMD system since a k-6 in the 90's and it making me wonder why i strayed away. It just works, its easy, and its cheaper. electrical here is so cheap its a non-relevant data point.

>> No.69497062

>>69497051
>electrical here is so cheap
Kek, the FX power chugging is a meme, 125w is not alot, intel chips have been 130w since the bloomfield i7s and they use even more power when you overclock it, but no one seemed to care because intel can do no wrong.

>> No.69497075

>>69495628
TSMC is Taiwanese, who are the true people's republic of china

>> No.69497143

>>69497062
Pretty much what i noticed. In game I pull about 330-400 watts from the wall. Not really that bad when i consider that my newest laptop pulls 230watt from the wall when docked and can in no way play games or render out like the desktop

Well below an average intel and Nvidia based Cad workstation that i generally get and updates once ever 2-3 years. My last dell precision desktop fully loaded out could pull 890watts, and barely ties the CPU and doesnt come close to the GPU in cinnebench

>> No.69497196

>>69497075
Who do you think you are fooling?

>> No.69497205

>>69497143
>890watts
Dayum was that a dual CPU? I remember when Intel was pushing the 45nm Penryn based xeon quads to over 3GHz, the tdp was 150w and people would have two of them in their mac pro or whatever workstation they had in 2008.

>> No.69497229

>>69490864
I would have thought FX's poor sales would have been punishment enough--almost forgot they got sued.

>> No.69497263

>>69491183
God
DAMN

>> No.69497266

>>69491183
*SX

>> No.69497290

>>69496691
Movies like that one aren't anti-Jewish and hence can be made even today.
>>>/pol/

>> No.69497313

>>69497229
FX didn't deserve to be squashed in the market. 4 cores 4 modules did perform pretty decent, it performs 10% faster than a first gen i5 at 4GHz stock 4 cores 4 modules, and if you overclock it to 4.6, which is easy even on stock cooling with such few cores running, it performs like an i7 920, and if your operating system's scheduler knows how to use it, it will switch between one core per module/ 8 cores 4 modules on the fly and you can have decent single core and really good multi core depending on what you're doing.

>> No.69497314

>>69493090
yes, its Hebrew tradition

>> No.69497316

>>69497205
Dual CPU, Dual quadro, with a 50watt raid card too.

>> No.69497327

>>69497313
>first gen i5
10% faster than an i5 760 at stock 2.93, i mean.

>>69497316
Holy shit, PSu couldn't even push that, the thing will shut off at 825w load which is pretty good for a 750.

>> No.69497348

>>69497327
it had a 1000watt slide in PSU that loaded from the back.

>> No.69497360

>>69497348
>1000w
This must've been a ballin precision, even a high end workstation like the HP Z800 only had an 850w supply.

>> No.69497372

>>69497290
Of course it's not antisemitic. And no, it can't be made today.

>> No.69497392

>>69491056
Hi amd

>> No.69497398

>>69497360
it was overkill, and sounded like a server. got it scratch and dent from a local business that went under

>> No.69497399

>>69491258
Kek

>> No.69497498

>>69497062
That power consumption of my i5-2500k bothered me.
I had a 7970 as well.
Thing would put out like 450W total.
Now I have a 1600X and Vega 56, and my PC is cold to the touch most time, cold air comes out of it. It's cooler than my monitor is by far.
I haven't measured from the wall, but I'm guessing it's around 50-70w in basic use, and around 100-180 during most gaming at 2560x1600. Approaching 300 for really demanding games and if not limit to 60fps.

I live in FL and really don't like the heat and I'm glad there was finally something worth upgrading to, and so cheap. My total upgrade was only around $705.

>> No.69497518

>>69497498
>7970
That thing is still a beast for how cheap you can pick a used one up but that 250w tdp is kinda alot.

The new ryzen stuff and cards are great if you're into tdp minimizing, the 2400G performs like a 5GHz FX-8350 and uses 65w I think.

>> No.69497596

>>69497518
The worst part is that it uses a minimum of 100W with a 2nd monitor plugged in, instead of the usual 25 or 28 or whatever.
Meanwhile the Vega 56 uses like 10-15.

Just the i5-2500k and 7970 at or barely above idle were hot as fuck.

Ignoring the heat consumption, yeah the 7970 was fine for 1920x1200 gaming on medium or high settings. But I wanted to play stuff in better quality.

>> No.69497680

>>69490864
This is the dumbest shit ever, pretty sure Intel is backing the class action law suit just to drain some money from AMD

AMD is going to easily win the case simply because every core is capable of doing an independent instruction.

>> No.69497714

>>69497680
>when your products are so terrible you have to go after a 7 year old product that wasn't bad anyway
Intel are the jewiest, recount the election motherfuckers ever, these are the same people selling defective iGPU i7 for the same money.

>> No.69497737

>>69491183
ETERNALLY BTFO’D
QUICK SOMEONE MAKE A SCREENCAP OF THIS

>> No.69497771

>>69491183
Gonna need to sue Intel for years of false marketing too.

>> No.69497819

>>69492147
>4 fetch
>4 decode
>4 fpus
>recognized as quad core in linux and windows
>it's an 8 core tho lol

>> No.69497843

>>69497819
not an argument

>> No.69497924

>>69497819
It is an 8 core, but setup in a way where the OS sees it as a hyperthreaded CPU because it does work that way.

>> No.69498142

>>69490864
>enlighten amd drones on what actually constitutes a core
You mean decide what holds up as a core in court. And that would be interesting. Ignore AMD/Intel for a moment and look at all the other CPUs we use today. Giving a definition of what a "CPU core" is gets pretty difficult.

>>69492147
I disagree. I already thought about Bulldozer as having some kind of "half-cores" when it was released because it had two INT cores sharing a FPU. And this puts the lawsuit in question because this wasn't some big secret they were trying to hide, pretty much every CPU review mentioned it.

>>69493046
>AMD and Intel have only ever produced single core CPUs -- after all, they share L3
I too am very curious as to how you're supposed to draw that line. Oh, your 3 year old 4 core CPU can't do AVX2 instructions at all? Well, then, I guess you've got scammed and have a 0 core CPU?

>>69496261
>>69496285
>>69496297
Wouldn't say Bulldozer is the same as hyperthreading/SMT, but it does raise an interesting point: If Intel decided that it's too hard to keep up with AMD Ryzen's increasing cores and just declared that their 6core/12thread CPU are 12 cores we'd all scream scam. So one has to draw the line somewhere.

>> No.69498250

>>69498142
AMD's bulldozer does work like hyperthreading, but the cores are physically there, yes it performs like a hyperthreaded quad core, but they're all there, we can debate about if it's a crappy 8 core for the money or not, but it is 8 cores.

>> No.69498276

>>69497819
What version of windows sees it as a 4 core? I have never seen an 8000 FX recognized as only 4 cores in an OS

>> No.69498279

>>69498250
Do you think hyperthreading uses fictional resources or something? The hardware is there for hyperthreading too, that's how it works, it uses hardware thats not being used at the moment to process other work, which is why the performance is only like 10-25% improved.

>> No.69498303

>>69493143
sauce?

>> No.69498322
File: 50 KB, 418x720, 4.5GHz 1866 bench.png [View same] [iqdb] [saucenao] [google] [report]
69498322

>>69498279
>10-25% improved.
Theere may be some extra logic on the die to double the threads, but it's not much, AMD FX with one core per module is 69% slower than 8 cores and 4 modules, I can get 730 at 4.5GHz on 8 cores and only 427 with one core per module.

>> No.69498696

>>69498322
730 to 427 is not 69% slower. It's 41.5%.
It *should* be half, if not for the way Bulldozer do.

>> No.69498763

>>69491049
They were the cheapest for my servers. They have plenty of power for that.

>>69495758
Almost a tank of gas.

>> No.69498827

>>69498696
Ah shit that calculation was a tad bit off, but yeah it should be half as fast, but it's not because there is a huge fetcher and decoder bottleneck when two cores in a module are both working, the ipc dips about 20%, I can score 109 points at 4.5 single core in a module on cinebench, but if you divide the 8 core score of 730 by 8, it's only about 91 points if both a modules cores are processing simultaneously.

>> No.69498902

>>69492420
I prefer to be a cheap fuck with money to spare than a poor fuck

>> No.69499025

>>69498276
In all windows it reports as a 4 core 8 thread chip, with 8 boxes for "threads"

>> No.69499106

>>69497680
>Blaming intel
The lawsuit was started when amd claimed that they had 8 cores, and it was against Intel's interested to bury amd.
AMD always had aggressive somewhat shady marketing tactics. It's what they where known for since day one, so you can't blame anyone else for that.

>> No.69499136

>>69497518
>That thing is still a beast for how cheap you can pick a used one up but that 250w tdp is kinda alot.
I put mine to rest, and wouldn't buy another one no matter how cheap it is. It served me well, but it's glory days are behind.

>> No.69499223
File: 30 KB, 326x297, KDPFXVa.jpg [View same] [iqdb] [saucenao] [google] [report]
69499223

>>69491192
I honestly love this post. Here, have a meme.

>> No.69499269

>>69492147
I give you a perfect 5/7.

>> No.69499330

AYYMD IS FINISHED & BANKRUPT

AYYMDPOORFAGS CONFIRMED ON SUICIDE WATCH

>> No.69499531

>>69499330
>Have inexpensive CPU that performs well for the money for a few years
>May get a check for a few dollars if AMD somehow loses which isn't likely
I give this post a 3.5GB/4

>> No.69499543

>>69499531
I give this post 8 out of 4.

>> No.69499548

>>69490864
Imagine being in court trying to explain what a core is and judge doesn't even understand ohms law.

>> No.69499566

>>69499548
That's to amd's advantage actually, but I'm pretty sure a judge has been to highschool.

>> No.69499615
File: 83 KB, 653x726, 1491349230554.jpg [View same] [iqdb] [saucenao] [google] [report]
69499615

>>69499543
Fug that was pretty good.

>> No.69499627
File: 133 KB, 600x400, 1454283815899.jpg [View same] [iqdb] [saucenao] [google] [report]
69499627

>>69491183
Nice one, anon

>> No.69499855

I bought the bulldozer because I'm poor, and I was really boxed in in terms of CPU options.
That said:
>one module is "two" """cores"""
>only one FPU per module
>only one INSTRUCTION DECODER per module

Who the fuck thought it was smart to make cores share the slowest and most bottlenecking parts of the CPU?
I loathe intel, but that doesn't mean AMD deserves a pass on this shit.
I only paid $65 for mine off a tray from some guy on amazon. Thank god I didn't pay more.
Bet your ass I will be after my 5 shekels from the settlement, too. Will be worth the 4 shekels postage I pay to get it.

>> No.69499872

>>69499543
it's 4 out of 8, Christ you incel fags can't even meme right
>>69499615
fug what a samefag

>> No.69499925

I remember when I got ahold of my first FX, a 6300, I had already did my research on the CPU and found out about the FP core/module thing, but that didn't stop me from enjoying it. The main point being DO YOUR FUCKING RESEARCH before you buy, goes for anything. If anyone truly feels conned then they probably have sub 90 IQs

>> No.69500033

>>69499925
The 6300 was a damn good buy before the price of the 8350 fell like a ton of bricks, there was a time where it was costing half as much with 3/4 the cores.

>> No.69500191

So many people defending bulldozer, yet i still get just as good fps sometimes better with my Phenom II x4 955 Black.
Anyone who defends bulldozer is just coping on their terrible purchase
IE tru story
Friend and i had the same gpu powercolor 7870 Myst (7930) since it uses the 7900 series chip.
I had the 955 black
He had the vishera fx 8350
Real test in games, he got maybe 5 fps more than me at any given time.
On older games like CS or WoW i someimes got 1 fps more.
And bulldozer is an energy hog, seriously the worst mistake ayymd made.
I hope ryzen 2 is like the phenom days.

>> No.69500213

>>69499872
>it's 4 out of 8,
No. Bulldozer doesn't have 8 cores, but amd claims that they got eight cores with only four hence 8 out of 4.
Stop being dense, and just wait for your settlement check brainlet.

>> No.69500236

>>69500191
There's no way a 955 at 3.2 would be getting better frames then a Vishera which is clocked way higher, and with the proper windows scheduler would be operating one core per module with "good ipc" as well as SSE4 instructions and a faster HT link to go along with it.

>> No.69500242

I'm happy about this. Shared resources between cores doesn't double the cores. Not sure why this lawsuit didn't happen sooner.

>> No.69500253

>poozen about to shit on israels market share
>lawsuit
BIG THINK

>> No.69500258

>>69500242
It did in 2014, though, they lost because it has 8 processing things that clock high but go kinda slow

>> No.69500268

>>69500236
My phenom II oced to 4ghz gamed better than my FX 83xx @ 4.6ghz especially in arma 2 multiplayer. Games were playable on the phenom 2, but not with the higher clocked fx.

>> No.69500451

>>69500242
>Shared resources between cores doesn't double the cores
It doesn't halve them either, which is essentially what the lawsuit's claiming.

>> No.69500486

>>69490864
what constitutes a "core" in the first place?

>> No.69500953

>>69500486
It's largely the things around the AGUs/FPUs themselves, like instruction decoding and scheduling. But there's a bunch of grey area.

It's pretty well agreed that GPU cores aren't really cores because they only have a single scheduler. But for everything in between, it's hard to say.

>> No.69500981

>>69490864
This is just some lawyers looking for something stupid to sue a company over to justify their law degree.

>> No.69501163

>>69491056
Where is the PSP?

>> No.69501228

>>69501163
A perfect CPU doesn't need backdoors

>> No.69501246

>>69492147
intredasting

>> No.69501279

>>69490864
Based class action will make AMD fail and disappear along with their stinking shit, leaving only superior Intel products to reign supreme. And Peace will be on Earth once again. Amen.

