• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** AMD "Zen" thread (inc AM4/APU discussion) ***

CAT-THE-FIFTH said:
Well,when I mean that I hope it is:
1.)Better clockspeeds,so hopefully the improved 14nm process being used means they can get closer to 4.5GHZ(instead of the 4.1GHZ maximum now).
2.)The memory controller is improved in both hardware and firmware,so they can get better clockspeeds and memory compatability.

Now things which would be nice,but hard to say if it will come with the refresh:
1.)Noticeably better IPC.
2.)Improvements to the bandwidth between the CCX units.
3.)Better AVX support.
Click to expand...

When it comes to the efresh on the same platform I think a lot of people will just get the new CPU, once you own mobo+ ram. It's not expensive just to get CPU for £200-350.
 
So Destiny 2 launch version has been tested:

http://www.guru3d.com/articles_pages/destiny_2_pc_graphics_performance_benchmark_review,6.html
http://gamegpu.com/action-/-fps-/-tps/destiny-2-test-gpu-cpu

Looking at the results,Ryzen seems to be generally better than Haswell in the title(above the Core i7 4770K and Core i7 4790K),and around the Core i5/Core i7 6000 series although not much separates the CPUs. The higher clocked Core i5/Core i7 7000 and 8000 CPUs seem to push ahead though.

Not bad performance!!
 
CAT-THE-FIFTH said:
So Destiny 2 launch version has been tested:

http://www.guru3d.com/articles_pages/destiny_2_pc_graphics_performance_benchmark_review,6.html
http://gamegpu.com/action-/-fps-/-tps/destiny-2-test-gpu-cpu

Looking at the results,Ryzen seems to be generally better than Haswell in the title(above the Core i7 4770K and Core i7 4790K),and around the Core i5/Core i7 6000 series although not much separates the CPUs. The higher clocked Core i5/Core i7 7000 and 8000 CPUs seem to push ahead though.

Not bad performance!!
Click to expand...

Something weird going on there. The 1300x is right behind the 1800x....
 
gavinh87 said:
Something weird going on there. The 1300x is right behind the 1800x....
Click to expand...

It seems to like clockspeed and cache. The Ryzen 5 1600X and Ryzen 7 1800X both have 16MB cache and the highest clockspeed of the Ryzen CPUs in the chart.Haswell E and BW-E also do quite well too.

Edit!!

Look at the Core i5 6600K and Core i7 6700K. Both are more or less the same in the game,with clockspeed and cache probably accounting for the slight differences I suspect.
 
Are those AMD's numbers? Because they picked the lowest scores possible for the i7-8550U https://www.notebookcheck.net/Intel-Core-i7-8550U-SoC-Benchmarks-and-Specs.242108.0.html

I'm curious if they actually get those numbers with a similar 15W TDP as the 8th Gen Intel APUs or if it's AMD marketing in action again. If they're actually competitive on TDP then they might have a winner, but I somehow doubt it given they have to use Vega for the iGPU part.
 
AndreiD said:
Are those AMD's numbers? Because they picked the lowest scores possible for the i7-8550U https://www.notebookcheck.net/Intel-Core-i7-8550U-SoC-Benchmarks-and-Specs.242108.0.html

I'm curious if they actually get those numbers with a similar 15W TDP as the 8th Gen Intel APUs or if it's AMD marketing in action again. If they're actually competitive on TDP then they might have a winner, but I somehow doubt it given they have to use Vega for the iGPU part.
Click to expand...
Thing is, Vega is clearly heavily overclocked out of the box. Just like Ryzen is very power efficient at around 3 GHz but not at 4 GHz, IGP Vega will likely be running at clock speeds that make it much more efficient than the discrete Vega cards.
 
The question is if Zen + Vega can match the 15W TDP Intel 8th gen APUs given the pretty significant process disparity between GloFo's "14"nm and Intel's 14nm processes. I would be extremely impressed if they somehow managed to achieve that kind of performance in a similar TDP range, but I fear they might be comparing a higher TDP part to the 15W Intel ones.
Either way, I'm eagerly waiting for reviews, because if they are competitive with similar power consumption numbers as the 8th gen Intel APUs, they will have a winner on their hands. If on the other hand they have a ""15""W TDP part like some of their older APUs...
 
DragonQ said:
Thing is, Vega is clearly heavily overclocked out of the box. Just like Ryzen is very power efficient at around 3 GHz but not at 4 GHz, IGP Vega will likely be running at clock speeds that make it much more efficient than the discrete Vega cards.
Click to expand...

Yep,and here is more info from AT and Hexus:

https://www.anandtech.com/show/1196...md-apus-for-laptops-with-vega-and-updated-zen
https://hexus.net/tech/news/cpu/111473-amd-announces-ryzen-7-mobile-apus-vega-graphics/

Considering how efficient Ryzen is,especially at lower clockspeeds,it seems plausible. The Stilt has done extensive testing and shown that upto 3.3GHZ Ryzen is pretty efficient:

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/

8Rch6JF.png

850 points in Cinebench 15 at 30W is quite telling. Or not telling, but absolutely massive. Zeppelin can reach absolutely monstrous and unseen levels of efficiency, as long as it operates within its ideal frequency range.
Click to expand...

Its no wonder when you look at the GF 14NM process - its orientated more towards mobile,and lower clockspeeds,so its not surprising it does better at lower clockspeeds.

It also fits into what we see with the AMD server parts which seem to pretty efficient too.

