Intel has been absolutely blown the fuck out of the CPU mark

File: ol4pzlk5vsez.png (374KB, 877x963px) Image search: [Google]

374KB, 877x963px

>>61833033
Ready to OC?

>>61833033
reported for anti-antisemitism

File: image (1).jpg (52KB, 484x361px) Image search: [Google]

52KB, 484x361px

>>61833033
>It's dead, burry it.
It's dead, curry it.

arstechnica is a joke nowadays

>>61833285
t. intel

File: badgoy.gif (11KB, 171x200px) Image search: [Google]

11KB, 171x200px

>>61833033
oy veyy this article is hate speech!

delete this now before my boys shut it down

File: 40-Cute-Otter-Pictures-20.jpg (116KB, 600x450px) Image search: [Google]

116KB, 600x450px

>>61833033
>go to compare CPUs
>benchmarks are just a bunch of game performance tests
>barely any industry benchmarks

File: 1472778992834.jpg (35KB, 325x346px) Image search: [Google]

35KB, 325x346px

>>61833033
OY VEY shut it down

File: 1497565895125.jpg (112KB, 691x771px) Image search: [Google]

112KB, 691x771px

>>61833033
FUCK YOU GOY

File: 1495358050273.png (495KB, 1070x601px) Image search: [Google]

495KB, 1070x601px

>>61833033
DELET

posting crying wojaks needs to be a bannable offense, it's so tiresome with every thread devolving into redditesque post chains of crying wojaks from both sides

File: 1502055755975.png (343KB, 672x794px) Image search: [Google]

343KB, 672x794px

>>61834894
UMA

File: 1502052999915.png (145KB, 633x900px) Image search: [Google]

145KB, 633x900px

Lower FPS in games
Segmentation fault

File: i9.jpg (434KB, 1906x1536px) Image search: [Google]

434KB, 1906x1536px

>>61834894
Agreed. Smiling Wojaks are better.

>>61834920
MACACO

>>61834920
http://phoronix.com/scan.php?page=news_item&px=Ryzen-Segv-Response

>AMD was also able to confirm this issue is not present with AMD Epyc or AMD ThreadRipper processors, but isolated to these early Ryzen processors under Linux.

>>61834973
Fake news, goy

|
|>
|
|3
|
|

anand tech's review was surprisingly bad toward AMD. If you want to game, don't by the $1000k amd cpu and the crazy motherboard, buy the $329 intel 7740x.

Even in the Chrome compilation tests, where I expected Threadripper to crush all, it lost badly to the (same priced) 7900x.

So amd fanboys, this is your high water mark. The i9 stuff is going to dominate/destroy Threadripper. Yeah, you are going to have to pay a lot of $ for it, but these are HEDT systems anyway.

Note that I write the above personally considering a threadripper system for its ECC workstation support.

How did they measure sea depth?

>>61836944
Anandtech's article is pretty crap to be honest. Almost every test is some single thread focused web/office/normie photo editing shit, others are old-ass benchmarks (PCMark 8?), etc. Half the shit on here isn't even relevant to the target market of this CPU.

>>61837029
They didn't even post any Ashes of the singularity benches. That's how yiou know they're intel shills.

>>61837029

Then who is the proper customer for this cpu, and what would be better benchmarks?

I look at the selection of motherboards available at launch, and I only see enthusiast targeted stuff.

The most workstation oriented task (beyond the video rendering stuff) is is their hand rolled 3D Particle Movement v1 benchmark, that is probably better thread optimized (Threadripper biased) than what a normal scientific / engineering guy would be running as he develops his codes.

I have some niche commercial codes at work that might work very well on Threadripper (PIV processing).

>>61837070
ok, AoTS

File: 10 cores uses more power lol.png (486KB, 1920x1080px) Image search: [Google]

486KB, 1920x1080px

>>61836944
You have to spend $400 more to get to the point where Intel wins by about 10%. After that, it's $700 for 16% and $1000 for 22%. Based on some estimations I've made, YMMV.

Also, 7900X power draw is already ahead of TR 16 core. How much crazier do you think this is going to be for the 12+ cores?

>>61837206
3D rendering and video encoding, mostly. Who cares if i9 wins in some office suite benchmark that's tuned for an officelady i3 with 8 GB of RAM?

Sir kindly do the needful and delete this

>>61833033
D-d-doesn't matter, it loses in mobilemark and gets lower fps than a pentium g4560 in Skyrim at 320 x 240 resolution.

File: 1436819647086.jpg (11KB, 242x250px) Image search: [Google]

11KB, 242x250px

I'm sick of AMD ripping me another asshole

REEEEEEEEEEEEEEEEEEEE
Even the housefire isn't far enough, HEDT of our 2 sku's BTFO
NO this can't be happening !!!
Not again
NOOOOOOOOOOOOOOOOOOOOOOOOOOO!!!!!

Coffee lake has to make this domation end and end fast.

>>61833033
>Rectangular CPU

Absolutely garbage

Next they'll do slotted CPUs

AMD a shit

>>61837219
>1920X consumes less power than 7800X with half the cores
>it also completes tasks far more quickly, making the delta in total energy used significantly greater
JUST

Old AMD Fanboy here - Athlon era. I've been running Intels for the past decade because AMD simply hasn't kept pace. I hope this new gen of processors is able to steal the crown. Unfortunately I just built a new rig a few months ago so it will be a while before I'm ready to upgrade again.

>>61836944
Their chromium compile test is pretty damn suspicious, Ars has threadripper beating the 7900x in chrome compiles so I dunno wtf anand is doing differently.

>>61833033
>burry it
bury it

>>61838220
>so I dunno wtf anand is doing differently.
Pretty much all their tests put either the i9 or i7-7740K in the lead. You do the math.

Oy!

>>61833033
>$1000 processors
>representing any major portion of the cpu market
please
most people are gonna be buying mid-range i5s or ryzen 5s

File: 1501855657747.png (208KB, 882x758px) Image search: [Google]

208KB, 882x758px

NO THIS CAN'T BE HAPPENING!!! I JUST SOLD MY LIFE INSURANCE TO BUY AN I9!!!

