Only 29% better multi-core performance

>only 29% better multi-core performance
>only 19% better single-core performance
>302% higher price
Why do people still unironically buy intel again?

Other urls found in this thread:

forums.anandtech.com/showpost.php?p=38204350&postcount=1048
forums.anandtech.com/showpost.php?p=38198889&postcount=996
forums.anandtech.com/showthread.php?t=2465645&page=41
anandtech.com/bench/product/1684?vs=288
en.wiktionary.org/wiki/proportional
twitter.com/SFWRedditImages

Because we're not poor...

muh emulation.

On a serious note, I'm not touching anything derived from the reason AMD is doing so badly right now. Have Zen beat a 3570k while giving double the cores and quadruple threads and I'll switch immediately.

>Unrealistic expectations: The post

Zen's going to be good, but not that good.

> Synthetic benchmarks equating to realworld performance

That's pretty low expectations, user. It should theoretically be just under Skylake following 40% increase and comparing Intel's offerings as well, but I'm not following that shit.
Releasing a product weaker than Sandy or Ivy Bridge would be terrible for everyone.

I'm going to make a Zen gaymen musheen and nobody can stop me.
NOBODY.

It will still be enough to put intel on suicide watch. Pic OP posted is an excavator CPU so that + 40% better IPC is what Zen will be.

So theoretically a quad-core Zen chip with a CPU frequency of 3.5 - 3.8 GHz will have:
>multi-core passmark of ~7800
>single-core passmark of ~2480

In addition to that rumour is Zen has even surpassed that 40% better IPC.

Anyway the 8-core Zen chip will be a fucking monster of a CPU.

>going to be good, but not that good.

>good
>not being able to beat a shitty budget 4 year old Intel

According to you it's going to be a complete failure.

>budget 4 year old Intel
On what planet has any i5 been budget? Or are we going by the Intel marketing where you need an i7 to make a webpage?

$250 CPU is budget, if you consider that all the decent CPUs have always been over $400. Even the AMD ones. Did you forget the $900 Fx-9590?

Zen: reaching 2013 Haswell's IPC perf in 2017.
Somehow, it'still A Huge Win for AMD

Passmark shitposting deserves an instant perma ban

Excavator is 50% behind Haswell in X265 encoding.
It is over 80% behind Skylake.
The Stilt did tons of testing with all chips at equal clocks. AMD's BD family arch is astoundingly far behind.

Look at Sandy Bridge vs Kaveri benches on anandtech instead of regurgitating utter stupidity.

Zen will perform around Ivy Bridge - Haswell per clock. Summit Ridge is competing against Haswell-E and mid range Broadwell-E. It is not a mainstream desktop chip.

>Passmark shitposting deserves an instant perma ban
It's the most accurate synthetic benchmark out there.

>Excavator is 50% behind Haswell in X265 encoding.
8-bit HEVC encoding but not 10-bit HEVC encoding, which is what people are mainly interested in.

>It is over 80% behind Skylake.
Source: my ass

>Look at Sandy Bridge vs Kaveri benches on anandtech instead of regurgitating utter stupidity.
You dumb user, it literally says >I'm not following that shit.
I will believe nothing until the product is out.

Also, please post your sources on x265 encoding.

It isn't accurate for anything, you tech illiterate child retard.
Your post is just even more proof that you deserve a perma ban for being so fucking retarded.

forums.anandtech.com/showpost.php?p=38204350&postcount=1048
forums.anandtech.com/showpost.php?p=38198889&postcount=996

See above, dipshit.

>8-bit x264 and x265 encoders
gr8 b8 m8, I r8 8/8

>8-bit
What are you even doing on Sup Forums?

>10bit encoding is going to be magically less intensive and a weaker CPU will perform better relative to a much stronger CPU!

Grasp at more straws, children.

>compare notebook cpu with 15W TDP to a desktop cpu with 84W TDP.

Try again, retard.
They were all set to the exact same clock speed. The intel chips were underclocked, and the Carrizo chip had its cTDP raised to the limit and was set to the same clock speed.

Can you show us proof that carrizo chip was constantly running at 3.2 GHz during the test with no throttling whatsoever?

Read the fucking thread. It was discussed.

AMD's APUs only heavily throttle the CPU when the IGP is active. The IGP wasn't loading during the test, and The Stilt monitored clock speed as well.

>Read the fucking thread. It was discussed.
I'm asking for proof if the chip maintained a constant clock speed of 3.2 GHz during the test. Can you provide it or not?

Perhaps because an i5 from 2011 outperforms this $70 chip?

>posts single post view
>demands people to read the rest of the thread
You can't be this retarded.

Read the fucking thread. Its on the very next page.
forums.anandtech.com/showthread.php?t=2465645&page=41


>how do I click a single link to go back to the page view!
Its like you really are retarded.

Because they like to throw money to their ((greatest ally))

That reminds me, Skylake was a huge fucking flop huh? Literally no IPC gain at all. A Haswell i5 clocked 100 MHz higher shits all over a Skylake i5.

i didn't actually know it scored that well. Just how mad would it be to buy one of these if I mostly mess around with linux and hobby programming?

