AMDeus Vult!
AMDeus Vult!
>1fps more than haslel
BRAVO
R
A
V
O
is that min fps ? 0.o
Yes, thats a min fps.
what happened here?
Fuck this graphs are all over the place.
IN HOC SIGNO VINCES
>haswell above skylake
wtf lol
yes, and since gamers are retarded they do not know how to properly read graphs
Nothing. Just another retarded AMD post where OP fails to realize a GPU limitation will hit way sooner than CPU bottleneck. Ryzen is 8% slower than Kaby while running on 20% slower clocks. That's exactly what correctly ran benchmarks show.
for twice the price
$500 Ryzen CPU beat by $240 Intel CPU
Just lmao, Ryzen is so fucking bad
But this is 1080p, some graphs show Ryzen being 30% behind the pack some show it being 10% behind the pack, this one shows it being 10% ahead of the pack.
simple for honor just came out, everything else is two-4 years old
also even in bf1 min/avg difference is smaller than for 7700k
IPC is equal and the clocks are only 15% lower. You were probably looking at pre-BIOS patch benchmarks. The MSI patch for example boosted performance in games by 17% on average.
Blame motherboard manufacturers
But the big thing here isn't average framerate, it's the fact that it has a massive advantage over any Intel chip in terms of minimum framerate, which is a far more important metric in most cases.
THIS IS INCORRECT IMAGE AND BENCHAMRK.
FIRST ITS NOT 640X480.
SECOND IS WHO THE FUCK CARES ITS A BROKEN SHIT GAME.
THIRD. WHO THE FUCK CARES ABOUT MINIMUM FPS WE ONLY CARE ABOUT MAX = GOOD MIN = BAD.
AMD SHILLS OUT>
Sir, you should of follow our instructions... I kindly ask you to remove that image.
25% higher minimum frames.
G O D C H I P
t. shill hoping you don't notice the min framerate difference
It is already decided. Ryzen is bad in Far Cry 3 800x600 so ITS TRASH IN GAMING DO NOT BUY OK?
Source? I can't find it.
>no source
Whatever you say AMDrone.
Two posts above yours.
DELET
20% higher minimum framerates, not 10%.
...
When was the MSI patch?
Is there a source for this benchmark?
DELET THIS
Literally who cares about minimum FPS, lol.
DELETE
ty
Last MSI update I see is from March 1st. Shouldn't reviewers have had that update?
Everyone ? Max fps is shit if it's not stable
They didn't start work on their reviews a day before launch.
That user is obviously being faceticious.
Everyone knows that minimum, 0.1% percentile, and 1% percentile are the most important benchmarks next to frametime graphs.
>AMD is quite literally always behind intel
why even bother?
intel will roflstomp them in a few months anyway
DELET
This. AVG fps is far more important than min FPS.
>1080p
Who the fuck plays in 1080p lul.
Are we finally free? is it the day when AMD finally won?
Tech City also ran For Honor. They got much better results for the 7700k that's basically a 6700k.
However it seems the only real difference in their tests is that pclab.pl picked a more complex scene in the single player to benchmark that has tons of trees.
minimum is what matters the most, if anything
dank meme.
isn't it the whole point of such tests to find the most taxing scene and confirm which part performs the worst? And apparently cheap 8c is not a meme.
>minimum is what matters the most
no, the gap between avg and minimum is what matters
smaller the better
No. What matters the most is average and variance.
I would think so, definitely.
It could also be that in the Tech Spot tests, that older board BIOS made Ryzen underperform in the less complex test.
In Tech Spots test they got a similar 133 minimum FPS but over 150 max in a less complicated scene.
Hopefully more reviewers revisit soon.
ok. Here's your 0fps cpu
This cannot be right!
Min fps is the holy grail of cpu test atm, max fps is just muh ghz. we dont live on Pentium era anymore
>he didn't notice min fps
I'm sure you like min 60fps up to max 140fps in your games
I'm torn. Minimum take rate is definitely more imporant than average frame rate if you already reach an acceptable average FPS threshold, but only some of the benches demonstrate this performance. What is going on? And don't say motherboard BIOS isssues and a memory bug.
Intel, the new waitfags.
>people should be ok with stutters and random FPS dips just like me
Window dont recognize zen hypertrading core, window treat it like real core. this can be fix by window update or bios
Jesus Rajeet at least speak passable English
>testing muh minimum in a Ubisoft game
Zero consistency
Maybe they should test Ryzen with integrated graph-
OH WAIT
Yet another value added feature that Intel gives us over AMD
windows thinks ryzen L3 cache is 128MB, it's actually 20 for one thing
still look at frametime graphs, kaby is all over the place, zen got smooth line
Its sorted by max fps and the inner bar is min fps
Other way around bud
1700 has similar performance. this is insane value. almost every reviewer noted ryzen was smoother than kabylake in spite of lower fps. if i were building a game pc today it would be with 1700
And generally speaking, higher mins = smaller gap. Yes you're right tho, you want steady framerate and many reviewers noted ryzen is steadier.
Sup Forums is if that is what Intel delivers.
>he says confidently without posting a frametime graph
Yeah, no.
Fuck intel for force my throat to their "integrated trash". Why I need to spend money for thing I don't give a fuck?
forums.anandtech.com
Apparently SMT is freaking great if it's not limited.
>X99
Its impressive given this is AMD's first production implementation of SMT and it is way better than Initel's initial attempt. Then again learning from Intel's mistakes is probably how AMD got here.
Still, the grandmaster of SMT remains IBM.
>he says confidently without posting a frametime graph
INTELLIOTS BTFO
can someone ELI5 me on what this means? What negative performance impacts happens with that error?
>crusades failed
>amd failed
Hmmm
>no i7 7700k
>no i5 7600k
>no i5 7500
>no i5 7400
AMD is strawmanning Intel by not using high value high gaming performance CPU's but instead random to try to make themselves look good.
*takes a look at Jerusalem on Google Maps*
>Israel territory
Judeo/Christian values won over islam
>paying more for less performance than Intel
>being swedish
Actually it is a Icelandic currency and we deport immigrants, unlike Swedes.
>being poojeet
Why do they always show fps? It's not a good way to measure performance. Shows how retarded gamers are.
kek
Frametimes are hard to understand.
And fps are easy to misunderstand.
GPU bottle neck
>tfw i5 4690k @ 4.4 ghz
>within 1 fps of a $500 ryzen 16 thread cpu
feels good man
"framerate"
Except that For Honor test is absurdly consistent. Most of the pentiums have similar minimums that scale with OC.
Looks like an obvious SMT issue. Look how shitty the i5-7500 with only 4 threads is.
It means that AMD's SMT gives a 50-90% greater improvement over HT when it's utilized properly.
You can see this in Cinebench where nT performance is about an extra 10% higher per core over Intel.
The 1800X is being the 10 core 6950X with 13% higher clocks on the intel part in multithreading because SMT is so good.
But in games? The games apparently are specifically optimized for HT and there's problems with the Windows scheduler, at least in win10, that are giving SMT a 8-12% performance drop.
But you can see in many games since Crysis 3 and newer get a 20-45% performance increase with HT.
If SMT were working, you'd expect a 30-85% performance increase. Which means the 4c/8t Ryzen part that's coming, even with less IPC and lower clocks, may end up outperforming the 7700k in newer games even it's optimized the same.
It says right there that the Ryzen part is like 25% less than the 6800k?
That's cheap.
Are you too stupid that you can't see that 53.900 < 69.900?
That does not explain the minimum frames being shit for intel, you moron.
is this a smt/bios fixed benchmark? are there other benchmarks for this game?
>FPU massacre
True enough, but the second you truly slam memory bandwidth (possibly latency as well x99 will pull ahead.
NOOO! THIS WAS SUPPOSED TO BE ANOTHER BULLDOZER!!!!!!
>True enough, but the second you truly slam memory bandwidth (possibly latency as well x99 will pull ahead.
True, but very very very few applications can use more than 60-90Gbp/s of memory bandwidth.
Nope seems this is still with SMT enabled. It's just the benchmarker used a more complex scene in the single player than Tech Spot did which Intels seem to stutter with.
>The most correct way of benchmarking a device is showing the worst result.
Oh ok.
Do you have any other benchmarks that show results like these or just this MOBA Ubisoft game
P-PLZ DELET