>>61836944
7740X is shit-tier. You are getting a platform that you cannot take full advantage. It is the same silicon as the 7700K with a mild bump in turbo clock (7700K can easily bridge the difference with a very modest OC).

Get a fucking 7700K a run-of-a-mill Z270 board. You save $200-300 bucks over a 7740X+X299 that you can't even fully utilize.

File: 1502015079074.png (187KB, 568x612px) Image search: [Google]

187KB, 568x612px

>>61834894
Go back to your safe space

>>61833093
Sure but doesn't going too high on voltage shorten the silicon lifespan too, even though sufficiently cooled? I would prefer to stay below 1.4v

I feel like upgrading now. How often do you guys upgrade?

>>61833033
Oy vey delet dis schmuck

>>61842761
thread ripper are some of the top binned amd cpus, hitting 4.0+ at 1.375 or lower should be near cake walk for it.

>>61833033
Will I be able encode memes at 60fps? Not one reviewer bother to run vp9 encoding tests.

File: [Leopard-Raws] Battle Girl High School - 01 RAW (BS11 1280x720 x264 AAC).mp4_snapshot_14.12_[2017.07.05_19.58.55].jpg (33KB, 360x360px) Image search: [Google]

33KB, 360x360px

Intel status:
Ripped to threads

>>61838220
>I dunno wtf anand is doing differently
Different Chromium versions and different compilers, to start with.

>>61845125
What a day to be alive.

File: 1494363947188.jpg (66KB, 685x741px) Image search: [Google]

66KB, 685x741px

>>61845125
STOOOOOOOOOOOOOOOOOPPPPPPPPPPPPPPP IM GOING TYO FUCKING KILL MYSELF

File: 1501209230534.png (103KB, 311x300px) Image search: [Google]

103KB, 311x300px

>>61845125

>>61833033
im done
i cant take this forum anymore

File: 1426994438235.png (760KB, 853x506px) Image search: [Google]

760KB, 853x506px

There is not an intel chip priced $80-$1,600 worth buying

>>61842761
But how much lifespan is 1.4v shortening? 1-2 years @ 24/7? Maybe. But by then you'll be replacing the cpu long before it reaches the end of its lifespan.

>>61845125
Wait for the (((update))) hehehe

>>61838220
You do realize same company now owns both Tom's Hardware and Anandtech, no?

File: 1502335120309.jpg (25KB, 480x480px) Image search: [Google]

25KB, 480x480px

>>61833131

kek'd

>>61845125

Quite possibly the greatest chart I've seen in 5 years of /g/.

I cried a little when I saw this.

File: EDFFE5D9-F55D-4278-9429-5329621B8BE6-16271-00000ECE2350761A.jpg (68KB, 768x960px) Image search: [Google]

68KB, 768x960px

>>61833033
Funny how this was relevant in 2011 and now it's relevant again.

>>61845712
With the proper bios, threadripper will likely run with ecc memory. years down the road you could still run it as a home server or render box. Better not wear that silicon out too soon.

>>61845994
t. destroyed poo in joo

>>61844779
Because reviewers are shit at actually testing technology sanely. Who the fuck gives a shit about random gaming benchmarks for a new CPU?

>>61845125
>$999
NOOOOOOOOOOOO WHAT WERE THE GOYIM THINKING???
DELID THIS

>>61845994
>Intel
>New chips increasing anything but laughability

Intel released worst platform to date x299 with worst CPUs to date based on Nahelem arch.

>>61845994
>2011
>throw more half-cores
>2017
>throw more glued together cores

>>61845994
Now it's the oposite. Amd Improved their IPC by 40%, kept the same core counts and intel is scrambling to get more cores leading to devastating thermal effects.

>>61845994
The biggest difference between now and then is that it's no longer 4c/4t or 6c/8t but 16/32

>>61846737
>5 years to improve 40% over that abortion bulldozer
>somehow it's a feat to brag about

>>61846737
AMD did well, but they are still behind on IPC and clockspeed.

They are competitive though because they offer their products much cheaper, core for core Intel still has the faster products though.

File: CPUZ.jpg (76KB, 417x413px) Image search: [Google]

76KB, 417x413px

>>61846756
Well, intel took 10.

>>61846763
>behind IPC
Ryzen has more IPC than Nahelem ever has had.
What Ryzen is lacking is clockspeeds.

>>61846772
Except core 2 was good on release unlike faildozer.

>>61846776
Ryzen IPC is at about Broadwell level, not skylake/kabylake level

>>61846796
Yes, 10 years ago it was great.

>>61846797
Broadwell is about 2% lower than kaby
Fuck off.

File: Review-chart-template-2017-final.001-640x480.png (51KB, 640x480px) Image search: [Google]

51KB, 640x480px

Crispy

Can someone tell me why would one buy the most expensive Threadripper at 1K when the cheapest EPYC is $500 and you get 128 PCIE lanes instead of 64?

>>61846823
More like 8%, eat shit

>>61846843
more like .2% when it's actually productive programs

Kabylake is an abortion

>>61846860
>more like .2% when it's actually productive programs

That's not how IPC works you tech illiterate mong

>>61846879
it kinda is, because of un-unified registers, any Nahelem can't have full instructions only two, so others need to be broken down to half width or wait until full registers free up, because productive software uses quad instructions, poor Nahelem can't keep up.

File: Untitled.jpg (288KB, 1920x1080px) Image search: [Google]

288KB, 1920x1080px

>>61846797
broadwell to kaby is 2-3 % slower
sky to kaby is 0% if negative some times
amd stated that 40% improvement was on sandy bridge but on launch they stated 52% literally bringing zen on kaby/sky level and we already have seen that zen beats often kaby in various benches even on games nowdays 1600x is on par with 7700k

>>61846842
anyone?