AndreiD said:
The question is if Zen + Vega can match the 15W TDP Intel 8th gen APUs given the pretty significant process disparity between GloFo's "14"nm and Intel's 14nm processes. I would be extremely impressed if they somehow managed to achieve that kind of performance in a similar TDP range, but I fear they might be comparing a higher TDP part to the 15W Intel ones.
Either way, I'm eagerly waiting for reviews, because if they are competitive with similar power consumption numbers as the 8th gen Intel APUs, they will have a winner on their hands. If on the other hand they have a ""15""W TDP part like some of their older APUs...
Click to expand...

LOL,the GF 14NM process is mobile orientated - its based on a Samsung 14NM mobile orientated process. Have you seriously just ignored all the testing of desktop Ryzen by people like The Stilt,or even AT themselves who showed the lower clocked AMD server parts being quite competitive with Intel lower clocked server parts in performance/watt too??

Plus trying to name a BD based CPU in this thread - what one based on a very inefficient uarch made on GF 28NM which performed worse in many cases than GF 32NM SOI??

Or the fact BR is being compared to 14NM Intel CPUs so was more or less two generations out of date on the process node side.

I mean why are you trying to name drop Bulldozer in this thread??

Compare the Ryzen 7 1700 to an FX8350 and come back to me.
 
Last edited:
Yup more results needed, as I can't see Ryzen +Vega performing that we'll and keeping below 15w.

We all know that Vega for the discrete market was pushed quite hard to get the results AMD wanted, we will just have to wait for more results to see if the new APU'S are power efficient at the same time as being fast, or will it be either or.
 
@CAT-THE-FIFTH GF 14nm is the same process as Samsung's 14nm, not just based on it, but I'd rather not gobble up AMD's marketing claims which as history has shown are often pretty questionable.
I reckon we'll see 3rd party reviews in a few days and we'll see if AMD's claims are on point or not. Epyc is another can of worms when it comes to AMD claims and what actual reviewers like servethehome got in testing.
 
jigger said:
That's desktop quad core and discrete graphics card levels of performance for 15-25 watts?
Click to expand...

To put in context,over on Reddit someone noted the 2700U scores beat an FX8350. Regarding the IGP performance it seems that 1 CU is disabled so all the rumours of the IGP having 704 shaders is correct. So its 640 shaders,but as long as the RAM is in dual channel the performance looks very solid.
 
One of the reasons I'm weary is that most of their performance numbers include Intel counterparts while for battery life they only compare with the previous gen FX-9800P:
https://images.anandtech.com/doci/1..._graphics_press_deck-legal_final-page-038.jpg
That's usually pretty telling for AMD marketing since they'd usually pounce on the occasion to compare with Intel SKUs.

But of course it's best to wait for 3rd party reviews.
 
AndreiD said:
@CAT-THE-FIFTH GF 14nm is the same process as Samsung's 14nm, not just based on it, but I'd rather not gobble up AMD's marketing claims which as history has shown are often pretty questionable.
I reckon we'll see 3rd party reviews in a few days and we'll see if AMD's claims are on point or not. Epyc is another can of worms when it comes to AMD claims and what actual reviewers like servethehome got in testing.
Click to expand...

Dude,you don't need any marketing when we had 6+months of Ryzen CPU testing,and AT tested Epyc and you might want to bury it but AT showed it to be a solid performance and more importantly performance/watt was competitive.

The Stilt,who has been around for years and is widely respected showed on a launch sample of the Ryzen 7 1800x,that the CPU is very efficient at lower clockspeeds. I mean look at various reviews too on the Ryzen 7 1700 at it stock clockspeeds.

The fact of the matter is you can deny it all you want,but Ryzen is efficient and there is plenty of evidence to show it will performance fine at lower TDPs just from user testing. Also the GF process is not exactly the same as they licensed the base process and its moved on a different path.

Plus its incredibly pointless to even compare Ryzen to any BD based CPU. BR only got down to some level of efficiency as they paired down the amount of cache,etc to save on power and transistor budget. Plus the GF 28NM node was a bulk process,it actually regressed in certain metrics over the previous generation 32NM SOI node,meaning AMD couldn't even hit the clocks they expected. They were then fighting Intel 22NM and 14NM processes which had Finfets and those help in mobile situations.

If anything dGPU performance is going to be more dependent on whether OEMs will ship the laptops with dual channel RAM or the ability to do so. This was the singular reason why AMD IGP performance in reality might have varied more than anything else and according to the AT article it looks like dual channel RAM is on the table for the launch models.

AndreiD said:
One of the reasons I'm weary is that most of their performance numbers include Intel counterparts while for battery life they only compare with the previous gen FX-9800P:
https://images.anandtech.com/doci/1..._graphics_press_deck-legal_final-page-038.jpg
That's usually pretty telling for AMD marketing since they'd usually pounce on the occasion to compare with Intel SKUs.

But of course it's best to wait for 3rd party reviews.
Click to expand...

LMAO,you are clutching at straws!! Why are you so spooked at the launch?? All you are trying to do is bury this in the most negative way.
 
CAT-THE-FIFTH said:
To put in context,over on Reddit someone noted the 2700U scores beat an FX8350. Regarding the IGP performance it seems that 1 CU is disabled so all the rumours of the IGP having 704 shaders is correct. So its 640 shaders,but as long as the RAM is in dual channel the performance looks very solid.
Click to expand...

Will be interesting to see how the market responds and how AMD price chips. I know a lot of people that would be very interested in a note book with this kind of performance and decent battery life.
 
jigger said:
Will be interesting to see how the market responds and how AMD price chips. I know a lot of people that would be very interested in a note book with this kind of performance and decent battery life.
Click to expand...

It looks solid - my main concern is that AMD make sure the laptops,at least ship with the option of having dual channel RAM. That is the area which can potentially become a problem in regards to IGP performance. I think some arguments could be made for chassis with sufficient cooling too,but TBH,that also affects Intel too.
 
You must log in or register to reply here.
Back
Top Bottom