The real reason most reviewers tested with 2133 and 2400 RAM on Ryzen

>the real reason most reviewers tested with 2133 and 2400 RAM on Ryzen

Other urls found in this thread:

youtube.com/watch?v=RZS2XHcQdqA
intelpentium4litigation.com/
youtube.com/watch?v=L4K7eIEAJx0
crucial.com/usa/en/memory-performance-speed-latency
twitter.com/SFWRedditImages

what's your favourite anime?

I'd love to know how they got 3600 CL16 memory to work. I can't get my system to run 3200 CL16, stuck at 2667 :(
I did get a 15 fps frame boost in fallout 4 going from 2400 MHz to 2667 MHz though

what is your favourite anime

Please delete your post.

Thank you.

>use 3600MHz RAM on Poozen that only works on 0.001% of chips/motherboards and it probably can't even pass memtest
>use slower RAM on Intel
>Poozen is faster
Woah...

Mobile Suit Gundam. I'm old.

Stay mad, corelet

They used 3200 CL14 on both and Ryzen overall won there.

Diminishing returns after 2666, though, except in two games

lol corelet BTFO

I keep seeing such wildly varying benchmarks and I don't know what to believe anymore.

Any legit reason for Intel not to solder the IHS? Ryzen's die is about the same size.

Hmm, why would you solder the CPU if jew shit TIM is cheaper and the goyim are buying it anyway?

who do you believe, first world country or India ?

It's irrelevant, the margin of difference is minimal enough that you'd never notice unless somebody forced you to run benchmark comparisons

>yet another fake benchmark
5/10 for the effort though.

bad bait

Kek, yes, go ahead and buy those super expensive ram with 32Gb more expensive than your fucking CPU.

...

If we are going by performance/price ratio, we have to count the expense for the motherboard and the ram after all.

>b-but AMD is so cheap

Let's see the 7700K at lower RAM speeds as well.

This. The 7700k still fucking obliterates literally everything Ryzen has to offer in terms of bang for buck.

>RAM prices going apeshit this year and still climbing all the way till Christmas.
>Ryzen users need to get super ram for their CPU to get the most out of it.
This shit is going to add quite a bit of expenses to getting a Ryzen system.

>poorfags
KEK

l m a o intelavivs are the poorfags now

you mean all users

>implying Intel's current offering don't still perform amazingly well at lower memory frequencies

So you could avoid warranty, intel hate free performance from overclock, same reason why K processor exist. jew at intel want you to pay for that free stuff

REEEEEE

What's the different between 1700 and 1700X? 5% more fps?

>in gaming

right now, luck of the draw and motherboard. The taichi board went from 3200+oc to 3733+oc

once bioses get their shit together the speed will go up but fuck me the wait is hard.

Where the fuck does this retarded image with white text on a background of bright screenshots come from?

You got 3 issues.

1) memory related, an issue that will get fixed in time

2) ccx and the interconnect being tied to memory speed, possibly fixable and able to run independent, but assume it can't.

3) something in windows.

So you got people who run the memory at 2300~ speed because they cant get it to work faster

You then have games able to run on 2ccx that currently kills preformance

and you have some shit going on in windows. The above is likely ideal scenario. 3 games that take advantage of more then 6 cores, along with along with a system that can get faster memory to work, and likely post some windows fixes.

>i-it's okay. Intel is better for poorfags that can't afford good memory.
>We don't care about the absolute best performance anymore
>7700k is good enough!
>just wait for Coffeelake!

k e k

>the real reason
Shitty BIOS was the real reason.

youtube.com/watch?v=RZS2XHcQdqA

It's easy for anyone to manipulate FPS benchmarks if they are not using the in game benchmark. AMD did this in their Ryzen demo with Sniper and BF4 by displaying the skybox more often. It simply not a true benchmark if the variables are all over the place

>MindBlank Tech
>literally who tier reviewer
Yeah, no. Meanwhile nearly every single big and well established reviewer out there came to the same conclusion that Ryzen is shit at games

ItsOnlyOKWhenIntelDoesIt™

intelfags now becoming conspiracy theorists just like the muh russians liberals

intelpentium4litigation.com/

Forgot link

yea... no they don't.