>>61846901
>it kinda is
No it's not, stop talking horseshit and making yourself look like more and more of an idiot

>>61846903
>talking about IPC
>he shows gaming benchmarks
Tech illiterate cunts please leave

>>61846910
>Ignore the whole post
Aaand your gone

>>61846843
It does you gaymer child
http://www.agner.org/optimize/blog/read.php?i=838

>>61846905
because you dont account for the overly expensive motherboards that needs its own super duper PSU and the ecc memory and and and...
the problem isnt the 500 epyc BUT
i9 currently costs 2k grand at this money you buy the 7550p a 32 core monster there is literally no reason to buy anything intel for the first time since k7-k8 era

>>61846928
>talk bullshit out of your ass
aand you're gone

>>61846922
typical from intel shills not to comment on the content of the post

>>61846951
When the contents are horseshit, you don't have to respond to them.

We were talking about IPC and he literally suggested it varies depending on the load. If you show that level of stupidity and ignorance about the topic at hand, you aren't worth responding to

>>61846941
>overly expensive motherboards that needs its own super duper PSU and the ecc memory and and and...
Dont they use the same socket?
As for ECC memory its fine im using those since LGA 775 days and i have two 1200W PSU that do nothing and gather dust.

File: 1478227130-4.0.png (73KB, 600x515px) Image search: [Google]

73KB, 600x515px

>>61846965
horseshit reviews incoming
https://www.pcper.com/reviews/Processors/Intel-Core-i7-7700K-Review-Kaby-Lake-and-14nm/Clock-Clock-Kaby-Lake-Skylake-Broad

http://www.anandtech.com/show/10959/intel-launches-7th-generation-kaby-lake-i7-7700k-i5-7600k-i3-7350k/8

>>61846972
yeah but i doubt amd will just enable epyc support on hedt(im sure its just a bios hack but i wont really support my theory 100%) given the price they simply covered every single possible solution leaving the true high end pricing for intel on hedt and on that price well epyc is far more sexier than a housefire

>>61846999
So where is the IPC of Ryzen in here? You know, the topic at hand?

>>61847030
let me guess you were on comma for the past 6 months you havent seen a single benchmark right?

>>61847026
you are not making sense,I wasnt talking about intel I am talking why would one get 1K TR over $500 EPYC.

>>61847046
Again, where are the benchmarks of Ryzen IPC?

>>61847053
because the epyc platform is far more expensive

>>61847070
In what sense?
It can be used on the same boards.

>>61847082
no it cant just unless you think amd will just hack their server sales which they wont

>>61833033
>performs more or less the same in video games
I'm glad I'm never going to have to update my CPU.

>>61833033
https://www.youtube.com/watch?v=osSMJRyxG0k
intel needs to die or get its shit together and stop breaking the law

>>61847054
if you wanna for us to take you seriously i suggest you wake up and remove your blue bubble there are TONS of benches already since zen launched
if you cant find any then RMA your self

>>61847087
oh so you have no idea then,got it.

>>61833093
>OCing threadripper
If you like 300W+ power drawn.

File: untitled-1.png (30KB, 687x788px) Image search: [Google]

30KB, 687x788px

>>61847176
no i actually do have a very good idea i was hoping that you had more iq than a macaco
but i was wrong

>>61845125
FUCKING DELETE THIS RIGHT NOW

>>61846965
You dumb fuck do you seriously think that those tech reviewers truly calculated "IPC" on an instruction level to find that 8% figure? They merely set the same clock speeds and equal core/thread counts, run different programs and average the result to find a performance delta disregarding any codepath differences etc, so yes, testing with a different set programs will result in a different "IPC". If you want a proper instruction level calculation refer to Agner Fog's detailed study

http://www.agner.org/optimize/microarchitecture.pdf

As he remarks
>The Ryzen processor put AMD back in the game after they had lagged behind Intel for
several years. The Zen kernel has a throughput of five instructions per clock cycle, which is the record so far

There are different factors when it comes to performance, different architectures have different advantages and bottlenecks and higher instructions per cycle may not give higher performance if you hit a bottleneck or nonoptimized code

>>61847202
>ask about EPYC and TR
>get Intel reacher
UMA DELICIA jej

>>61845125
OYYY VEYYY REPORTED FOR ANTISEMITISM
DELET PLS

>>61847087
>amd makes profit on motherboards...

you're not that bright are you?

>>61847272
Delete yourself, cretin.

>>61846756
>>5 years to improve 40% over that abortion bulldozer
It's not 40% over bulldozer, but excavator.

>>61846965
>he literally suggested it varies depending on the load
Because that's how it works, you fucking gaymen child.

File: AMD_Threadripper_1950X_review_Better_than_Intel_in_almost_every_way_Ars_Technica_UK_-_2017-08-11_11.09.19.png (77KB, 795x601px) Image search: [Google]

77KB, 795x601px

Friendly reminder that when 4K gaming becomes a thing it will be GPU bound and which processor you use will be more about to features than clock speeds.

AMD wins on this front in terms or price, power usage, thermals and unlocked features.

>>61847793
>>amd makes profit on motherboards...
who the fuck said that ?
give me one reason as to why amd would let ANYONE to buy epyc for the TR4 thus elliminating the whole TR lineup?

>>61846772
>half of frequency
>half of sc performance
That's intel for you, 10 years of architecture improvements.

>>61848035
>give me one reason as to why amd would let ANYONE to buy epyc for the TR4 thus elliminating the whole TR lineup?
How does it eliminate the TR lineup when highest clocked Epyc has only 3.2GHz turbo?

>>61848093
did you just seriously answered back talking about a server cpu and got only the clock meme ?

are you retarded?
it has double the cores
double the pcie lanes
no motherboard on tr4 is built around that
and yet here we are discussing such a stupid stuff

>>61848116
>did you just seriously answered back talking about a server cpu and got only the clock meme ?
It's two different use-cases, moron. Threadripper is for workstations where you want many, but highly clocked cores. Epyc is for servers where the per-core performance is less important than the total throughput.

>>61848135
multithreading perf on both cases are what is most important and in both ways epyc demolishes tr4

File: speedy fan.jpg (29KB, 800x800px) Image search: [Google]

29KB, 800x800px

good job AMD for literally blowing the Intel jews out of the water.

>>61847943
Stop spamming this shit, you have 0 idea what you're talking about.

File: 1501099942960.png (20KB, 493x280px) Image search: [Google]

20KB, 493x280px

>>61847054

>>61848760
Right now post the non cherrypicked one as we all know it's at Broadwell level, so it can't possibly be ahead of Kabylake

>>61847194
>OCing X299
If you like 400W+ power drawn.

>>61848772
UMA
M
A

Stay triggered

>>61845994
Amd have memed themselves right up Intel's ass and it's bueatiful

>>61848772
Your fucking idiot .Intel has worse IPC. Intel has better single-threaded performance due higher clocks.

>>61848937
They have better single threaded performance due to higher IPC and higher clocks you dumb fucking cunt

>>61848941
The chart doesn't support your theory, Shlomo.

File: Intel Celeron Slot 1 366 SL376 01.jpg (240KB, 1000x454px) Image search: [Google]

240KB, 1000x454px

>>61838036
Slotted cpus you say?

>>61848419
I understand perfectly well what I am talking about. 'For now' if you want to game at 4K the processor becomes less relevant in terms of clock speeds as most of the load goes onto the GPU. Even at 1440P there is not 'that' much of a difference bwtween a 7700K and AMD's offerings. I know that in the future when GPU's are capable of running 4K at high frame rates then the 'current' line of CPU's will become bottlenecked. It's why some shills are benchmarking at 720P or lower to state one CPU is slower than another as more is thrown onto the CPU.

When 4K gaming is better specced Zen+ will take over from there anyhow.

tl;dr You got baited.

File: Unbenannt.png (49KB, 827x178px) Image search: [Google]

49KB, 827x178px

>>61845000
nice

>>61849097
>never mentions "for now" in original post
>"U got BAYTID!!1!1@1!"

On an actual serious note here, does anyone know if X299 has the same DDR4 RAM training issues as X99 does, and does X399 / threadripper have the same issues?

>>61849191
ram compability on X399 has greatly improved as it uses the same arch as Ryzen , there have been no reported issues do far , GN noted that it could boot 3200 and 3600 off the bat

same goes for X299 afaik , 3200 off the bat

>>61849243
I think that went in a wrong direction
I mean that X99 takes up to 50 seconds to even POST, apparently because of DDR RAM training

I wondered if the same issue persisted in X299 and was introduced into X399, I noticed seemingly similar issues with X299 when I watched a GN review

>>61847194

EK Water Blocks got it to 4.2 across all 32 threads, and it still draws less than 300W collectively. That's unquestionably impressive.

>>61834932
UMA DELICIA!

>>61849260
no , only if you enable ram training it will take 1~3 seconds longer to boot

>>61834373
*delid

>>61846828

>6c/12t more than 7900X
>only a 45W difference

Jesus christ, Zen's efficiency absolutely slaughters Skylake-X. Intel's 12, 14, 16, and 18 core CPUs "may" perform better than the 1950X but the power draw and heat production will be absolutely insane and their boost/OC potential will be utterly garbage. I'd expect to see the 18c chip with the GPU accounted for to probably hit 850-900W.

>>61833033
2017 AMD wins CPU market. so call "pc enthusiasts" praise it like their new god, shits on intel.

10 years of AMD domination later...

AMD pulls an intel past 10 years, same core count same pci express lanes nothing new. "pc enthusiast" crawls back to intel woah woah we need new savior.

mean while at amd and intel HQ they are laughing at so called "pc enthusiast" spending their life fortune on their stupidly over priced cpu they will never use efficiently.

U ALL LOSE INTEL AND AMD FAGS

>>61849327
is that just for X399, or also for X299?

File: 234785639.jpg (31KB, 177x278px) Image search: [Google]

31KB, 177x278px

>>61847202
I wish I could have sex with a 7700K

File: 23895704395.jpg (28KB, 488x424px) Image search: [Google]

28KB, 488x424px

>>61845994
They finally got it to work for them

Guyz I want a Ryzen CPU but Zen2 is supposedly going to blow everything available today including Zen1 out of the water due to lower wattage and 5 ghz stock speeds. As I am a poorfag I don't want to invest too much in something that's going to be obsolete in 1-1.5 years. Should I wait?

>>61846879
Are you trying to say that IPC doesn't differ between programs?

File: 1501563433261.gif (2MB, 570x570px) Image search: [Google]

2MB, 570x570px

>>61849717

File: 1502091921105.png (350KB, 672x794px) Image search: [Google]

350KB, 672x794px

>>61847202
Please delid this

>>61833033
Remember the Bulldozer series OP? Or are you're PTSD pills blocking that memory out?

>>61846922
Why does games not have with IPC to do? If you have two processors running the same game with the same number of cores at the same frequency, and one runs it faster than the other, then by definition that processor has higher IPC.

>>61847273
>The Zen kernel has a throughput of five instructions per clock cycle, which is the record so far
This is true, but something you only care about when personally writing hand-tuned inner loops in assembly. Not irrelevant in a general sense, but irrelevant for this thread. No programs run at five instructions per clock generally. No programs saturate Intel's four, either. The vast majority run at somewhere between 1 and 2 on average.

>>61845125
DELET

>>61849793
Zen 2 is socket compatible with Zen 1 so you can just buy it now and upgrade the CPU when the time comes.

>>61849355
>Intel's 12, 14, 16, and 18 core CPUs "may" perform better than the 1950X
That "may" remains to be seen, at least. The Xeon-like base frequencies of those parts doesn't really set the highest expectations. It shall be interesting to see how well they boost in practice, given the 7900X power-draw numbers.

>>61846972
>Dont they use the same socket?
Not really, TR uses socket TR4, Epyc uses socket SP3. They're supposed to be physically the same, but electrically incompatible.

>>61849793
I am not a poor fag, but the OCed 2500k is still okay for now and I expect to be okay until March 2017 when Zen 2 is released.

Zen 2 looks promising and I hope it will be the 2500k of this generation so that I can run for 6 years too.

File: S_AMD-K7600%20(ES)%20(front).jpg (26KB, 384x192px) Image search: [Google]

26KB, 384x192px

>>61838036
>Next they'll do slotted CPUs
I wish they did. I've kinda missed them.

File: 4389574309.png (263KB, 745x1402px) Image search: [Google]

263KB, 745x1402px

>>61847202
As an Intelcuck I'll admit that AMD has made major leaps in per-core performance since the dark days of the FX 9000 series.