>> No.69501655
File: 296 KB, 640x785, 1548267022413.jpg [View same] [iqdb] [saucenao] [google] [report]
69501655

>>69491056

Sorry, I can only count to four.

>> No.69501690

>>69501655
Just remember that there are two cores in each box you counted.

>> No.69501767

>>69490864
Delete this! AMD IS GOOD GUYS! They never did anything wrong, and if they did it's nothing compared to other companies! It doesn't even count! AMD is my friend!

>> No.69501857

>>69501655
well we know you're not an nvidiot at least but you still seem a few brain cells short of human

>> No.69502006

>>69491183
#REKT

>> No.69502036

>>69490864
To be fair, fake news media aside, Bulldozer was a complete mess. Not only that, but most AMD fan boys at the time even mentioned that the marketing for it seemed sketchy and unimpressive.

I got into AMD right before Bulldozer was announced and I remember it clearly. Ryzen line really brought AMD back into the game and Raven Ridge / Bristle helped too, but Bulldozer was just... Bull Dozer.

Interested to see how this court case plays out.

>> No.69502346

>>69502036
Bulldozer is what persuaded me to go AMD. I had a lot of fun overclocking my old Core 2 Quad. I was inexperienced at the time and made a lot of mistakes that eventually lead to serious stability issues that I couldn't resolve. I looked into Ivy Bridge and decided that the price was too much. I ended up building a new budget computer that was based around a Bulldozer APU as a stop gap until Intel prices became more reasonable. It stayed as my main machine for 4 years. The whole computer cost me around $250. The price to performance ratio was simply outstanding. After spending a decade paying for what I thought was the best from Intel, I couldn't believe that I could get so much for so little money. I now run an all AMD set up based around an 8350. Currently I have no plans to upgrade to Ryzen because the 8350 is more than enough for what I do.

Bulldozer wasn't the best. It has glaring flaws but it was never as bad as it is made out to be.

>> No.69503166

>>69501279
Yes I loved it when things were peaceful when intelaviv l wouldn't make anything better than 4c8t on mainstream mobos with fucking paste under the ihs.

>> No.69503269

>>69493213
>>intel cause everything bad that happened to AMD
he never mentioned Intel...
What a dumb caricature anon

>> No.69503336

>>69502346
>glaring flaws
Apart from the weird design and ipc thing, FX did nothing wrong, soldered spreaders, the entire range is unlocked, all the boards are unlocked, a 970 board cost less than a p67 board, the 900 series chipsets with the SB950 have all sata 6gb/s ports that don't die like the non b3 sandy bridge boards, there's just no fuckery with the and fx platform but people want to make controversy.

>> No.69503867

>>69490913
>Because it unveils AMD's marketing strategy of misleading bigger numbers because it rightfully takes their customers for gullible tech illiterate retards.
amd shills here have proven over and over again how tech illiterate they are. i imagine it only gets worse for those with less of a clue. I hope the lawsuit is a success.

>> No.69503917

>>69503867
It likely wont work because the cores are physically there, they're shit, a 4c8t Intel quad will beat it, but it is eight cores, and can prove this by showing 4 core benchmarks being damn near half as fast.

>> No.69503992

>>69503336
bulldozer had great IPC,
its lack of FP throughput, placed it behind intel cpus.
Why did amd designed BD that way?
Because you were supposed to offload heavy FP workloads on the GPU. This is what HSA was supposed to do on both amd builds and every Arm device.
an 8350 could get more IPC(instructions per CLOCK) and better performance even than a 4790k on integer workloads.

>> No.69504028

>>69491183
/thread

>> No.69504158

>>69492147
What it meant in the 80's isn't relevant. For the last 30 years it has meant integer + FPU.

If you marketed something as having "Bluetooth" simply having an image of Haraldr Gormsson, you likely would lose a fraud case even though Bluetooth technology did not exist for most of human history.

>> No.69504172
File: 193 KB, 1085x1018, 1517621496224.png [View same] [iqdb] [saucenao] [google] [report]
69504172

>>69497737
here ya go.

>> No.69504279

>>69503992
It did not have good ipc when you consider when core, but if you consider a bulldozer core as two cores like me, it actually does have decent ipc.

>> No.69504492

>>69496570
>an FX core is technically the two integer cores in that module
/thread

>> No.69504641

>>69490864
I wish I could get half of what I paid for my fx8350

>> No.69504964

>>69495426
>8350s were strong as fuck if you didn't use them
Kek

>> No.69504989

>>69504964
Not sure people even would say that, an 8350 with a 7970 is a damn fine 1080p gaming experience, I would've gladly paid the 1000+ dollar for this stuff 6 years ago because my computer would still be whipping the consoles ass.