You have at best a mixed opinion that its good enough and the extra cores are worth the trade off for most reviewers

gaming reviewers look at the 1700 and oc, and wonder if future games will work better as they move to 4 core + and newer apis

and work related reviewers who pretty much tell you intel is dead for workstations.

all of them have the caveat of bioses are shit, and windows may need patching.

well we are several bios patches in an 1 possibly major windows patch for ryzen, new reviews keep coming, and many of the big reviewers refused to re test when a new bios hit before launch.

Not to mention most reviewers got asus motherboards who literally shit the bed so hard Sup Forums went from they are great to never trusting them again.

While I do agree, shit is suspect at best, we do know ryzen is going to jump in performance each time higher speed memory is used, what's good to know is cas latency doesn't matter much.

youtube.com/watch?v=L4K7eIEAJx0

while that's true, the in game benchmarks are so fucking inconsistent across the board, who would ever trust that?

>it's another "no name youtube channel somehow gets better benchmark results than everyone else" episode

Well shit and i was really thinking about waiting to get the ryzen 5. well fuck it ill just get what i was planning on
Intel i5 7600k or 6700k overclock it with a EVGA GTX 1080.
NZXT accessories all the way.

and he is baseing this off a simulated cpu with 2400mhz memory... what a great opinion he has.

lol?

That guy in the video has been proven to be a moron repeatedly. And you're just as stupid for buying his bullshit.

The 1600 is way better than any i5.

no, the reviewers used shit ram cause the mobos to begin with were shit and need bios updates
fucking retard AMD and Intel fanboys crying corruption in youtube "media"

No, the "everyone else" was using 2133 and 2400 DDR4 on Ryzen while using 3200 on Intel.

actually the most corrupt reviews I found were from Arstechnica and other tech journalism shit sites, youtubers are actually much better sources for a good idea of what the baseline is by comparing several reviewers compared to just reading tech shit propaganda

I know the youtubers are better
generally the actual media is way more corrupt, in youtubes case it's usually just enthusiasts or uninformed idiots

Anandtech and Phoronix were good as usual.

Youtube was still full of people who had no clue what they were talking about, even if they were probably less bias on average.

except it's not, but i guess you're too fucking retarded to see the flaw in comparing a 8c16t chip with a 4c8t one in tests that aren't multithreaded.

go look at the simulations they did for 4c8t ryzen vs 7700k, it's very close and they're pricing it at HALF the cost of intel.

DELETE

>literally who tier reviewer
Do you have anything intelligent to say about the methodology here being off, or are you just a mindless shill?

Activating Intel Management Engine, anyone who speaks ill towards Intel will have their personal information given to the lowest bidder come Dec 9th.

surprising that you're not a tripfag cuz you sure are as smart as one

When are we gonna get a Ryzen benchmark with something like this?

Why don't we use timedemos for benchmarks anymore instead of this "teehee I'll play for 5 minutes on this system and then 5 minutes on this system" with unequal results?

because these new DRIBLE AYY games cant do that. also they usually have a specific part of the game that they run to make it as even as possible.

>0.001%
>anything than isnt anus
nice try merchant

>CL19

now tell us how that effects the performance by listing a benchmark

higher speed is useless with higher latency

It's irrelevant, the margin of difference is minimal enough that you'd never notice unless somebody forced you to run benchmark comparisons

Wrong. Lower latency does help improve slower speeds.

If you have to raise speed by raising latency it may be running faster, but if it's waiting much longer for commands it doesn't matter how fast it's going

got a asus prime x370 pro which with stock bios runs my 3200 ram at 2133mhz. the latest bios update allows for 2666mhz somewhat stable but fucks up the cpu temp sensor adding 20celsius and sometimes the system freezes.

for anyone interested dont touch the asus prime its a giant unstable fuck up. dont expect your ram to work anywhere near its labeled performance stable.

post a benchmark

thats not true
if you up the latency and increase the speed you get higher performance on ryzen because the L3 is quite big

theres an old linus video about ram timing

it really doesnt matter in real world applications due to the speeds you have to run the dimms at to get the lower latency

inversely, speed doesnt really matter either after a point because you have to set such a high latency

Hey, retard, how about you post the formula for true latency and examples if you want to prove your point instead of anecdotal bullshit?