>>61849704
both , as its Impediment by the mobo manufacturer and not AMD/intel

>>61850121
march 2017? ah fuck i got hosed , bought a R5 1600 then

>>61849793
Depends on what you have now

>>61850001
fun fact , the intel "i9" 14~18 cores are rebranded xeons clocked higher

>>61849135
>Threadripper is out 'now'

What part of this did you not understand? You can run 4K on certain games 'now' and if you have SLI/CF easily on some AAA titles (depending on how well SLI/CF is supported since single GPU is always best).

Stop being so butthurt.

>>61849793
If you need a pc now, buy a r3 1200, you can upgrade next year to zen2 with the same motherboard

>>61850489
They're don't even have higher base clocks, just higher turbo.

>>61849957
>between 1 and 2 on average.
likely for compability to Phenom / FX and core2 and core i
id expect programs a few years from now to saturate em

File: 1499806979124.png (107KB, 2250x2324px) Image search: [Google]

107KB, 2250x2324px

>>61845125

>>61850566
It has nothing with "compatibility" to do, or even with optimization options. It's partly about the intrinsic ILP of the algorithms in question (you literally can't get more than one instruction per clock out of a series of dependent instructions), and partly about the CPU's ability to extract that ILP from any particular algorithm (due to its ability to hide cache latencies, branch prediction, &c).

>>61850123

That would be kind of cool. Need to upgrade? Still have a slot open? Just add some new CPU(s) that add to your already installed processing power

File: SquaresAndRectanglesK.jpg.cf.jpg (20KB, 399x214px) Image search: [Google]

20KB, 399x214px

>>61838036

Technically squares are rectangles. Boom roasted. BTFO.

>>61838036
Surprised how common core hasn't renamed rectangles to straightangles for dumbfucks like you.

>>61851805
Thanks, doc.

File: 1502378315806.gif (127KB, 600x600px) Image search: [Google]

127KB, 600x600px

>>61845125
RIPtel

>>61851805
Thanks doc

>>61849912
OP was probably still in his diapers when zambezi launched.

File: ZEN.jpg (143KB, 512x960px) Image search: [Google]

143KB, 512x960px

>>61846828

>>61834920
>Multi-core doesn't matter!
>Price/performance doesn't matter!
>Power usage doesn't matter!
>Temperatures don't matter!
>Soldered dies don't matter!
>Stutters don't matter!
>Streaming doesn't matter!
>Data centers don't matter!
>OEMs don't matter!
>Hyperscalers don't matter!
>Upgradeability doesn't matter!
>*NEW* Locked platform features don't matter!
>*NEW* Synthetic loads don't matter!
>*NEW* Burnt pins don't matter!

>>61836733
Phoronix? Fake news? Are you dumb or so?

File: trash it.jpg (27KB, 948x711px) Image search: [Google]

27KB, 948x711px

>>61846828
>Adding a full GPU load into the results
>>61837219

File: intel fans enjoying skylake-x.jpg (263KB, 1600x851px) Image search: [Google]

263KB, 1600x851px

>>61854840
Wow it's almost like the exact opposite of what ended up happening.

I don't care,i always have and always will buy Intel no matter what.
You can't put a price on brand loyalty

>>61854934
good goy

File: dolphin-1700[1].jpg (128KB, 577x719px) Image search: [Google]

128KB, 577x719px

>>61854934

dis.

If you want to play the emulators such as Dolphin. Intel is the best one for gaming, playing emulators, web, etc... AMD is good for workstation, not home user.

>>61854934
>You can't put a price on brand loyalty
Not for the owners of the brand, perhaps. For you, you only have to look at the MSRPs to determine the exact price to you.

>>61855036
>home users use their pcs solely for gayming

>>61850476
Zen 2 isn't coming out until late 2018 or early 2019. Next year we'll probably have a Ryzen refresh.

>>61855036
Intel is only good for Netflix. For everything else get Ryzen

>>61854888
amd shills really do have text files with pasta ready to go

>>61855113
Fuck off schlomo.

>>61855036
That benchmark is just dickwaving. Almost anything can run Dolphin these days. Even old laptop CPUs.

File: edd.jpg (90KB, 620x670px) Image search: [Google]

90KB, 620x670px

>>61855113
Did you really just post your shitty "rigs" as a signature?

>>61855202
Truth hurts

File: AMlosi.jpg (259KB, 670x768px) Image search: [Google]

259KB, 670x768px

>>61855184

>Ryzen is not good enough..
>then wait for Zen 2
>waiting TM

Well, I am waiting....

>>61854892
FAKE NEWS, GOY

>>61855245
suddenly half that list doesn't matter when vega is concerned
really activates my almonds

>>61855294
>Ryzen is not good enough..
t.JUSTnich

>>61855307
Suddenly GPUs matter when talking about CPUs.
Really activates my deflector shields.

>>61855307
There are at most two things on that list that would actually apply to Vega (power usage and temps).

>>61855334
>yes goy don't look at that failed product result of everything we accuse intel of doing wrong

File: 546875988.png (1MB, 2000x1367px) Image search: [Google]

1MB, 2000x1367px

>>61855386

>>61855386
>>61855402

>>61855358

FRESH COFFEE LAKE LEAKS

>i7-8700k
>35% better ST over 1700x
>15% better MT over 1700x
>soldered IHS

File: 1446667091224.png (8KB, 196x171px) Image search: [Google]

8KB, 196x171px

>>61855671

>>61855671
>soldered IHS
You almost had me for a second there.

>>61845125
KIKERIPPER

File: blewityadid.png (218KB, 500x374px) Image search: [Google]

218KB, 500x374px

>>61855671

>soldered IHS

>>61854934
This is the saddest shit I have ever seen. You are willing to hand your hard earned cash to a Company who'd gladly kill you if meant saving 1$ on the cost of production of a CPU and all for what?

>>61833033
thread ripper is good in the sense that it beats Intel over priced jew high end hardware which you didn't need AMD to release anything to know you are getting your wallet raped.

