AMD Status: Ass Blasted Beyond The Milky Way
>>61338254
Thanks for the i9 release dude. Now I'm going to get me a threadripper.
>>61338083
The stupidity of this board will never cease to amaze me.
>>61338277
>>61338288
>being this mad
Intel are the last people that should be making fun of others for using glued together dies.
>>61338083
is that real? are we living in idiocracy?
>>61338306
>dem retarded arguments
AyylMaoD and Incel inbreds are really fun to read
>>61338342
AYYYYYYYY
>>61338355
You know what?. I think I'm going to build my new Ryzen rig this week just to piss you off.
>>61338369
>>61338355
inconsistent performances because of the 4 glued-together desktop dies
>INCONSISTENT PERFORMANCES
>4 GLUED-TOGETHER DIES
LAUGH AT AMD NAO
>>61338306
AMD is trying to re-enter the server market segment, Intel continues to deliver 20+ years of uninterrupted data center innovations while maintaining broad ecosystem investments. Xeon CPU architecture is proven and battle tested, delivering outstanding performance on a wide range of workloads and specifically designed to maximize data center performance, capabilities, reliability, and manageability. With next-generation Xeon Scalable processors, Intel expects to continue offering the highest core and system performance versus AMD. AMD’s approach of stitching together 4 desktop die in a processor is expected to lead to inconsistent performance and other deployment complexities in the data center.
>>61338379
Should I get a 1800x or 1700?. Either way I'm getting the Microcenter mobo combo.
>>61338083
Can we just have an Ayymd vs incel general? I feel like it's the same vocal minority always posting this stuff
How much more REKT can AYMD get?
>>61338421
No side ever listens to each other. I'm confident AMD is the absolute best and jews think Intel is the best. That's all their is to it.
Either way I will take opinions (whether you hate AMD or not) on if I should get the 1800x or 1700?
>Intel mentioned AMD by name in one of their marketing slides
Don't the top companies usually avoid that to avoid giving other companies any advertising/brand recognition?
>>61338333
It's real. Official Intel sanctioned shitposting is now here. https://www.computerbase.de/bildstrecke/78687/
>>61338446
>top of the line server CPU vs underclocked mainstream consumer CPU
what are the chances actual server guys got pissed at intel for insulting them with these slides?
>>61338083
what the fuck
this reads like some low effort shitpost on /g/
>>61338467
Usually is the key word here.
Intel has nothing until 2020/2021 and they know only too well. This is the beginning of hilarious damage control that will only get more desperate as AMD continues to optimize their new architecture that already matches what Intel has been perfecting for nearly a decade.
>>61338533
Where to you think they got the "glued together" stuff from?
>>61338528
"Actual server guys" aren't the ones making the big money decisions and placing orders for thousands of chips.
>>61338533
Intel has outsourced its marketing department to /g/ it seems.
>>61338398
1800X isn't worth it. If you go 1700 go B350, if you go 1700X (which is more likely to hit 4ghz) go X370
>>61338355
Doesn't it scale better than intels shitty ringbus tho?
>>61338594
Yeah but how far does the 1800x go with easy simple overclocking?. I just use the built in software to overclock because i don't have the patience to fuck with all the bios settings. What should I get it to with a decent air cooler?
>>61338588
Exactly. Those decisions are made by accountants who run every big company these days, and AMD can always undercut Intel due to modular architecture with huge yields.
>>61338632
>Those decisions are made by accountants
My god, how big of a fucking retard are you? You need to stop watching the big bong theory.
>>61338446
> OFFICIAL FUCKING SLIDES
> $2220 server processor
> Versus
> $430 desktop processor DOWNCLOCKED BY 40%
Omg send help I am dying from laughter
>>61338645
Oh wow you sure convinced me you autistic NEET.
>>61338395
10,000 lightyears in MS Paint
>>61338588
pretty sure CTO has the last word
>>61338744
but lightyears is a unit of measurement for length not a time
>>61338668
Clocks are irrelevant at that comparison.
Learn to read bitch
>>61338774
thatsthejoke.avi.jpg.scr
>>61338482
what the fuck?
>>61338505
>4% difference for web server work, the one thing xeons should dominate effortlessly
>>61338277
>>61338314
>>61338467
>>61338528
>>61338668
NO SHUT UP
>>61338824
What if Brian really lurks /g/?
>>61338859
>if
>>61338859
>if
Come on, these slides are pure condensed /g/ shitposting.
>>61338083
hope you're joking
>unless you like overpaying
>>61338780
Clocks aren't irrelevant. Different CPUs can handle different clocks at different thermals and powers.
Which the Ryzen stock clock is 3.6 ghz, boosts to 4 ghz, with 95W TDP.
The Xeon stock clock is 3.2 ghz, boost to 3.7 ghz, with a 130W TDP.
So with a lower TDP (which likely means it also has lower power usage) and 25% the price, the Ryzen has a 12% clock advantage, which would seem to put it on par or better than the Intel given Intel's chart, except Java, which seems to be an optimization issue and will likely to be fixed in the near future, seeing as Ryzen is so new.
In other words, Intel is absolutely BTFO.
>>61338306
I'm not expert, but isn't Intel's ringbus performance subpar?
Intel posting shit that doesn't mean a fuck and in some cases outright lying. The real test is in what software is being used. Ryzen scores highly here in terms of cost for what oit does. It beats Intel in a load of real software use tasks soundly and those it does not the price difference alone makes up for it. Intel is so butthurt they are hiring monkeys to lie for them.
>>61338931
Yes, hence the shitpost slides and FUD damage control.
>>61338083
>Intel workshop
Do they give free coffee?
>>61338931
The next xeon will use a grid design instead of the ringbus.
>>61338962
It's just more of the same shit. Mesh of ringbuses.
>>61338902
4 DIES GLUED TOGETHER THOUGH
>>61338446
>Server side Java
How is that a valid benchmark? Who is the pajeet doing this slides?
>>61338982
This is fucking rich coming from the assholes who've been putting glue under their heatspreaders for years now, and it's now spread to their HEDT platform as well. With disastrous results.
>>61338083
Does this means that Intel...is...DESPERATE?
>>61338446
>the city of '''Intel'''
>tfw I have some glued-together desktop dies at work and they are still usable 10 years later
>>61338594
I read that x370 is going eol and they'll be stopping or have already stopped bios updates. Check it out if your really planning on getting one
>>61338902
Are you mentally retarded?
>Clocks are irrelevant at that comparison.
Do you know what that means?
add more bingbus :---D
DELID
E
L
I
D
>>61339498
add more clock sbeed and threbs! x---DDD
keep chibset and manufabuturing probcess xddd
>>61339431
Nice FUD. It's literally just Asrock end-of-lifing one board (Asrock Gaming K4 X370).
>>61338083
>i've got them this time!
>>61338355
> EPYC Desktop
King of Threads
Hi, Gerald.
Take this down. Delete your post.
>>61339559
Hey just making sure you don't get it bud. I've got an asscock myself
>>61338314
Or glue altogether as they use some jewglue under their heatspreaders
*breathe in*
AM4
Skylake-X and Ryzen R7 1800x.
Constantly 100 ns or an average of (40+140)/2 = 80 ns?
On the other hand, EPYC may have higher latencies since they have 4 dies. And the Xeon 28 core too since it has also 28 cores in the mesh instead of 10.
>>61338083
>Austism: The Thread
JEWTEL IS FINISHED AND BANKRUPT
JEWTELFAGS CONFIRMED ON SUICIDE WATCH
HOW CAN JEWTEL FAGS EVER RECOVER?
>>61338446
AHAHAHAHAHAHA
It's like Intel switched sides with AMD, now they are the ones doing those stupid slides.
>wintel cuccs full damage control
The future is LAMD get on board.
>>61338446
aren't those slides supposed to be pro intel? did AMD outsource their marketing staff to intel?
>>61338446
>Our CPU that costs 5x as much is better if you compare it to an underclocked AMD product
How will AYYMD and their 4 Glued-together desktop die ever recover?
>>61338446
>compare $10K Xeon to 1/4th of an EPYC chip with all the enterprise features disabled
>win by 4% in web server
HOLY FUCK HOW BAD IS THE DAMAGE
OH GOD
IS INTEL, DARE I SAY IT, FINISHED AND OR BANKRUPT?
>>61338355
>monolithic die bretty good?
>>61338369
I was going to build some internet machine with a shitty Pentium (or one of those Atom boards) since the new APU's aren't out
But now I will either build some obsolete FM2+ or AM1, maybe even go with a overpowered R5 1400 to get something modern
Seriously this shitfest slides only tells me the serious amount of problems their systems will have, if this is the best they can do for their enterprise consumers paying hundreds of thousand of dollars I can't fandom the shit they sell to consumers paying hundreds
>>61338306
top kek
>>61338355
>This is how Intel tries to sell Xeon server room melters now
Please Intel, think of the poor power companies that you're hurting with that processor.
>>61346772
Hurting? They're probably celebrating.
>>61346789
>be coal plant owner
>data centers now need dozens of times more power for Xeon
>profit
>Glued together dies out perform non-glued
>Position it as a negative
Intel you fucking blew it
>>61338083
>glued together
LOL Intel has entered the shit posting business
>>61338083
>glued together
At least the glue is good, sand merchant
>Intcels
>>61338446
>$2220 part vs $430 part
>not at least 100% better in every case
>worst case is a 39% difference
>for a part that is 516% more expensive
Ayy lmao
What you nerds don't understand is that enterprise clients pay for quality. Uncompromising quality.
It literally doesn't matter if Intel chips are 5 times more expensive. Because the money saved on stability, B2B-level support, a holistic ecosystem, and long-term maintenance far outweighs the cost of some tiny chip.
Face it. AMD is not ready for the enterprise market.
>>61344470
>>61347988
Seems like some people don't know how to read the chart.
This is not a chart on which chip is faster. It's a chart on how much more the performance of the chip increases when you turn SMT on. Basically intel is saying that performance will never be worse off or inconsequential when you enable hyper threading
It's a valid point that should have been better stateded. Since they're comparing deltas, they should have used a comparable part.