>> No.69504995

>>69504989
Not sure why*

>> No.69504999

>>69496238
>FX really doesn't give a shit about how fast your memory is, I don't know what it's deal is but I'll take it.
Well your NB frequency is pretty low, people OC it to like 2500-2600 and maybe then you'd see more difference with faster memory

>> No.69505022

>>69490864
people trying to claim cores are not cores is whats embarrasing. you seriously have no argument whatsoever. embarrasing how tech illiterate you are

>> No.69505037

>>69505022
I've taken a glimpse into the mind of an amd retard and it's not pretty.

>> No.69505046

>>69490913
i was well aware my 8-core fx only has 4 FPUs when i bought it and i literally dont care. the gullible ones are people that think that amd uses fake cores.

>> No.69505052

>>69504999
Checked, but even if I crank it to a 225 bus with 2475MHz memory controller speed, it doesn't score any faster than with it at 2200MHz stock. I now people got huge gains with the bus overclocking on the Phenom II, but AMD FX does not work that way. I feel like having 16GB of ram made it faster than overclocking the ram ever could.

>> No.69505064

>>69491113
so its an 8 core 4 fpu cpu. notice the part where theres 8 cores. sure theres only 4 fpus, but, theres 8 cores. fuck youre dumb

>> No.69505079

>>69496285
>Yes it is indeed AMD's take on HT
>>69496315
>Marking it as a quad module, 8 thread CPu would've saved everyone the trouble
But that simply wouldn't be true either, it would be perceived as something equal to HT while it's much more than HT. All HT is about is just some extra registers to do away with context switching penalty of fully-software multithreading.
Bulldozer modules has actual doubled execution units with just some stuff shared, while HT shares everything aside from some on-chip memory to store data of the thread.
Bulldozer module can run two threads simulatenously while HT physically CAN NOT do that and only switches between two threads, just does this switch faster than if done fully by software.

>> No.69505099

>>69505046
Do you really think it limits to FX case you amd retard?
How about
-Polaris with 8gb ram it can't effectively utilize because the gpu shits itself at higher resolutions anywaY
-Ryzen 2700x which comes OC'ed to the hilt from the box to advertise 4.35 ghz turbo which diminished it you actually wan't to achieve it across all cores
-"""12nm"""" which for all intents and purposes is worse than intels 14nm, featuring larger transistors
-naming their chipsets with numbers bigger than intels to create an impression they're superior

AMD as a company is a perfect match for people deep into THE dunning-kruger curve - dumb as brick mouthbreathers who believe they have a grasp on technology.

>> No.69505199

I wish they had released a Streamroller or Excavator AM3+ FX CPUs. Any idea why they never did that? AM3+ was kinda shortlived considering Bulldozer architecture still had 2 more iterations and Excavator supposedly had significant IPC improvements.

I read somewhere that the 28nm process they moved to at the time sucked and couldn't clock as high as the 32nm, so all the IPC improvements would be nullified by a lower clock anyway. But was it really the case?

>> No.69505214

>>69493225
>AMD
>Underdog
>what are the rights to x64 computing and console graphics

>> No.69505241

>>69505214
>what is market share
>what is assets value
>what is market capitalization
>what is ebitda
>what is sales

>> No.69505327

>>69505199
The FX 8370 is probably the best update to the FX series because it's a binned chip that clocks higher, my chip can only do 4.5GHz at 1.4v but an 8370 would for sure do 4.6 or more, it's turbo is 100mhz higher anyway.

>> No.69505560

>>69505327
Yeah I know, I have an 8350 and am fine with that for now, just wish they had released an Excavator FX in like 2014/15 :M Then the history would maybe remember Bulldozer as something a little better

>> No.69505587

>>69505560
People are never gonna remember FX as something really good for the money because multicore doesn't matter™.

>> No.69506597

>>69505587
Nah we gotta admit AMD just fucked up with that one. A shrunk Phenom III with an updated instruction set would still be meh but it would have raped Bulldozer in all orifices, but AMD chose to meme us instead.

>> No.69506719

>>69506597
>shrunk Phenom III
Phenom II didn't clock that well, 4.2 was the most it would do on a 6 ore and it needed alot of voltage, an FX 8350 can get the same performance at 4.4 with way less voltage.

>> No.69506800

>>69491474
I bought a fx4130 brand new for $30 and its running devuan good right now

>> No.69506837

>>69490864
What the shit, a CPU core is a CPU core, performance notwithstanding.
What is the judge trying to argue?