AMD already released their other cpus which are worse than Intel. How do you plan to take over the CPU market with a CPU that cost as much as an entire high performance rig ?

>>61855402
>AdoredTV
I like his content ;-;

File: my sides.png (173KB, 640x400px) Image search: [Google]

173KB, 640x400px

>>61856090
>How do you plan to take over the CPU market with a CPU that cost as much as an entire high performance rig ?
You mean the 7980XE, right?

>>61856090
>AMD already released their other cpus which are worse than Intel
>worse
>by pricing their octocore slightly lower than Intel's best quad
No one will deny Intel is still best single threaded but the 1600 and 1700 are amazing value and anally rape Intel in anything multi threaded.

>>61856133
Noone said he was bad, but he's not exactly subtle about his bias.
https://www.youtube.com/watch?v=osSMJRyxG0k&t=370s

>>61836944
Spend all the money on a i9 and can't afford to get a AMD TR too? What are you poor?

>>61842761
Then look for a stable OC at a lower voltage if you're worried?

Thank you based shit wreckers.

File: br-intel-shill.jpg (21KB, 267x297px) Image search: [Google]

21KB, 267x297px

>>61845125
EXCLUA

File: 1zyelv4.jpg (48KB, 600x762px) Image search: [Google]

48KB, 600x762px

>>61845125
god damn, it's like a violent massacre that only laster a few minutes

>>61845125

Oy vey!

Delete your post goyim, stop being an anti-semite.

This is the coming of the second shoah!!

>>61859102
Rolling to post my waifus in smug postures.

File: chiyo-3.png (443KB, 560x580px) Image search: [Google]

443KB, 560x580px

>>61859145
Tarnation. I have no frogs to post.

>>61859102
>based tiger mom
top kek

>threadripper silicon is top 5% Ryzen chips
>threadripper is more cores than most people can use
>buy threadripper and disable one of the dies
>Having highest clocking Ryzen with largest heatsink so also lowest temperatures
>get mad respect from the internet

>>61859102
bless me with digits, chip magician kevin sorbo

>>61859262
anyone actually done that yet?

>>61859344
TR has built-in "compatibility" mode that lets you do just that. Disables the other die. It's for software that flat out won't work if it sees 16 cores 32 threads or just explodes in flames because they never thought anyone would run it on something that powerful.

>>61859363
It doesnt disable the other die, it turns off SMT and enables NUMA memory accesses. Naturally, SMT and the NUMA setting can be independently toggled of each other.

>>61859344
No doubt benchwhores and extreme overclockers will.

>buying from a company who can't even properly bin or test their shit.

enjoy your crippling CPU bugs like Machine checks due to compilation.

>>61859401
Oh yes all this delicious FUDge I just can't get enough.

http://phoronix.com/scan.php?page=news_item&px=Ryzen-Segv-Response

>AMD was also able to confirm this issue is not present with AMD Epyc or AMD ThreadRipper processors, but isolated to these early Ryzen processors under Linux.

Kill yourself.

>>61859262
>>61859344
seems a bit dubious, because even if you have more surface area, you're still burning more power on the GMI links that would otherwise just be turned off.

it would be much smarter to just do Silicon Lottery or whatever cherry picking and throw the net saved money from the cheaper chip/platform at a custom loop.

>>61859262
Literally why, though? If you just don't use the cores on one die, they're power-gated off anyway, so you get the same benefit with the flexibility to still use them if you want to.
Also, it would make more thermal sense to spread your workload across the dice.

>>61859433
wow so one particular poor design flaw is not in another iteration of the CPU.

Even shittel has all sorts of fucking erratas and microcode fixes floating around but usually normal retard consumers don't start seeing critical CPU flaws causing userspace programs to "segfault"

buying AMD is a risk.

>>61859395
>To address this problem, Game Mode essentially tells Windows to recognize only 8 of the 16 CPUs in the system. An updated Ryzen Master Utility lets you switch between Game Mode when it’s needed for older games, and Creator Mode when you want all of your CPU cores and more memory bandwidth.

http://www.pcworld.com/article/3214635/components-processors/ryzen-threadripper-review-we-test-amds-monster-cpu.html

>>61859478
http://techreport.com/news/26911/errata-prompts-intel-to-disable-tsx-in-haswell-early-broadwell-cpus

https://www.extremetech.com/computing/244074-intel-atom-c2000-bug-killing-products-multiple-manufacturers

https://hothardware.com/news/critical-flaw-in-intel-skylake-and-kaby-lake-hyperthreading-discovered-requiring-bios-microcode-fix

Eat shit.

>>61859440
What link traffic would you be using if you're basically turning one of the two dice off?

>All of these intel shills.
What happened to my beautiful /g/?

What about poo in loo?
Housefires?
AMDpoors?

My friends what has happened to us?
Don't you all remember?

File: landscape-1433528197-delish-hot-fudge-sundae-poke-cake.jpg (87KB, 1600x800px) Image search: [Google]

87KB, 1600x800px

>>61859401
>>61859478

>>61859499
http://www.legitreviews.com/intel-x99-motherboard-goes-up-in-smoke-for-reasons-unknown_150008

>>61859499
and in most of the cases, they were fixed with microcode because Intel actually knows how to build CPUs

the Atom shit isn't for consumers but stupid retards like Cisco who need to perpetuate nonfree x86 shitware on edge routers and shit like it.

>>61856604
I don't get why people say this. He's against bad practices, not a blind fanboy. He made entire videos about what did amd fuck up.

>>61859169
I have no frogs but I must post

>>61859559
>Every so often people ask me "why are you so down on Intel", so let me explain why.
I'm not even saying it's a bad thing, but he's literally saying that he's biased against Intel.

>>61859513
I assume that the link stays up because otherwise you'd not only cut off not only half your RAM but half your PCIe lanes too.

tfw no undervolted threadripper with passively cooled massive copper heatsink

sigh...

>>61859557
Hahaha this damage control is pathetic.

The only one fixed with microcode was the HT bug. C2000 bug was "repaired" by soldering some shit onto the mobo and will die eventually anyway in a few years. TSX was disabled permanently. Also the exploding X99 mobo was not really "fixable" as it exploded already.