These slides are embarrassing.
>>61348080
t. Shlomo
>>61341140
Skylake for untuned software that you want to "just work", Epyc for int-heavy stuff you can fuck around with schedulers/affinities/mem allocation to get highest theoretical gainz.
Also low core count VMs is clearly a win for Epyc. Just keep spawning 2 or 4 core instances all day and give zero fucks, the performance will be great.
>>61348138
Paying 5 times the cost of a ryzen should get me better SMT performance.
>>61348271
>>61348271
>Paying 5 times the cost of a ryzen should get me better SMT performance.
Sure. That's what the chart proves.
>>61338083
What would your marketing slide look like, /g/?
>>61338446
8C/16T vs 8C/16T
clock for clock the AMD processor is way slower.
>>61338467
Their only x86 competitor and VIA doesn't count.
>>61345920
Get an AM1 build, it's fun.
>>61338306
did intel hire a 4chin tard?
>>61349560
anon, they do not compare it to EPYC 8c/16t taht has different feature set compared to desktop ryzen
hooo boy that's some good shilling notes
>>61338446
>4 dimensional benchmark alchemy
>>61349700
UPSET
P
S
E
T
>>61338342
I'm not an inbred! My sister is though
>>61348080
>>61348138
It's still misleading bullshit because AMD's SMT has been shown to be more efficient than Intel's HT.
http://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/15
>On average, both Xeons pick up about 20% due to SMT (Hyperthreading). The EPYC 7601 improved by even more: it gets a 28% boost on average.
>intel literally unironically shitposting IRL
>>61349701
We take all competitors seriously, and while AMD is trying to re-enter the server market segment, Intel continues to deliver 20-plus years of uninterrupted data center innovations while maintaining broad ecosystem investments. Our Xeon CPU architecture is proven and battle tested, delivering outstanding performance on a wide range of workloads and specifically designed to maximize data center performance, capabilities, reliability, and manageability. With our next-generation Xeon Scalable processors, we expect to continue offering the highest core and system performance versus AMD. AMD’s approach of stitching together 4 desktop die in a processor is expected to lead to inconsistent performance and other deployment complexities in the data center.
>>61352897
>stitching
Hello friend, while I agree entirely with what you posted, I'd like to quickly like to inform you that the correct term that HQ would like us to use is "glued" - Please use it next time, instead of "stitching". Thank you for your cooperation, and your continued efforts to spread the word of the superiority of wonderful Intel™ products.
Yours sincerely,
Drone #44963-b28
>>61338083
>>61338355
>>61338446
>>61348080
>>61348271
>>61349560
>>61351270
>>61352897
>>61338083
This can't be real. If it is it's some of the funniest damage control ever.
>>61353118
It is
>>61353089
JEW KILLER
>>61349560
good thing the AMD processor can increase its clocks by 80%
>>61353089
And the Intel chip is TWO TIMES MORE EXPENSIVE than the AMD one it's losing against.
>>61338083
>holistic
Intel is akin to homeopathy?
>>61349582
The thing is when you're the #1 company any time you specifically mention your competitor you're just giving them additional recognition. When you reach the top there's no where else to go except down basically. The top company usually uses a more vague term like "the competition" or "the next leading name brand" rather than naming a specific company. The smaller competitors will occasionally use the name of their larger rival when comparing their products because by comparing themselves with a larger company they're getting their name out there and they're also explaining what their product is for.
intel is getting extremely desperate
>>61338446
intel is weeping
>>61338591
what the fuck! thats insulting!
/g/entoomans would do a much better job than jewtel did
>>61354510
Indeed, it seems more like AMD outsourced their marketing staff to the Intel staff.
Now if only Keller could rub off some of his magic on Raja...
>>61356022
Raja?
The dude behind like every good Radeon period (R300 and R700/Evergreen family)?
>>61356065
I mean for Vega. It has been kind of underwhelming.
>>61356097
You mean FE?
The card with disabled everything, even fucking discard accelerator from Polaris (the fuck did they even disable it?)?
That's not an indication of Vega performance.
We have to wait and see if pajeet can pull another R770.
>>61354510
My bet is they hired /v/ to make these slides
>>61338083
Literally everyone and their mothers is giving Intel shit for these slides. What the fuck are you on about?
>>61338288
it's a memer trying to get replies
no one can defend those intel slides
>>61356544
Triggering retards like you so i can laugh with a thread that hasn't died since yesterday.
>>61338446
>dat 40% downclock
>dat price difference
INTEL LITERALLY NOT FIGURATIVELY ON SUICIDE WATCH HOLY SHIT
>>61339498
do cpu engineers play a lot of factorio?
>>61349338
>no critic's goy
>>61349338
>>61356874
>Holistic
More like Holocaustic.
>>61357049
>Oy, vey, brother!
>>61338083
>glued together die
that's _literally_ what Intel did though
not a fanboi, both suck
>>61357412
Intel didn't have Infinity Fabric.
sexy glue coming through
>>61356643
>I-I was only pretending
This shit is real.
Fuck you Intel.
>>61358640
what website is that?