AMD RYZEN Demo Event - Beats $1,100 8-Core i7-6900K, With Lower TDP

[Brusfantomet]

AMD stock seems to be down 2% in after market trading as of writing this, witch i find weird.
Most people would consider the performance demonstrated from zen to be good, but maybe the market is taking the numbers with a "truckload of salt"?

except for this guy

https://en.wikipedia.org/wiki/Jim_Keller_(engineer) originally from DEC done work for AMD and apple


Lets wait to zen hits the actual market its not like AMD has used PR before ...

anyone else remember the Althon XP's....

Think he was referring to graphical designers.

 

[Assimilator]

You're missing the bigger picture:
It will cost LESS than a i7-6900k as well.

AMD hasn't released any pricing information on Zen/Ryzen yet, so I'd like to know what evidence you're basing this statement on.

And you are just using a single application at the time?
There are OS subsystems, multiple drivers, a web browser and some other stuff you are running, perhaps an anti-virus program and a pile of other "hidden" crap running in the background, in addition to the task you are "working" on. Even if your little game only uses 3-4 threads, more cores will help reduce stutter and other problems, even if it doesn't show up in average framerate benchmarks. Anything less than 6 (real) cores for a power-user today is not going to cut it.

If a "pile of other hidden crap" is slowing down your PC, it's almost certainly not because it's tying up your CPU, it's because it's reading from or writing to a mechanical disk. In which case you can be running a 128-core CPU and your PC as a whole is still gonna be slow. CPUs and operating systems are extremely good at task scheduling and context switching, precisely because that's their main use case.

And what is a "power user"? I consider myself a power user because I generally have half a dozen browser windows, with dozens of tabs in each, open - plus at least two instances of Visual Studio - and I play a game at the same time, and my 4-core 4-thread i5-3570K is still perfectly adequate for that. Power users don't need more cores/threads, people who do video rendering do, and that is a very small percentage of the market that explains why core counts have stagnated and higher core counts command higher prices.

Let's be honest, AMD has been trying to force the "MOAR CORES BETTAR" Kool-Aid down our throat for years and it hasn't worked out, for the simple reason that writing code that takes advantage of parallelism is difficult and hence the majority of apps around (old and new) don't do it. High core counts are and will remain a niche market for years to come.

 

[LucidStrike]

I am happy. Even though I am not going to build a new AMD system any time soon, the introduction of more competitive product will force the price of intel's HEDT down. If in one year I can upgrade to 6950X with $500 I will be thrilled!

Unless you already have an LGA 2011-3 board, it doesn't make sense not to at least consider switching to a comparable Zen chip.

 

[LucidStrike]

that's what people said with 7970 as well. at this point i will just waiting until the actual hardware hit the market.

That said, it's exactly the approach AMD 'just' took with the RX lineup, focusing on gaining marketshare with high price-performance ratios.

 

[renz496]

It doesn't need to be as fast as the 1060 in every title, average is a better indicator of speed. You can go look anywhere, it's not just hardware canucks re reviewing after AMD's new driver.

hardware canucks test was not made using latest Relive driver either. it depends on what tittle being tested in the suite. one of TPU latest GPU review that include both RX480 and GTX1060 (done in mid november; the test on hardware canucks also using AMD november drivers) the relative performance chart still put 1060 6GB ahead of RX480. and the changes from the new driver is not that big. from launch driver to relive both TPU and guru3d shows average around 2-3% performance improvement for RX480. and i also check the numbers from 1060 and RX480 launch drivers versus much more recent one. while RX480 did gain more performance on much recent test the same also happen with 1060. so RX480 is not the only card gaining more performance. hardware canucks chart probably going to paint a different result if they include a few new tittles that favors nvidia cards such as Watch Dogs 2 and Dishonored 2.

 

[eidairaman1]

Unless you already have an LGA 2011-3 board, it doesn't make sense not to at least consider switching to a comparable Zen chip.

That said, it's exactly the approach AMD 'just' took with the RX lineup, focusing on gaining marketshare with high price-performance ratios.

Do not double post, use the multiquote button

 

[renz496]

Polaris 10 doesn't have double rate 16bit FP feature while both PS4 Pro and Vega 10 has double rate 16 bit FP feature.

Native 16bit FP support is important for incoming Shader Model 6.0.

AFAIK FP16 support in games is not that new. if anything using FP16 probably going to reduce image quality. on mobile (like SoC) the reason to use FP16 is to save power and bandwidth. can you provide the link talking about the importance of FP16 in upcoming SM6?

 

[P4-630]

Wow, seriously? Do you know which ones they are?

GTA V, CAPS showed it once with his xeon.

 

[DeathtoGnomes]

There were a few of us that watched the livestream from Discord. Sad more of you all were not there to comment.

 

[bug]

Hm, AMD is at it again.
Nice play on "Zen with no boost vs intel with boost". When all cores are being used, intel doesn't boost either (except for a short while - a few seconds, till it gets hot).

But overall a solid showing.

 

[R-T-B]

8 core 16 threads is finally becoming mainstream. This is definitely good news for everyone. We have been stuck at 4 core 8 threads for far too long.

Agreed, but I expect a 6-core mainstream Zen before 8-core gets there. Still good.

Unless you already have an LGA 2011-3 board, it doesn't make sense not to at least consider switching to a comparable Zen chip.

His system specs are literally a click away, you know?

 

[YautjaLord]

Aside from that Handbrake graph for both ZEN & - if my memory serves me well - 6900k, everything else seemed like "What, where's fps counter, no Lisa, Mrs. Su, what sh!t 68fps in 4k in Star Wars Battlefront Rogue One DLC you talk about? Are you guys serious?" Seemed like more of a hype train, but i digress. I'm still hungry for AM4/X370 mobos & benchmarks, oh & i don't buy that Blender demo on ZEN vs 6900k sh!t either. Waiting for you guys to actually benchmark the beast yourselves, man i am really hyped.

Other than that - it's good that AMD gets back into competition? Bring on the benchies, who's gonna bench this thing? Bring Kaby Lake & Skylake to compare, 6900k is Skylake right?

 

[bug]

Other than that - it's good that AMD gets back into competition? Bring on the benchies, who's gonna bench this thing? Bring Kaby Lake & Skylake to compare, 6900k is Skylake right?

6900k is actually Broadwell, but performance-wise, Skylake isn't any faster.

Edit: 6990k is actually 6900k

 

[YautjaLord]

6990k is actually Broadwell, but performance-wise, Skylake isn't any faster.

Now i'm completely confused. They pitted ZEN vs 6990k? I heard something like 6900k. Thanx regardless, maybe it's my hearing & fact that i remember more what CPUs were used in Dota 2 while streaming. 6700k was OC'd to 4.5GHz & struggled (hype?), ZEN & 6900k(?) ran smoothly. And in AMD fashion the bald guy engineer(?) said 6900k cost twice more than ZEN. lol 1100$ or something like that?

P.S. ASUS, MSI & Gigabyte still don't have AM4/X370 mobos in their mobos list. I checked. lol

P.P.S. Loved the ZBrush demo with ray-tracing & 53 million polygons models stuff.

 

[P4-630]

 

[R0H1T]

Now i'm completely confused. They pitted ZEN vs 6990k? I heard something like 6900k. Thanx regardless, maybe it's my hearing & fact that i remember more what CPUs were used in Dota 2 while streaming. 6700k was OC'd to 4.5GHz & struggled (hype?), ZEN & 6900k(?) ran smoothly. And in AMD fashion the bald guy engineer(?) said 6900k cost twice more than ZEN. lol 1100$ or something like that?

P.S. ASUS, MSI & Gigabyte still don't have AM4/X370 mobos in their mobos list. I checked. lol

P.P.S. Loved the ZBrush demo with ray-tracing & 53 million polygons models stuff.

Well there's no 6990K, although there was the AMD/ATI 6990, it's just a typo from the poster.

 

[ratirt]

This ZEN looks pretty promising. Not sure if I should get one when it shows up on the market or wait a bit longer. Seeing what it can do is great although waiting for reviews and more benchmarks might be wise idea to clarify what that CPU can really do. Anyway I think I'm getting an AMD platform Grrr I can even fill that chill crawled up my spine when thinking of this I really wanna go with AMD and try this thing with all features it gives

Just an add.
Finally maybe we will get mainstream 8/16 c/t instead 4/8 Gotta sell my still good 3770k

 

[TheinsanegamerN]

Hm, AMD is at it again.
Nice play on "Zen with no boost vs intel with boost". When all cores are being used, intel doesn't boost either (except for a short while - a few seconds, till it gets hot).

But overall a solid showing.

My ive bridge i5 has 0 issue maintaining it's maximum rated boost speed on all four cores without any tinkering on my end. A halfway decent cooler will keep those chips boosting 24/7.

 

[bug]

My ive bridge i5 has 0 issue maintaining it's maximum rated boost speed on all four cores without any tinkering on my end. A halfway decent cooler will keep those chips boosting 24/7.

What i5 exactly and what "boost" frequency does it maintain?

 

[YautjaLord]

Have slightly mixed feelings about the showcasing of BF1 as well, meaning - couldn't for a f*ckin' life of me see the actual fps counters/graphs/stats/etc... - that one can be called a con(?), but the fact guy said both rigs had Titan X[P] in 'em should indicate something. And the fact it ran that not yet released Rogue One SW Battlefront DLC (or any other game) in 4k - good news. Right? Overall, I enjoyed it more with occasional "wtf? you guys serious? dafuq y'all talking about?" moments in-between.

P.S. I see 4k LED IPS monitors in my PC store drop prices, time for ZEN-based, air-cooled (for now), 4k gaming/Blender'ing rig. Yeah, i use this awesome, awesome app for some project i have for Carma: R & Carma: Max Damage. Can rival 3DSMax actually & you can't beat the awesome price of free. lol

 

[HD64G]

hardware canucks test was not made using latest Relive driver either. it depends on what tittle being tested in the suite. one of TPU latest GPU review that include both RX480 and GTX1060 (done in mid november; the test on hardware canucks also using AMD november drivers) the relative performance chart still put 1060 6GB ahead of RX480. and the changes from the new driver is not that big. from launch driver to relive both TPU and guru3d shows average around 2-3% performance improvement for RX480. and i also check the numbers from 1060 and RX480 launch drivers versus much more recent one. while RX480 did gain more performance on much recent test the same also happen with 1060. so RX480 is not the only card gaining more performance. hardware canucks chart probably going to paint a different result if they include a few new tittles that favors nvidia cards such as Watch Dogs 2 and Dishonored 2.

You didn't check the last page in the HW Canacks reveiw eh? Because there is a comparison table there that perfectly describes the benefit that RX480 had vs the GTX1060 from the June2016 until November's driver release. So, while you are correct that 1060 gained also from newer drivers, RX480 gained much more and so it closed the gap to just over 2% for the games HW Canucks tested and to just over 4% for the gamelist TPU's @W1zzard uses. And it is an established fact now that CGN clearly wins the nVidia's arch there with the gap being visible for almost all games.

 

[ADHDGAMING]

if they have better solution they will charge for it. heck even when they not they still charge 1k for 9590.....initially.

Sue made it a point to mention the Intel Chips price for a reason .. Pay attention she was dropping clues. it will certainly be cheaper.

 

[Assimilator]

Sue made it a point to mention the Intel Chips price for a reason .. Pay attention she was dropping clues. it will certainly be cheaper.

You have no understanding of how companies work, do you?

 

[dozenfury]

FYI, another site (AT) they did some CSI-ish examination of the presentation, because scores were suspiciously not lining up with where they should be in that Ryzen demo and they noticed something. In the AMD Ryzen demo, the video shows that the render sample value is changed to 100 instead of the default 200, which dramatically changes the scores. So if you download the demo from AMD and just hit go, you will have the default 200 render sample value and will have a slow score. You will need to change the render sample value to 100 to get an accurate comparison, and that's assuming they also didn't change anything else.

Incredibly sneaky, and more reason to wait for real third-party unbiased benches. I'm very hopeful it turns out to be good, but stuff like this raises some major red flags especially considering AMD's history of letdowns.

 

[bug]

FYI, another site (AT) they did some CSI-ish examination of the presentation, because scores were suspiciously not lining up with where they should be in that Ryzen demo and they noticed something. In the AMD Ryzen demo, the video shows that the render sample value is changed to 100 instead of the default 200, which dramatically changes the scores. So if you download the demo from AMD and just hit go, you will have the default 200 render sample value and will have a slow score. You will need to change the render sample value to 100 to get an accurate comparison, and that's assuming they also didn't change anything else.

Incredibly sneaky, and more reason to wait for real third-party unbiased benches. I'm very hopeful it turns out to be good, but stuff like this raises some major red flags especially considering AMD's history of letdowns.

This is the reason I mostly avoid these articles altogether. In most cases it avoids me the unrealistic hype.