>> No.69506844

>>69506800
>I bought a fx4130 brand new for $30 and its running devuan good right now
Ew you could at least upgrade it to some Piledriver baby

>> No.69506863

>>69506837
I'm just amazed (((people))) would even start shit about AMD FX chips, considering how old and irrelevant they are for budget build use.

>>69506844
This, a brand new 8350 is 69.99 now. can't recommend anything else if you own an AM3+ board, keep it somewhat relevant for a little while.

>> No.69506883

>>69506597
>>69506719
It seems that AMD did ponder shrinking Phenoms for a while, and the result was Socket FM1 - a very underwhelming shit platform with 32nm shrunk K10 CPUs with shitty low clocks and no L3 cache

>> No.69506890

>>69491183
Never has an Intel shill been so BTFO.

>> No.69506901

>>69506863
>considering how old and irrelevant they are
That's not a valid point considering AMD has released new Excavator based APUs as recently as late 2018

>> No.69506927

>>69506901
Yeah I know they still make Bristol ridge which is FX, but when it comes to building budget rigs, the 4c8t 2400G does dominate the FX-8350, even though I like the 8350

>> No.69506985

>>69505037
> it has 8 cores and 4 fpus therefore it only has 4 cores, even tho it has 8, it only has 4
u r a retard

>> No.69507019

>>69505099
>the gpu has 8gb but since the gpu is slow you shouldnt say it has 8gb even tho it does

> the cpu can clock higher on fewer cores whuch is useful in some cases but im an autist and that option triggers my autism so im going to project my autism now

plz just leave

>> No.69507053

>>69491969
holy shit someone should sue them

>> No.69507085

>>69507053
Pentiums didn't even have L2 cache, how can one consider those CPUs?

>> No.69507494

>>69505199
Excavator could've been a decent upgrade. It was already better than the Piledrivers despite the reduced cache.

>> No.69507625

>>69507494
They should have just dropped it mid development and stuck to phenom or something.

Or focused 100% on APUs.

>> No.69507701

>>69507625
>Or focused 100% on APUs.
They did for some time.
AMD was still competitive in laptops compared to Intel on a highly inferior process node for some time. It wasn't until Intel dropped to 14nm and had a 2.5x node advantage that they couldn't compete anymore.
Also they had later excavator based cores only in APUs that were significantly superior to Piledriver. As far as I'm aware, they did have slightly higher IPC than Phenom II, whereas Bulldozer and i think Piledriver as well were lower, and they were only really held back by the 28nm/32nm node.

>> No.69507770

>>69507701
They still produced a lot of FX chips and a few server chips at that time though, if they didn't spend money on that they may have even been able to get Ryzen out sooner.

>> No.69507798

>>69507770
Ryzen took what.. 4.5 years? That's pretty normal for a new arch and I think a lot of the wait was on GloFo's 14nm. They still had wafer supply agreements at the time, afaik. And wafer supply agreements are also a lot why they kept making bulldozer and Opterons even though only GIS and some other niche industries used them.

>> No.69507828

>>69490864
inlel shills are getting desperate

>> No.69507870
File: 26 KB, 728x380, intel-meltdown-spectre-patch.png [View same] [iqdb] [saucenao] [google] [report]
69507870

Don't mind me OP, I'll just be browsing without backdoors. IDK about you but I don't want the Kremlin knowing what pornhub videos I watch.

>> No.69508350

>>69491192
This. Corporate dick sucking and fanboyism has always made me cringe.
I know Intel sucks but I still bought a 3570k 7 years ago because that shit was a good product and not because I like Intel.

>> No.69508478

>>69507701
>>69507770
>>69507798
As for lack of Steamroller and Excavator high end CPUs, 28nm was really shit and didn't clock very high. Any IPC improvements were lost on lower clocks.
Even if we assumed they had the basic Zen idea at the time already, had they released it on 28nm it would've been a disaster.

>> No.69508783

>>69501655
You correctly identified computational modules (the other retard counted l2 cache blocks) but you failed to see two mirror image ALUs and a side mounted FPU in each of them.

>> No.69508860

>>69508478
Ya, that's what I was saying.
They could still have low clocks for lower power consumption in laptops to a degree where high performance isn't expected.
But suppose they couldn't compete with the 6700k with another 180W part even though Intel is happy to compete with the 2700X with a 180W+ part.
So even though they already had a higher IPC core design they could have made as an 8 core desktop part, it just wasn't worth it for the small market.

>>69508350
Yeah. I've bought close to a dozen CPUs over my lifetime, and only 2 have been intel. They were the best at the budget, an E8400 and i5-2500k, at those times.
But although I could have technically afforded and upgrade to a 4790k or 6700k, I didn't get those because it was hard as fuck to justify so much more money for such a small improvement unlikely how easy it was to justify upgrading to Ryzen.

