Discussion: AMD Ryzen

Page 6 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Another case where an 8 core is within 33% of maxing out a GPU even at 1.2Ghz.
DX12 again showing efficiency at using more cores.

1461783532QJ6xQ9Rbp6_4_5.png
 


Which shows again that 4-core will remain the best choice for gaming. In the first place almost everyone agrees that AoS is a best-case game, not all the games will see same benefit from extra cores/threads. In the second place they are comparing chips at same frequencies instead using stock frequencies.

The best case happens when both chips are clocked at 1.2GHz, which doesn't represent a gaming setup. Even then the 5960X is only 34% faster, which is just the difference in stock clocks between the 5960X and the best quad-core (the 4790k).

In fact, when both chips are overclocked at same 4.5GHz the quad-core ties with the octo-core even under DX12.

Therefore a higher clocked quad-core will tie or win over a lower clocked octo-core.
 


Does it mean you are at a disadvantage with more cores in games? Apart from the lower clock.
 


That benchmark compares chips with same microarchitecture (Haswell), therefore the only differences are clocks and the amount of cache, but as the 4.5GHz benchmark shows the quad-core and the octo-core perform the same (difference is less than 1%), therefore extra amount of cache is not providing any advantage to the octo-core.

However, I think Zen will have less IPC than Haswell. Taking Excavator scores and multiplying by the 40% claimed by AMD I got Sandy Bridge level of IPC. Add to that almost everyone is expecting AMD to achieve lower clocks <4GHz due to using the 14LPP process node. Thus I think higher clocked Skylake or Kabylake 4C will be the favorite choice for gaming over an 8C Zen at lower frequencies.
 
My speculation is that AMD will once again fight with more cores. So the Zen 6C/12T will be priced against the Intel 4C/8T. And the Zen 8C/16T against the Intel 6C/12T. They can do that because there is no GPU wasting space in the die. So it is important so see modern games actually taking advantage of more cores so that AMD can compete even if there is a clock speed and IPC disadvantage.
 


What hype? XD

The fella in the video doesn't know what IPC actually means and he is just re-capitulating some stuff from rumors and marketing slides. Meaning, the same things we all know in here; I think.

In any case, the bit that was interesting is "they'll have them by Q4 2016". He is either mis-interpreting "mass production" with "mass availability" or just bundling them as the same thing.

Cheers!
 


The author contradicts himself; he first says it is "the Zen successor", then latter forget and writes about "7nm Zen desktop CPUs". In any case it is evident this Starship processor uses Zen+. It looks as Zen is going to disappoint a bit and the hype about its successor Zen+ already begun. Some remarks about Starship:

(i) 48-cores on 7nm doesn't looks good. We are waiting 64-cores on 16nm from competitors.

(ii) No one can believe that Globalfoundries will have 7nm ahead of everyone else.

(iii) Assuming AMD is using the same multidie approach as Zen, this means that the 7nm Zen+ die would have 12-core, which agrees with the original (missed) target for Zen.
 


Producing some few test chips on your lab is one thing. Producing ~600mm² commercial chips in mass is another.

IBM roadmap server
14nm --> 2017
10nm --> 2018
7nm --> 2020

openpower-power-roadmap-new.jpg
 
I think it's much like the way the internet has worked for some time, pics or it didn't happen. Or in this case benchmarks or it didn't happen. 😛 Not just some one off es chip with a demo game that's never going to see the light of day, real bench's with real existing gpu's using real games or other software depending on the intended purpose.

The hype stuff is old, from both amd and intel. Pretty charts don't mean squat, it either runs programs or it doesn't. Theory doesn't get it done. We all know the campfire stories from amd, they're going to dominate the market with a 512 core 1024 thread chip plugging along at 300mhz per core etc. hehe.
 


If you read the link, they showed a whole waffer of chips produced in 7nm. That is not what you imply to be as some sort of "lucky lab experiment".

Plus, that roadmap does not do the same linking you do there. It only displays: "10nm and 7nm between the 2018 and 2020 timeframe". They could perfectly skip 10nm if they nail 7nm before 2020, which I see they are trying to do. Like I said, that "lab prototype waffer" is already 1 year old. Think about that for a moment.

Cheers!
 


<Mod edit - snark>

Scientists at IBM Research have made the first working prototype computer chips with circuit features as small as 7 nm. The milestone was achieved in collaboration with GlobalFoundries, Samsung and researchers at the SUNY Polytechnic Institute Colleges of Nanoscale Science and Engineering.
Click to expand...

Industrial production of server-like chips is not expected before 2020. As the IBM roadmap shows 14nm next year, 10nm in 2018, and 7nm in 2020. They have said that 7nm comes after 10nm

7nm.png


If you have a link where IBM has stated that will skip the 10nm node and will go directly to 7nm or are trying to do that, this is the time to share with us.

Their roadmap agrees with ITRS (International Technology Roadmap for Semiconductors) which predicts 7nm for around 2020 as well.

https://en.wikipedia.org/wiki/7_nanometer

It is needed to understand the difference between full industrial production, development, and research. Foundries labs are always many years ahead of industrial production

rd.png
 
All your posts were deleted because this thread is mean't to discuss Zen EXCLUSIVELY. If you see a few of us mods deleting posts but don't understand why, please PM one of the moderators for assistance. Don't ask why your post was deleted on the discussion thread itself because this thread is not mean't for discussion other than Zen.

Thank you!
 


AMD didn't dismissed it. They didn't comment on boards and costs. From the above link: "AMD stated that the "Zen" processor platform is on-track, and the company will not comment on "customer specific board-level solutions."

The company that commented on the rumor was ASMedia, but their answer to the Zen issue was weak. Their designs could have passed formal certification and still have the long-wire USB issue requiring the extra board circuity that will increase costs.

Moreover, it is evident that there is a delay. The AM4 mobos were supposed to appear time ago, much before Zen. In fact ASMedia has claimed that Zen chipset has not still started production: "The AMD Zen chipset will be in production by the end of the year and will go on sale in 2017."

http://techfrag.com/2016/06/22/amd-zen-chipset-price-issues/

Now, it would be glad to know how Zen could be released in Q4 2016 if the chipset will be for sale in 2017. I claim again that Zen has been delayed to 2017.
 


Bolded mine.

AMD has no issues with their chipset (according to what I have read) and it seems the problems stem from some partners and not all.

I still classify it as a rumor and salt is required, since AMD has stated otherwise.

Cost is something of another conversation.

Cheers!
 
I see zen is floating on a lot of rumors. Including estimates and rumors on performance.
If the flagship Zen 8core-16thread performs 50% or more of 6900K for $500-$600, it should be a good deal.

I suppose Zen will come with the wrath stock cooler?
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom