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

[Gerard]

Just as well they tidied up the cables later on, they were all over the place.

 

[gavinh87]

Just as well they tidied up the cables later on, they were all over the place.

All the advances computer tech has made and we are still using ****** front panel connectors

 

[Glaucus]

Apparently Raven Ridge will be released towards the end of the year IIRC,but the only leak so far has been of a ULV one.

Laptops with desktop Ryzen CPUs have already been released,but they are above £900.

thanks but damn them for being so slow, was expecting the apus to be out about nowish.

 

[CAT-THE-FIFTH]

thanks but damn them for being so slow, was expecting the apus to be out about nowish.

Yeah,especially since the APU would probably be a one CCX design,so potentially could be quite a decent CPU itself!!

 

[Fishbait]

All the advances computer tech has made and we are still using ****** front panel connectors

It's easier to enter from the front than it is from behind

*ahem*

I meant as in controllers etc, take DOOM. If you decide to play DOOM, you have to remove the controller or you're screwed M&KB wise & same for some other games.

Beats having to get up & reach around...

 

[gavinh87]

It's easier to enter from the front than it is from behind

*ahem*

I meant as in controllers etc, take DOOM. If you decide to play DOOM, you have to remove the controller or you're screwed M&KB wise & same for some other games.

Beats having to get up & reach around...

Haha, I meant the power on, reset switches etc.
There needs to be a standard layout so we can just plug 1 cable in. This current way has been the same since forever.

 

[subbytna]

@CAT-THE-FIFTH

I think in games both the i5 and i7 CoffeeLakes will behave very similarly, also very similar to the 7700K, as pepole keep reminding us "games don't really use more than 4 cores" 4 cores with HT is probably more true but i think the i5 2 extra real cores over the 7700K will enable it to keep its pace, easily.

The more interesting prospect is the i3 8350, if thats 4 core HT it will also behave in game in much the same way as the i5 and i7, in which case Intel may end up cannibalising their own i5 and i7 line up. think about it

To put it into perspective, and not looking for reasons to bring this chart up again, this is how Intel's 12 thread CPU's fared against their own 8 thread CPU's, Again, i think CoffeeLake will be much the same as SkyLake, at least in games.

(again ignore the red box)

Isn't the i3 a 4 core not HT

 

[humbug]

Isn't the i3 a 4 core not HT

Dunno.

 

[subbytna]

Dunno.

Spotted originally at Anandtech Forums and later further detailed over at Videocardz.com, it is now clear that Intel will push for the higher core count with its upcoming 8th generation Coffe Lake CPUs. Unlike the Core i5 lineup, which will supposedly consist of quad- and six-core CPUs, with and without hyper-threading, the Core i3 lineup will be quad-core SKUs without enabled hyper-threading.

We've already had a chance to see some information regarding the Core i3-8300 and it appears that the Core i3-8350K will be quite similar, featuring 8MB of L3 cache and 4.0GHz clock speed. Unlike the Core i3-8300, the Core i3-8350K will have a somewhat higher 91W TDP and be unlocked.

The Core i3-8100, could be the cheapest Core i3 SKU and work at 3.6GHz, have 6MB of L3 cache and the same 65W TDP, as the Core i3-8300.

 

[Fishbait]

Haha, I meant the power on, reset switches etc.
There needs to be a standard layout so we can just plug 1 cable in. This current way has been the same since forever.

Oops, my warped mind bad.

 

[humbug]

Anand yup

ok thx

 

[pete910]

Now whose using Americanisms

 

[humbug]

 

[RavenXXX2]

AMD to surge more than 40% because its new chip will take share from Intel:

"Our industry checks show improving mindshare/shelf-space for AMD's new Ryzen desktop-PC processors, incl. 30-50% share at prominent e-tailors, well ahead of AMD's 11% current desktop unit share," analyst Vivek Arya wrote in a note to clients Monday. "The key conclusion is that AMD has momentum which should gradually translate to share gains."

https://www.cnbc.com/2017/08/15/amd-to-surge-more-than-40-percent-analyst.html

 

[Simon Belmont]

Well done AMD.