>> No.69508940

>>69507870
the only reason I regret upgrading

>> No.69508995
File: 132 KB, 838x734, 46ec1cb2f9a197c530deb8185fe8ae61.jpg [View same] [iqdb] [saucenao] [google] [report]
69508995

>>69503992
It did not have good IPC. It was not designed to have good IPC. It was designed as a speed demon. It was built to be narrow and high clocked. However Globalfoundries did not deliver on 20nm and they never got the clocks they wanted. There is a reason this architecture holds the CPU clock speed world record.
>>69504279
Are you actually retarded?

>> No.69509623

>>69503992
>>69508995
The int IPC was high. The fp IPC sucked.

Same way my 1600X was like 4x faster than a $3000 xeon in a lot of tests. There are many varied instructions that a CPU can be better or worse at.

>> No.69509681

>>69499223

Thanks for the upvote, fellow redditor!

>> No.69509828

>>69503867
But intel does the exact same thing, yet nobody seems to care.

>t. 8 """core"""" cpu with only 4 physical cores

>> No.69509877

>>69509623
The FX cores do not have high IPC. They were never designed to have high IPC. They are coincidentally good at VERY specific task due to there primitive nature but this isn't by design and isn't particularly useful.

>> No.69509909

>>69504172
Someone send this to the judge/AMD's lawyers

>> No.69509910

>>69509877
lmfao i love how you're trying to argue without clearly knowing the specifics of how IPC is and can be measured. You think it's just some generic average or based on one specific application because you're retarded and can't read through the context of what others have posted to know that's clearly not the case.

>> No.69509976

>>69495986
I wouldn't consider 1.4v high at all. I was pushing 1.45v through my FX-6300 just to keep it stable at 4.5GHz.

>> No.69510035
File: 90 KB, 477x616, 1456647016105.jpg [View same] [iqdb] [saucenao] [google] [report]
69510035

>>69509910
FX cores are narrow AF and due to there extremely long pipeline are prone to cache miss. This wasn't a mistake that FX has shit IPC. This was by design. They were expecting 5+ghz out of Globalfoundries at FAR lower power draw.This is This is a FACT about the AMD's FX architecture. It is a FACT that it was designed to be clocked much higher and the entire architecture was never designed for high IPC. Read a fucking book and learn something about microarchitectures you absolute ape.

http://www.lighterra.com/papers/modernmicroprocessors/

>> No.69510156

>>69491183
I'll make the screencap

>> No.69511127

>>69490864
AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

AMD TO DECLARE BANKRUPTCY
>AMD TO DECLARE BANKRUPTCY

TO BE PURCHASED BY CHINA

>> No.69511266

>>69509828
*4 modules with 2 cores per module

>> No.69511304

>>69511266
I don't know how much more we can clarify that it is eight cores, /g/ is fucking retarded, these are the same people who deal with Intel's bullshit because they watched a dick suckem video that said it was good.

>> No.69512021

>>69511266
>>69511304
>2 cores per module
Absolutely false.

It is 4 modules (4 physical cores) with hyperthreading (which turns it into 8 LOGICAL cores ON PAPER).
In reality however, it is still just 4 physical cores with some meme feature that makes them take instructions a little faster.

They are advertising 8 cores, your OS will show you 8 cores (if you have hyperthreading enabled in bios) BUT YOU ARE STILL GETTING JUST 4 PHYSICAL CORES.

>> No.69512039

>>69490864
Someone explain this to me it's literally my first thread on here I've ever posted in.

>> No.69512056

>>69512021
Zen's AMD's first microarch to support SMT. Look at die shots, there are 8 cores and you cannot deny it.

>> No.69512063

where are you niggas with the screencap? jesus. i'm lazy as fugg

>> No.69512074

>>69512056
I'm talking about how Intel is doing the exact same deceptive marketing with their Intel chips but no-one seems to care about that.

>> No.69512106

>>69512021
AMD knows that an FX "Core" is a module, and that it operates like a 4c8t chip, but technically they are physically 8 cores that need to be leveraged by software to get the thing rocking so they say eight core processor, it's not like an i7 where you can turn off the hyperthreading, you can only disable modules, and have a 2, 4, 6, or 8 core chip.

>> No.69512152

>>69509909
Which will get them absolutely nothing.
It doesn't matter what the 386's core was. It's what the market now considers one to be.
Both Intel and AMD have marketed a core to be a full core for the last 20 years, not Faildozer's abortion.

You can't just decide on your own to reduce the definition.