>>61859627
Which would make it equal to the 1800X in memory channels and still have more PCI-E lanes than it, which sounds in line with what >>61859262 wanted to do.

Intel focuses on mobile/laptop tech anyways, it's what they are good at

Core m shit is sneaky m actually really good now

>>61859664
Can't wait to see what Ryzen APUs will do to that.

>>61859655
yeah but no reviews of TR's ghetto mode made any mention of cutting off PCIe etc.

>>61859634
It's not damage control. I'm more than aware of the various faults in Intel hardware.

the difference is that Intel has a significantly LARGER QA budget than AMD and this Ryzen gcc bullshit is pretty much proof that AMD's QA is worse than intel.

something like building software seems pretty standard usage scenario and amd fucked up.

>>61859677
Consume tons of power and so no battery life

Intel is good at low power with high performance, but for desktop that doesn't matter nearly as much

File: lisa-monkey.png (647KB, 605x603px) Image search: [Google]

647KB, 605x603px

>>61833033
(UMA) Delicia mode

>>61859690
Yes, well, the point there being that if you just don't use that part of the memory or those PCI-E lanes, they're effectively disabled, and there's no traffic to them going over the IF links.

File: wat-2.png (362KB, 700x700px) Image search: [Google]

362KB, 700x700px

>>61859697
>ryzen
>bad efficiency

>>61859711
Not that but they just can't into mobile processor

>>61833033
That fucking beast of a processor for only $550?
Holy shit.

>>61859720
The Jaguar cores are pretty swell from what I've heard.

>>61859691
>Intel has a significantly LARGER QA budget
Isn't this proof they're more of a bunch of fuck-ups if they have even more high profile defects?

>>61859720
We'll see when Raven Ridge launches later this year. I fully expect the power efficiency of Zen to shine through. It's made on a low power node to begin with.

>>61859691
>something like building software seems pretty standard usage scenario and amd fucked up.

>certainly not the workloads most Linux users will be firing off on a frequent basis unless intentionally running scripts like ryzen-test/kill-ryzen.
>As I've previously written, my Ryzen Linux boxes have been working out great except in cases of intentional torture testing with these heavy parallel compilation tasks.

>>61859624
Oh, if you mean that, I agree. I thought you meant being an amd fanboy.
Being biased against Intel is completely normal and understandable.

>>61859749
No, the 12 core is $800 and 16 core is $1000.

>>61859855
>Being biased against Intel is completely normal and understandable.
While I do agree, and I've already bought Ryzen, I still have to say that Intel's business practices shouldn't overshadow the engineering achievements that they have, like being among the very first to an OoO architecture despite their instruction set weighing them down, and always leading the way on the fab side of things.

>>61855218
how bad would a cpu have to score to not be good for emulation?

File: nm process details.png (31KB, 707x797px) Image search: [Google]

31KB, 707x797px

>>61859896
>always leading the way on the fab side of things.
Not for long.

>>61859916
Well aware, but historically.

>>61859901
Dunno but even my old i3-3220 with a GTX 750 ti can run it fine.

>>61859901
Dolphin is one of the most well optimized emulators out there. You'd have to go pretty far out of your way to find something so crappy it wouldn't work on it.

>>61859916
What does "cell height" mean?

>>61859949
Oh, I know that, but my old phenom II 955 non oc was struggling with mario sunshine in areas, no idea what it scored, but im curious where the cut off between able to play well and not being able to play well is, just so I know where this benchmark stop mattering and is just full on dick waving.

>>61859624
See I don't see anything wrong with that. Everyone has biases and as long as you're upfront about those biases and people are aware it's fine. It's when you try to hide your bias in order to push an agenda I have issues with.

>>61859970
Anything Sandy Bridge, or OCed Nehalem should be able to run it no problems.

>>61860023
probably but would still like a number just to know how far into dickwaveing we are.

>>61860039
Probably any higher than the 640 listed on that chart.

>>61856604
Bias !== wrong. That's one of the worst fallacies a person can have.

Marx was bias, but correct about many things.
Gobbles was bias, but correct about many things.

>>61859797
That's scary because then Intel would have literally nothing they beat Intel at other than games and with thread optimization like with vulkan they'll have literally nothing.

The only exception would be their ability to stick a 4k decoder in all their new chips.

>>61860005
>>61860081
As I've explained already, I never implied anything else. All I did was explain why he might fit on an "AMDrone starter kit". Not because he's an AMDrone himself, but simply because he'd be required watching for anyone aspiring to.

>>61850934
what? sockets work in exactly the same way, the only difference is ease of installation. Multi-slot mobos were just as rare as multi-socket ones.

>>61833033

Threadripper is a great CPU for a tiny segment of the CPU market. It's completely irrelevant to 98% of PC users just like Intel's offerings in this segment.

I'm glad AMD is competitive in this niche again, but don't get carried away.

>>61860139
I think it hinges on Coffee Lake being a good buy or not.

>>61860061
have 1700 got 648, running again clamping the afinity to the back 8 cores/threads seeing if that changes anything.

>>61860081

BIASED

A person cannot be bias. Why do people always fuck this up on the internet? Bias is something you have, not something you are.

>>61860165
Raven Ridge probably will be the best sold AMD CPU of this crop.
An "almost as good in CPU, but completely obliterating in GPU if compared to intel" SoC for cheapass laptops.

>>61860193
Their 1700 at stock gets 521.

>>61860165
That's why it was the second highest selling cpu on Amazon until it got sold out.

File: 74942.png (33KB, 650x337px) Image search: [Google]

33KB, 650x337px

>>61860227
>completely obliterating in GPU if compared to intel
There's no obvious reason to think that Vega Mobile will be worlds better than the current GCN crop. Especially not given the power consumption Vega is currently showing.

>>61845125
LMAO and it's fucking pissmark, Intel is totally BTFO.

>>61860316
just noticed my bench was an old version, just got a new one.

File: 1486211958044.png (235KB, 723x890px) Image search: [Google]

235KB, 723x890px

>>61845994

>>61846756
lmao, it ended up being 52% over excavator

>>61860379
we currently have no idea about vegas power consumption, all we have to go on is the 1000$ workstation thing

we need full rx vega to make a call on how efficient it is, and we need some to do low voltage tests to see how it scales down.