Now do something on the GPU side?

 

[Doobedoo]

Well done AMD.

Now do something on the GPU side?

Well they need more monies to be able to compete on both fronts. Both Intel and Nvidia are obviously able to throw much more resources at their respective markets than Amd are able to do.

If you want them to do better then support them.

 

[CAT-THE-FIFTH]

Have people seen this:

http://www.chinatimes.com/newspapers/20170808000038-260202

(AMD) this year Jedi big counterattack, x86 acceleration processor (APU) and server processor market share steadily rising, the graphics processor (GPU) is also sold out of stock. (2330) embrace, 7 nm APU and GPU orders will be in the first half of next year in the TSMC production capacity of the film production and production, and to speed up the process of micro-speed, has decided to return to the foundry leader TSMC (2330) The

AMD in the 20 years before the 28-nanometer APU and GPU chips, mostly by TSMC OEM production, in 2015 decided to APU and GPU product line into the 14-nanometer fin-type field effect transistor (FinFET) process, the product The line has been handed over to GMF. Therefore, TSMC last year and this year, although actively seeking 16-nanometer orders, but from the main line of the main product line is almost zero.

Restricted Gloriafeld production capacity

AMD launched a new generation of Polaris architecture GPU last year, the market reaction is applauded but not called, the main reason is the occasion of the launch of the occasion, although the massive, but limited by the foundry 14 nanometer production capacity is limited, can not enlarge the shipments, wait until Competitors Huide (NVIDIA) TSMC 16-nanometer process using the Pascal architecture GPU introduced after the AMD GPU foundry production has been out. As the Huida Pascal architecture GPU is a newer architecture, AMD is due to limited production capacity and lost market opportunities.

The second half of fear and see out of stock

AMD has introduced a series of Ryzen processors and EPYC server processors based on the Zen architecture this year, and now appears to have strong sales momentum, the new Vega architecture GPU also received market praise, but AMD is now facing the biggest problem is still Gloriafeld can fully support the required 14 nm and 14+ nanometer production capacity. The industry believes that Glover in recent years, there is no expansion of 14 nanometer production capacity, limited capacity in the case of limited, do not rule out AMD APU and GPU in the second half will still be out of stock situation.

Taiwan after the annual order to eat?

In order to solve the problem of capacity constraints, while speeding up the process of micro-speed, AMD next year 7 nm APU and GPU OEM orders will return to TSMC. According to industry sources, AMD will be introduced next year, several Zen 2 architecture APU processor, and a new generation of Navi architecture GPU chips, etc., will use TSMC 7nm process mass production. If the process is smooth, do not rule out the new generation of 719 nanometer process APU and GPU will also be TSMC won the OEM orders.

AMD has not repeatedly commented on the foundry strategy, however, the executive director Su Zifeng (Lisa Su) has repeatedly said that TSMC and Gloria Plaza are extremely important foundry partners.

Ultra-third quarter of last year, the first six degrees modified with Grosfond's wafer supply agreement, which is a modification of the provisions of the ultra-wafer to Girona Fond other than the foundry to buy wafers, in order to win Other wafer foundries, ultra-pay 100 million US dollars in cash to Ge Luo Fang as a price. In other words, AMD can seek other foundry production capacity support, and TSMC is the first choice for ultra-micro partners.

(Business Times)

So AMD is going to use TSMC 7NM next year for its APUs and GPUs due to restricted 14NM capacity at Global Foundries.

 

[jigger]

Have people seen this:

http://www.chinatimes.com/newspapers/20170808000038-260202



So AMD is going to use TSMC 7NM next year for its APUs and GPUs due to restricted 14NM capacity at Global Foundries.

So AMD could be at 7nm long before Intel?

7nm APU's would be a big deal.

 

[Yaayuh!]

If they're going for 7NM I can imagine it's going to be a while until we see them. Though it must be ideal for APUs.

 

[Dicehunter]

Anyone know or heard anything if Ryzen will by chance get something like Intels Adaptive voltage maybe through a bios update so it's not running at max volts while overclocked ? That's the only thing I miss about my Intel 4770K, Adaptive voltage even when overclocked.