>Athlon's core is X
>Athlon 64's core is X
>Phenom 1's core is X
>Phenom II's core is X
>Bulldozer's core is 0.5X AND LOOK WE HAVE TWICE AS MANY CORES AS INTEL!
>Piledriver's core is 0.5X AND LOOK WE HAVE TWICE AS MANY CORES AS INTEL!
>Ryzen's core is X

That's fraud.

>> No.69512192

>>69512152
And advertising something as 8 cores when it only has 4 is not fraud?
>but it's ok when intel does it

>> No.69512197

>>69491056
Those are the cache you dingus

>> No.69512200

>>69512152
But that's not what happened. There are eight cores in what they claim to be eight core processors.

>> No.69512230

>>69491183
Not so good old days reading messages like "program.exe cannot start because couldnt find a math co-processor."

>> No.69512257

>>69491056
Retard, those are not the cores jesus christ

>> No.69512270

>>69512257
Kek, intelfags talking shit about an architecture they know nothing about.

>> No.69512273

>>69491056
nigger you are literally pointing to l2 cache blocks.

hm

google says it's an fx 6300,

but the amd website says it's a 6 core processor..

wtf is going on I thought I understood what the issue was, IE AMD marketing multiplexing hybrid cores (ie piping instructions from two separate L2 memory blocks into one core for better efficiency) as being twice the number of physical cores, but

how the fuck do they call that a 6 core processor? do they have 2 legit cores and 2 chinked cores?

SOMEONE EXPLAIN IM HAVING A CRISIS HERE

not that I'd ever buy AMD but WHAT IS GOING ON

>> No.69512321

>>69512273
Okay, you have three modules, they have 2 cores each, if you use 3 cores or less, the ipc of it is that is a high clocked core 2 duo, if you have 6 cores working, the ipc is almost 20% worse, but you have double the cores, so it ends up being faster than if only 3 cores and 3 modules were being used.

>> No.69512330

>>69512321
>it is that is
is that of*

>> No.69512452

>>69490913
This is the only instance where this has happened, just how misleading are their Ryzen CPUs, now?

>> No.69512466

>>69512452
Even FX isn't misleading, it's 8 physical CPUs that are setup in a way where it behaves like a 4c8t chip, it's the non jewiest chip ever made.

>> No.69512479

>>69512321
I've been trying to make sense of your post.

So there's obviously 4 cores. I define a core as a block that contains one ALU and all the supporting infrastructure for that specific ALU.

What it sounds like to me what you're saying is that AMD expects at least one core on each chip to underperform or fail entirely, and then they disable it, leaving it with three working cores, or "modules" as you call them.

I don't understand why the IPC goes down if you enable another physical core, unless you have shitty power and/or thermal management.

simply because you multiplex two separate instruction pipelines into the same core doesn't mean you have more physical cores? but that's not the point.

I was just wondering why they were binning a 4 core chip as a 3 core chip and selling it as a 6 core chip.

I mean if they sell hyperthreaded cores as a standard feature they should have capitalized on that and marketed it as such, to claim an edge over intel.

>> No.69512496

>>69512479
The ipc drops like a rock because you have two individual core in a module with their own ALUs and all that god stuff, but they share a fetcher and decoder, it's cramming instructions through a single entry and exit point so it creates a bottleneck, FX 8350 with 8 core is still 70% faster than an FX-8350 with 4 cores and 4 modules running, and AMD can prove this via benchmarks.

>> No.69512532

>>69512496
you keep contradicting yourself

or are you just trolling?

>> No.69512553
File: 101 KB, 778x710, 4.4GHz Bench.png [View same] [iqdb] [saucenao] [google] [report]
69512553

>>69512532
I'm dead motherfucking serious, the 8 cores and 4 modules can never be double the speed of 4 cores and 4 modules because of the aforementioned bottlenecks, but the cores are fucking there.

>> No.69512632

misleading? yeah. But i wouldnt call it false advertising. a core is a core. recourses that a core utilizes are...you guessed it, recourses that the core uses. You see how those are two different things?

>> No.69512768
File: 112 KB, 840x860, AMD_Bulldozer_block_diagram_(CPU_core_bloack).png [View same] [iqdb] [saucenao] [google] [report]
69512768

>>69512553

IDK man

an ALU with two parallel integer clusters is still just one ALU IMO

>> No.69513210
File: 52 KB, 602x641, 7zip.png [View same] [iqdb] [saucenao] [google] [report]
69513210

if it's only "really" 4 cores, why does it scale so well up to 8 threads? remember: it doesn't have hyperthreading

>> No.69513237

>>69513210
It scales to about 1.7x when it comes 1 core in a module being used vs 2 cores in a module being used because the core is physically there.

>> No.69513237,1 [INTERNAL] 

>>69491183
kill yourself retarded samefagging subhuman

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
Captcha
Action