Single thread test

>2950X @ 4.5GHz = 195
>7900X @ 4.5GHz = 192
What does this mean?

Attached: cbr15-ipc-comparison-scores.png (1278x959, 87K)

Other urls found in this thread:

notebookcheck.net/Threadripper-1950X-benchmarked-at-5-2-GHz-on-all-cores-using-liquid-nitrogen.241680.0.html
twitter.com/SFWRedditImages

D-DELID THIS!!!

it mean that skylake-x is shit, and that intlel will soon lose their biggest advantage

>D-DELID THIS!!!
wow i cannot believe you made that joke.

source?

The comparison to the A series shows just how much of a leap AMD made with Zen. Intel is simply fucked.

2950x at 4.4 is beating 7700k at 4.5, impressive

reminder that this was only a refresh, a tick
zen 2 will be a tock

>7nm, better IPC and frequency
oh fuck

Good to know that the 1700 I bought a year ago is worth less than $100 now.

>they managed to do this while using a process node made for smartphone cpus, licansed from samsung and slightly tweaked by glofo
I don't even want to see what they can do with de node developed using IBM knowledge

>7nm
>40% higher clocks or 60% power reduction
>around 10% actual IPC increase
>more cores CCX rumoured
yeah...I think AMD is back

>What does this mean?
It means that you must delet this thread right now. Plz.

Attached: IMG-20180731-WA0005.jpg (720x1280, 195K)

>ayymd has better single thread performance than intlel at the same clock speed
Can you kindly remove this thread?
Sir, please, this will hurt my investment portfolio.

that's one of the drawbacks of going amd but i think that 1700 will last you few years unless you're a gaymer.

>40% higher clocks
keep in mind that this is compared to 14LPP, not 12LP
that'd put a 1950X XFR on the 5.8GHz range

a 10% IPC increase would put the 2950% @4.5GHz at ~215cb
the 8700K needs to run at 5GHz for that

3950X?

yeah, I was just making the comparison against a 14LPP cpu, which is what glofo is basing that 40% perf increase

DELIT IT REEEEEEEEEEEEEEEEE

that increase would put the all core turbo at ~5.2GHz as well
notebookcheck.net/Threadripper-1950X-benchmarked-at-5-2-GHz-on-all-cores-using-liquid-nitrogen.241680.0.html

It means Zen2 is going to wreck Intel's proverbial anus if AMD can pull this shit off with just Zen+.

>40% higher clock on 4 Ghz

>4 Ghz + 1.6 Ghz

>5.6 Ghz overclock

who gives a single fuck about single core perf anyway
its not like the difference is such that it can justify the money spent

>*NEW* single core performance doesn't matter!

pretty sure he was reffering to amd not intel moron

that's only if AMD do not add cores. 5Niggurhurzt with an IPC increase would still be quite the achievement, don't you think?

Samefag intel shill. AMD now has more cores AND better single core performance.
C O P E
O
P
E

pretty much that
it'd allow an 7nm EPYC 7601 do an all core boost of 3.9GHz on the same 180W TDP

of course it would, that's why I'm worried about those morons who want 16 cores on mainstream instead

>base clock of a 1300X would jump to 4.9GHz, boost at 5.2

this, enterprise would benefit far more of a massive core count increase while maintaining similar clocks than consumers
maybe that's why they're splitting production between TSMC and glofo

>maybe that's why they're splitting production between TSMC and glofo
bingo

2950X is BASED

A 6 core CCX would be pretty dope tho.
Just imagine 6c/12t APU's, daiyum!

>12 or 16 core dies for enterprise with small clock improvements
>8 core dies for consumers with massive clock improvements
let's hope this is true

>being this idiot
sigh

CHECKED

Attached: doug_dimmaDab.jpg (1204x1200, 134K)

i rather get a 4c and let all of the other area to be filled with the gpu

but that'd not be needed when your 4 core apu can do 5GHz base, 5.4GHz boost and 1.75GHz GPU clock along with a 10% IPC increase

>ass-blasted kike screeching autistically
a socket a year put the goyim in fear

Its based as Fuck. No cpu can touch the price/ performance of it. The closest competitor is the 7980xe which costs 1000$ more.

and since in enterprise core to core latency isn't that important, they could do 3 or 4 CCXs and cram 1 or two more memory channels in each die

it's just too funny that this one retarded Intlel shill ignores that CPU completely cause it beat the 7980XE with ease. fucking based and redpilled CPU

Attached: prove_me_wrong.png (1199x700, 1020K)

NOOOOOOOO
NOOOOOOOOOOOOOOOOO
STOP IT GOY
INTEL IS BETTER FOR 144HZ GAMING
YOU NEED THOSE EXTRA 10 FPS

Attached: 1510976023363.jpg (179x200, 8K)

this makes me think about something
does intel have anything else down the line besides ice lake? it'll be released a few months before zen 3, and then they'd need something else to compete with it

>inb4 AMD overtakes intel in clocks and IPC, so it'll be obvious that intel's compiler gimps amd cpus and they get sued again

>tattoo
Pass.

kek

Attached: 1534204099106.png (1920x1080, 611K)

This would be glorious.

b-but zen 2 will only perform slightly better than zen+ goys, it will still lose to the 9900k! intel is the best! t-trust us goys!

OYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYY VEYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYY

Attached: 1520633925774.jpg (691x771, 112K)

this is antisemitic, bankrupt and finished.

>Pass
just cut the arm off... the rest is still good.

It means Zen2 will top the market in a lot of benchmarks.

>2950X
>4.40 GHz 191
>4.50 GHz 195
So theoretically at 4.7 it should get 203 points, which is above Coffee Lake IPC wtf

who cares? we know the only benchmark that matters is passmark :^)

how much would it score at 5.8GHz?

And geekbench.
But only 12% of the subtests of the single core tests.
And userbenchmark, but only some subtests like quad core OC

should I hoard some stocks while they are still low now? Will AMD moon?

>5.8GHz
how?

Only gaming IPC matters.

I wonder, do the marketing departments of AMD and Intel analyze this? Does some poor intern have to create a portfolio of "antisemitic Intel memes todays youth seems to be using"? I would pay to see those briefings.

14nm 1950X has a max turbo of 4.2GHz
4.2+ 40% = ~5.8

Can't answer that for you

>Single core performance doesn't matter
>You have to pay higher price to buy the superior AMD
Am I dreaming? Is this reality? AMD won?

so basically a mixture of IPC, memory latency and raw clockspeeds, like something that zen 2 will crush intel on?

What timeline is this?

Well, there's the guy that had to explain Spurdo to some bald boomer of the FBI or NSA, so...

not yet, not yet
in 8 months

It took AMD a while to prep the ovens. Let the shoah begin.

AMD tricked intel into building their own ovens, they're only going to force them inside now

No one is going to let AMD stock explode, no matter how well they do for the next few years. Every rating body out there could list AMD has a strong buy, AMD could post $3billion in profit per quarter, and their stock wouldn't consistently break $25 per share.

AMD is a very weird way to spell IOTA

2019 some serious shit is gonna hit the fan for Intelaviv.
>But muh ayyGPU
>But muh weefee
>But muh *NEW* solder

Attached: 1534471569185.jpg (207x245, 44K)

Post source you scaly faggot.

Attached: Screenshot_20180519-210135__01.jpg (564x589, 65K)

>intel is gonna brag about better igpu while zen gets better single core performance
WHAT IS GOING ON

its above in ipc, but do to latency its "realworldgaming" ipc is still lower, but this honesly only effect like 10% of FPS if you have like fastest nvidia TI gpu

Attached: ugly.png (480x586, 134K)

>skymeme i7 6700k at 4.3GHz
OC benchmarks

>into the bin it goes

shit ipc

2950x is pretty goat but until it can reach 5ghz intlel will still dominate in single thread and gayming.
Where the fuck are you getting the 40% from? Do you seriously think 12/14nm -> 7nm equals a 40% clock increase?

glofo thinks this
tsmc is more conservative, 35%

Are passmark and cpu-z still trying to ((((((update)))))) their results?

40% or more frequency uplift vs 14nm LPP is straight from GloFo's press material.
The issue is that these statements aren't qualified. It could mean that the process simply has more OC headroom before hitting a voltage wall, not that they can raise fmax that drastically inside of normal TDPs.

they'll always do it, they have to protect their greatest ally

Attached: uBgThsu.png (1562x332, 129K)

glofo's press material strictly claims 40% improvement on the same power draw

and there is this

Attached: 6e530f3f294e920ca239c237738718bb_XL.jpg (750x563, 162K)

I'm already saving my neetbux for Zen2

Fucking behemoth mopping the floor with Intel

Attached: untitled-3.png (672x794, 44K)

but that's TSMC's 7nm, the one they're using on 7nm enterprise vega

>the rest is still good
except... if she has a tattoo, she's fucked in the head

Foundries don't make statements like that because power/frequency curves have several important inflection points.
Their bullet points are 40% or more frequency, or 60% or more reduced power at iso frequency.
They had a couple presentations on the process in 2017. Their keynote should still be on youtube.

that's why i said "and there is this."

cut the head off then

>8700k BTFOs everything AMD has
LMAO

oh, really? then why is pic related on the o

Attached: 7LP.png (781x620, 167K)

>intel shills still trying to save face
I wonder what they'll say when zen 2 comes out

Only netflix matters.