Latency in clock cycles / frequency = latency in actual time

Latency of DDR4-4266 @ CL19 = Latency of DDR4-3200 @ CL14.25 = Latency of DDR4-2400 @ CL10.7

infinity fabric needs fast ram and a better bios

wait a few months and the 7700k wont look nearly as good

>AMD system with the same RAM and a cheaper board basically matches Intel
>AMD CPU costs marginally more than that particular Intel CPU
>AMD system has twice the cores for superior longevity
>C2Q Q6600's were generally capable medium settings 1080p machines until 2013
>C2D E6600s were practically useless by comparison at the time of Sandy Bridge.

You. You have the smell of an immature child with no life experience.
Would you happen to be a virgin?

When you do the simple-ass math you see that Zeppelin is running it's inter-CCX uncore at 1.066-1.2Ghz at the stock rated settings. 45nm (Intel) designs could run their cache-to-RAM-communications fabrics at up to 4Ghz, almost eight damn years ago.

Frankly I see no reason why, except for power characteristics, the uncore cannot be unlocked to run at IMC speed. Maybe a new Ryzen revision with higher TDP will be released later this year with that capability at stock. Maybe they are just saving it for Zenver2 along with slightly better clock potential at the same/similar power draw, which would be a phenomenal win.

In any case if/when boards that can handle 3600 cl16 are commonplace, this speed is ~50% increase in cross CCX bandwidth and ~50% reduction in cross CCX latency.

Yeah, Q6600s are a good example.

When they came out, they were worthless. Single core performance was significantly lower and no games really used more than 2 at the time.

2-4 years later they had great resale value.

>Frankly I see no reason why, except for power characteristics
Because cross-CCX cache hits aren't nessisary for most workstation applications so they get much higher perf/watt. Gaming is different.

Q6600 were fine if overclocked, the stock 2.4ghz clock could easily be clocked up to 3.4-3.7ghz with a non-garbage motherboard

Correct a long time ago. Incorrect in 2017.

If I had to decide, Clannad

Also, stock clocks for that image.

Does this mean that I should delete my 8GB DDR4 2400MHz kit and buy 3000+ MHz 16GB kit on my Ryzen 7 1700 CPU?????

3000+ kit isn't guaranteed to work on Ryzen. Those are clocks verified for Intel XMP.

You need to wait until mid April to get a good idea of what RAM to get, probably.

How did he get the Ryzen 5? Oh simulat-

simulated shillmarks

crucial.com/usa/en/memory-performance-speed-latency

>2400MHz

DELET

Ram speeds have little effect on I7 performance.

Performance began to be severely limited by SSAO, post processing, draw call requirements and more robust memory usage by about 2012.

A 3.4Ghz 6700 using a 770 or thereabouts could only manage ~45FPS even with shadow and non GPU post-processing turned down or off (GPU dependent settings did not matter much). Lowered resolution wasn't an answer either. I shudder to think about using one in modern games.

less power draw at same clock
slightly less heat
easier to get it to 1800x level

downside no included cooler= ~$40

that reminds me, didn't they promise wraith max or something for 1700x/1800x?
It's hell to get AM4 coolers now.

INSHILLS BTFO

half or less improvement compared to Ryzen

dat Crysis and GTA5 shutterfest on 7700K

Four cores is enough.
It's ENOUGH!!!
Single thread performance is all that matters!

...

>1700 with -3200 MHz RAM versus 7700K with 3200 MHz RAM
>INTEL IS FASTER HAHAHAHAHAHAHAHAHAH

ok

7700K is a stuttering piece of shit, news at 11