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

[CAT-THE-FIFTH]

It seems Hardware.fr has shut down:

https://translate.googleusercontent...700201&usg=ALkJrhidYMw-c4PJA7Jqlhqh9sFWncTzQw

A shame!

 

[HeX]

AMD designed Zen 2 with Ice Lake in mind...

Our plan for the Naples-Rome-Milan roadmap was based on assumptions around Intel’s roadmap and our estimation of what would we do if we were Intel,” Norrod continues. “We thought deeply about what they are like, what they are not like, what their culture is and what their likely reactions are, and we planned against a very aggressive Intel roadmap, and I really Rome and Milan and what is after them against what we thought Intel could do. And then, we come to find out that they can’t do what we thought they might be able to. And so, we have an incredible opportunity. Rome was designed to compete favorably with “Ice Lake” Xeons, but it is not going to be competing against that chip. We are incredibly excited, and it is all coming together at one point.

Forrest Norrod (Senior Vice-President and General Manager of the Enterprise, Embedded and Semi-Custom (EESC) Business Group at AMD)

https://www.nextplatform.com/2018/06/20/amds-epyc-return-to-the-datacenter-ring/


Really can't wait for next years round of CPU's, AMD really in with a chance of taking the crown! Helluva come back story.

 

[NinjaCool]

This strongly suggests what many of us have said that AMD will take back the performance crown in 2019!
It looks like 2019 is a good time to upgrade

 

[Beren]

AMD sponsored content alert.

Not saying it is lies just aware of the bias being what it is.

 

[sideways14a]

Even the most pessimistic of outlooks for AMD next year are bound to show major share take, its not quite at the point where they cannot fail (i mean this is AMD after all, masters of the screw up) but they do have an excellent chance of securing some much deserved market share.

 

[nashathedog]

i7 8700K vs R7 2700X in 1440p Gaming!

 

[humbug]

i7 8700K vs R7 2700X in 1440p Gaming!

I don't get people who make reviews like this, its as if they have no clue what they are doing.

For a start they are using the most expensive Ryzen CPU they can find when it conmen knowledge that for gaming the vastly cheaper 2600 provides identical gaming performance.
Use 1440P which everyone knows is a really bad idea for testing gaming performance on the CPU because at that resolution the game in fact becomes GPU bound, and he must have realised this during his testing because strangely half of his games are running reduced graphics settings no doubt to get around the GPU being the bottleneck.

He also ran the Samsung B-die 3200 memory at 2933, which is 133Mhz slower than i'm running my 3000Mhz rated Hynix memory on a £75 B350 board, so you could easily add another 15% performance to his results and i don't buy his excuse that his X470 couldn't run the memory as fast as the X370, if anything has changed between X370 and X470 its that memory clock speed especially on Samsung B-Die has increased, most people are getting them to 3400 / 3600Mhz on X470.

 

[Troezar]

I don't get people who make reviews like this, its as if they have no clue what they are doing.

For a start they are using the most expensive Ryzen CPU they can find when it conmen knowledge that for gaming the vastly cheaper 2600 provides identical gaming performance.
Use 1440P which everyone knows is a really bad idea for testing gaming performance on the CPU because at that resolution the game in fact becomes GPU bound, and he must have realised this during his testing because strangely half of his games are running reduced graphics settings no doubt to get around the GPU being the bottleneck.

He also ran the Samsung B-die 3200 memory at 2933, which is 133Mhz slower than i'm running my 3000Mhz rated Hynix memory on a £75 B350 board, so you could easily add another 15% performance to his results and i don't buy his excuse that his X470 couldn't run the memory as fast as the X370, if anything has changed between X370 and X470 its that memory clock speed especially on Samsung B-Die has increased, most people are getting them to 3400 / 3600Mhz on X470.

What I take away from this is that in real world useage you won't notice the difference. I have used 1440p for 7+ years. I'm only moving up in resolution going forward so the CPU won't be a bottleneck in gaming, for everything else more cores will be a benefit.

Add in cost of the platform, potential upgrades, which I think he missed the point. You don't drop a new CPU in next year but in 3 or 4 years time you drop in the fastest 7nm CPU and max out the platform lasting you another few years.

Definitely convinced me that AM4 or TR4 will be my next upgrade.

 

[CAT-THE-FIFTH]

If a game is truly CPU limited it will show at 1440p - that is my experience of playing some games which show this,as the minimums go down the drain.

 

[humbug]

What I take away from this is that in real world useage you won't notice the difference. I have used 1440p for 7+ years. I'm only moving up in resolution going forward so the CPU won't be a bottleneck in gaming, for everything else more cores will be a benefit.

Add in cost of the platform, potential upgrades, which I think he missed the point. You don't drop a new CPU in next year but in 3 or 4 years time you drop in the fastest 7nm CPU and max out the platform lasting you another few years.

Definitely convinced me that AM4 or TR4 will be my next upgrade.

Yeah i just think if testing CPU performance its best to stick with 1080P, or even 720P, testing like that is done for good reason, al lot of what this guy did makes no sense.

 

[Troezar]

If a game is truly CPU limited it will show at 1440p - that is my experience of playing some games which show this,as the minimums go down the drain.

Which games do you find are CPU limited?

 

[Troezar]

Yeah i just think if testing CPU performance its best to stick with 1080P, or even 720P, testing like that is done for good reason, al lot of what this guy did makes no sense.

I agree if you just want to test the CPU performance however it is good to see real world testing. If like to see more reviewers test how people actually use their PCs. A blind test of the two systems would also be interesting.