>>61833033

To be fair, there is still no low-power ZEN for mobile and still no APU. Those are pretty important.

>>61860441
Soon(tm).

>>61860427
Even so, there's no reason to assume that it will be so much better than current GCN that it will "obliterate" Iris Pro.

>>61860456
the only reason iris pro has anything on amd is its cache, which vega is apparently really good with, if amd laptop has any amount of hbm, keep in mind they could just put 128/256/512/1gb on the thing and probably be very cheap, then all advantage intel could have is gone.

On that note what was the last revision of gcn on an amd apu?

>>61860379
It's actually fairly amazing that the i5 4690 reaches >50% of the A10-7850K with almost a quarter of its ALUs. Is Intel's graphics architecture actually really good?

>>61860316
Just got 562, but i'm also not on a fresh boot or leaving my computer alone for the bench to do its thing as its a 10 minute bench.

>>61860547
>the only reason iris pro has anything on amd is its cache
Not really. You'll notice that it has scaled fairly well if compared with the HD Graphics 4600 in the i7/i5 parts further down on the list. The 4600 has 160 ALUs, the Iris Pro 384, somewhat corresponding in performance scaling.

>>61860560
You're comparing apples to oranges, ALU count isn't important, it's how much they can do.
A simpler measurement is die size.

>>61860619
>it's how much they can do
Literally one FP32 add/mul/fma per cycle in both cases.

File: production.001-1440x1080.png (123KB, 1440x1080px) Image search: [Google]

123KB, 1440x1080px

>>61836944
I'm confused about the Chromium compilation test. Ars also did one, but with totally different results.

I really want to see Linux benchmarks, where it isn't hobbled by NTFS.

I also consider ECC basically mandatory at this point.

>>61860636
But we both know that's not the entire story otherwise AMD and Nvidia would have 10k ALUs per chip already.

>>61860650
I'm also very surprised by the compilation test. Considering the disparity between AnandTech's and Ars' results, I'm starting to doubt its validity as a benchmark.

>I also consider ECC basically mandatory at this point.
Why though? Have you ever actually encountered a bit error?

>>61860650
Half of Anandtech's tests are fucking ancient.

>>61860663
Yes, but the other side of the story is not about how much the ALUs can do, but about the other functional units on the chip. I'm just saying that since Intel has almost a quarter of GCN's ALUs and still match 50% of its performance, those other parts of the architecture must be really good.

>>61860650
It depends on both the compiler used to compile chromium, the version used, and what flags you use.

That's why I don't like compile benchmarks because they're way too easy to fudge up, unless they're controlled properly, they aren't in most cases.

>>61860681
You're looking at Ars Technica's Chromium benchmark.

>>61860707
>That's why I don't like compile benchmarks because they're way too easy to fudge up, unless they're controlled properly, they aren't in most cases.
It does seem reasonable to assume they've used the same settings and conditions on all tested CPUs, though.

>>61860680
Yes, I've seen them happen on my laptop twice (Thinkpad W540 with 32GB of RAM). Noticed when parsing extremely large CSV files. Probably would have never realized but for the format being so inefficient that most bit flips are noticeable as syntax errors.

File: Broadwell%20U%20Dies.jpg (2MB, 3512x2948px) Image search: [Google]

2MB, 3512x2948px

>>61860701
But then you come back to die size, the 5675c has 128MB eDRAM, 1100MHz boost vs 750MHz boost on AMD, the eDRAM alone is bigger than the GPU on the APU, not to mention the 48 EU which is enormous itself.
Intel's solution is way more silicon, and that was on 14nm vs AMD on 28nm

>>61859262
or get the 1900x which has 8 cores/16threads

>>61849793

Waiting for Zen means waiting until potentially 2019. You'd have to be retarded to sit around that long.

Zen 2 is AM4, so you can buy a 1700 now for like $300 USD or whatever it is and upgrade to a Zen 2 in 2 years without any issue.

>>61860789
I've been running two moderately busy servers with 32 GB RAM each for several years 24/7 now, and have never once encountered any error that could be ascribed to bit errors.

>>61860859
You'll notice I wasn't talking about the 5675C, but about the lower i5/i7 parts.

>>61856133

He's biased to a childish degree. You can hear him snicker when reading shit about Intel and AMD. No one who acts like that should ever be taken seriously or regarded as a good source of content

>>61859559
>He's against bad practices, not a blind fanboy.

He spent an entire video trying to spin the 1080Ti being the best card (at the time) as a bad thing somehow.

Anyone who uses dolphin, I have a question
Does it have issues with pre rendered video? been testing it a bit and find prerendered can cause that to fuck up but not normal gameplay.

>>61861523
I'm running an older version of dolphin and it plays pre-rendered video fine of you've selected the right video back-end for it.

>>61861540
So its not a cpu issue? great, thanks.

>>61861558
Do keep in mind I'm playing a old as hell version (4.0 build 6112), and not everythings as polished as the latest version is. On some games the D3D backend is better, others OpenGL is the one to pick.

>>61833363
The closest is cinebench, in which threadripper cucks the 7900X to death. Considering you're getting a shitload more performance for the same price, Intel failed hard.

>>61834920
>Muh 10% more FPS over 100
Kill yourself.

>>61838023
>Coffee lake has to make this domation end and end fast.
It won't. Intel is still under the impression that they'll maintain their monopoly and can crap on the consumer. Zen 2 is pretty much guaranteed to fuck even more shit up.

>>61860411
rekt

>>61860411
/thread

>>61847194

$1000 for CPU
$560 for mobo
$1120 for two nvme ssd's
$1510 for 128GB ram
$799 for gfx card

but I can't afford muh monthly electricity bill...

Try again, little kid.

>>61849690
>not realizing Intel is allowing AMD to get back on the market so they won't be hit with anti-monopoly laws

Red or blue we all go in the same direction.

>>61855402
Should add Hardware Unboxed, RedgamingTech

>>61862558
Blender, Cinema 4D and Adobe Premiere Pro too.