>only 32% greater single-core
>only 24% greater multicore
>275% of the price
>needs expensive motherboard instead of free one picked up from the side of the road
>needs expensive RAM instead of free RAM picked up from the side of the road
>doesn't come with free case and PSU like LGA 775 picked up from the side of the road

you forgot the best bit

>don't have to pay heating bills anymore because your Xeon will keep your family warm for free

OP is a literal retard, and the junk bench hes showing is akin to CPUBoss, the shit that morons on Sup Forums constantly spam.
Real world the difference between the Excavator Athlon II X4 845 and a Sandy Bridge i5 2500k can be massive.

anandtech.com/bench/product/1684?vs=288

The relevant discussion about why they were comparing a notebook cpu came like 5 pages before.

It's kind of funny to see retards like you argue this much over worthless data.

>Graph doesn't start at 0

This LMAO.

And?

>better process
>more power consumption
>only 3% more multicore performance
kek

The whole point of a bar diagram is that the length of the bar is proportional to the quantity it represents.

The relevant discussion is on the page I directly linked, tech illiterate retard.
One well known fanboy tried to call out The Stilt's testing methodology, The Stilt then showed that the chips were all running at the same clocks, making the objection totally absurd.

Its normalized to the Excavator chip.

Its just desperate shitposting after being proven wrong at every stop.

Bad goy, bad goy!

And? What part isn't proportional to the scale indicated?

All

The real difference of the values are less than 66% yet the bars of the i5 are up to eight times as long.

the point of a diagramm is that you can make visual comparisons without looking at the scale. You might as well just have a table of numbers without a diagram.

I didn't ask what part wasn't proportional to the scale you're crying for. I asked what part isn't proportional to the scale indicated?

>hurrrrrr
>da bars being long is mean!

The percentage differential is shown in exact numbers.

>he doesn't look at the scale
Leave Sup Forums. Now.

Look up what the word proportional means.

Huge flop?

BONDS BTFO RUTH

>Lower score than that HASWELL-H CPU

Haswell was TWO fucking generations before Skylake.

What part isn't proportional?

>55869861
The physical length of the bars.

How are they not proportional to the scale indicated?

en.wiktionary.org/wiki/proportional

>proportional
>At a constant ratio (to). Two magnitudes (numbers) are said to be proportional if the second varies in a direct relation arithmetically to the first.
The ratio of bar lengths is different from the performance ratio.

I said "physical length of the bars" not "length of the bars proportional to the scale indicated".

There is no haswell-h on there.

How are they not proportional to the scale indicated?

>"proportionate in proportion to nothing"
Do you know what "proportionate" means?

>graph starts at 0
Do not pass go
Do not collect $0.02 from intel

>"I don't know what 'zoom' is"

>he's upset he didn't get two dubs in a row

Just look at the first benchmark JetStream ST.
The ratio of bar lenths is ~5 but the ratio of the values is 1.3948.

5 =/= 1.3948

The bar length is in relation to the values over 90. How are they not proportional?

>create a distorted physical representation of a quantifiable measure to enable quick visual comparison of several results by modifying the scale in your favor
>defend your original submission by arguing the representation is proportional to the given scale when the actual discussion is about the scale not starting at zero

Why over 90 why not over 5000?

>all scales must be in the one most beneficial to my fanboy position
Graphs are technology. If you can't get over people actually using them you really don't belong on Sup Forums.

>Comparing Q1 2016 to Q2 2014
Nice try AMD shill

>It's AMD's fault that Intel didn't release any new 4 core CPUs so far this year

>what is skylake

>someone chooses a scale that is most benefical to their preferred brand
>argue that the chosen scale should be neutral to both brands
>that someone argues that I'm commiting the same fallacy
kek I didn't expect Sup Forums to deteriorate this badly.

>skylake
>2016
Pick one

Graphs are technology. If you can't get over people actually using them you really don't belong on Sup Forums.

Are you literally retarded?

Because when you have enough money, price/performance becomes less relevant, and absolute performance becomes more relevant. Price/performance is a more useful metric maybe if you're buying MANY CPUs, but just as useful is performance/watt.

>$250 CPU is budget,
Guess we're on a planet where $50-125 CPUs don't exist.
i5s are mid-range.

And don't even bring up CPUs that are not Xeons or Operons that are over $500, those are "enthusiast" chips for people with more money then sense.
By your logic the only top end cars are the ones in the millions.

>the only top end cars are the ones in the millions
But this is true. Nobody would call cheap shit like a Golf a top end car.

You have very obviously never driven a Golf GTi...

>122%
>bar three times as long

>comparing your results to a bunch of really old processors and one from two generations ago that is kicking your ass

You're beating the Haswell-E by clock speed, that's all. The boost speeds are 3.6 and 3.9GHz respectively and that is where the difference is.

There's practically no improvement clock for clock.

>all of this bitching and shit-flinging over artificial benchmarks
>absolutely little practical difference during times you run your OS to play games and shitpost on Sup Forums

There wasn't much shit flinging. The OP got called out for posting a worthless stupid bench by an user. That user then backed up his position, and the OP did nothing to backup his claims. Then it turned into semantic nonsense, and a few autistics arguing over graphs.

Thanks for the (you)