 

[RavenXXX2]

Heads up to anyone running the 2200G/2400G and overclocking, there is a frequency black hole between 1300Mhz+ up to 1499Mhz. On some chips this will cause a BSOD, so if you thought your chip was a crap clocker try 1500Mhz.

 

[CAT-THE-FIFTH]

Which games do you find are CPU limited?

Planetside 2 and Fallout 4(in settlements or any areas which have large concentrations of NPCs). I saw the same dips with a GTX960 and a RX470,and also with a GTX1080 at qHD.

 

[nashathedog]

I don't get people who make reviews like this, its as if they have no clue what they are doing.

For a start they are using the most expensive Ryzen CPU they can find when it conmen knowledge that for gaming the vastly cheaper 2600 provides identical gaming performance.
Use 1440P which everyone knows is a really bad idea for testing gaming performance on the CPU because at that resolution the game in fact becomes GPU bound, and he must have realised this during his testing because strangely half of his games are running reduced graphics settings no doubt to get around the GPU being the bottleneck.

He also ran the Samsung B-die 3200 memory at 2933, which is 133Mhz slower than i'm running my 3000Mhz rated Hynix memory on a £75 B350 board, so you could easily add another 15% performance to his results and i don't buy his excuse that his X470 couldn't run the memory as fast as the X370, if anything has changed between X370 and X470 its that memory clock speed especially on Samsung B-Die has increased, most people are getting them to 3400 / 3600Mhz on X470.

He's doing a direct comparison between the mainstream flagships so I haven't got an issue with that, Some people do seem to be having issues with memory speeds with the second gen Ryzens, I can't run my 3200mhz b-die memory above 3000mhz on my B350 board that had no issue running the memory at it's rated 3200 mhz with the 1600x I replaced with the 2700x. It's disappointing but not a game breaking issue as the performance difference with the speeds seems quite small in the majority of games, At least it has been for me. That said I'm currently running 3440x1440 with an RX480 so that's a much bigger limiting factor.

I didn't think much of his testing system though, As he mentioned towards the end, he was tweaking the games graphics settings to keep performance above 100 fps with the 1080ti, So that's going to exaggerate the difference more than it will for the real world gamer who's paying for that sort of hardware, the vast majority of those running a 1080ti at 1440p will be looking to keep settings as high as possible. I'm not a competitive twitch gamer so I'm happy with my fps being kept above 60 & I tweak my settings for that level of performance or a little higher for games like the F1's, Project Cars & the Dirts etc.

 

[CAT-THE-FIFTH]

BTW,it looks like China might fine the DRAM makers over price fixing:

https://forums.overclockers.co.uk/t...ng-potentially-heavy-fines-in-china.18823502/

 

[4K8KW10]

Ryzen 3 manufacturing both on TSMC and GlobalFoundries.

TSMC Begins Mass Production of 7nm Process, AMD Vega 7nm Confirmed, Zen 2 CPUs Expected Too – Production Capacity To Increase By 3 Times Next Year https://wccftech.com/amd-7nm-vega-gpu-zen-2-cpu-mass-production-tsmc/

Also, AMD can select and use Fabs of both TSMC and Global Foundries to create their next-gen processors. This is made possible by Global Foundries using similar 7nm pitches and SRAM cells that are very close in design to TSMC, allowing Zen 2 7nm processors to be developed on either Fab without major differences.

Later this year, GF will use immersion steppers to tape out its first 7-nm chip, an AMD processor. An IBM processor will follow with ASICs coming in 2019, said Patton.

GF made the size of its 7-nm pitches and SRAM cells similar to those of TSMC to let designers like AMD use both foundries. AMD “will have more demand than we have capacity, so I have no issues with that,” he said of AMD using the Taiwan foundry. via EETimes

 

[jigger]

Ryzen 3 manufacturing both on TSMC and GlobalFoundries.

TSMC Begins Mass Production of 7nm Process, AMD Vega 7nm Confirmed, Zen 2 CPUs Expected Too – Production Capacity To Increase By 3 Times Next Year https://wccftech.com/amd-7nm-vega-gpu-zen-2-cpu-mass-production-tsmc/

That sounds like a great move. Having GloFlo and TSMC compete against each other for work can only benefit buyers.

 

[4K8KW10]

It will be interesting to see how AMD will manage the production split between the foundries.
Imagine Ryzen 7 3700 being made both at Globalfoundries and TSMC but depending on the process, the chips get higher frequencies / lower voltages, etc.
Will they split the inferior chips to lower tiers?

 

[4K8KW10]

This story did not leak just today it has been in the press since April 2016. In fact AMD's initial run up in 2016 was mostly because of the announcement of the THATIC Joint Venture. The Hygon x86 Dhyana has been discussed since it appeared this past April or so for Linux validation. So the question is, where have you been?

Intel really can't do anything about it as they agreed as a part of the 2009 Order and Settlement Agreement from the FTC to allow both Via and AMD the right to form Joint Venture's to produce x86 silicon. This is summarized quite nicely in the Federal Register.

Why should you be worried about China? UMC and TSMC are already have ties to China and they both fab a huge amount of the worlds processing IP. You don't think that they already have the litho masks?

The more people who have computers in China and access to information from the west, the better off we'll all be.

China Finds Zen: Begins Production Of x86 Processors Based On AMD's IP https://www.tomshardware.com/news/china-zen-x86-processor-dryhana,37417.html