updatebitsandchipsAM4 Socket will be µOPGA and it will have 1331 pins

[DrMrLordX]

Why don't you give us a detailed run down on why you are so optimistic on Zen based of the TECHNICAL details we know so far.

Even though you omitted the question mark, you're answering a question (that wasn't even directed at you, incidentally) with a question. Dirty pool.

 

[AtenRa]

I somehow missed the confirmation. Is there a link?

If voltages drop from node to node, and Bitlife slides listed 95W SR, are there enough pins left to provide the needed currents or is the package fed at higher voltages, which gets further stepped down for the cores?

Well its not official yet, it was mentioned here.

Perhaps its only for OC and we will not see more than 95W TDP SKUs.

 

[cbn]

People don't use hyper 212 on 5820k? What makes the wraith limited to 125w cooler?

Hyper 212 EVO is rated at 180W so it could be used.

And as far as Wraith goes here is a really detailed review on it:

http://www.tomshardware.com/reviews/amd-wraith-cpu-cooler,4450.html

(AMD old 125W cooler on the left, Thermalright Macho Rev.B in the middle, Wraith on the far right)

Old AMD Stock CPU Cooler

AMD Wraith CPU Cooler

Thermalright Macho Rev.B CPU Cooler

Based on these results it appears to me Wraith could easily work with a 140W CPU by either accepting some temp increase or fan speed increase (or perhaps AMD could swap the Delta QFR0912H 92mm x 25mm PWM fan for a 92mm x 38mm PWM fan).

 

[MiddleOfTheRoad]

That's not why they stick with AMD, that is an arguing point used to try and explain away lack of performance.

There is no need to explain away a lack of performance -- you could argue that there really isn't an AMD consumer CPU that isn't a budget offering at the current pricing levels. Their products are good performers for the respective prices. Budget chips are a compromise -- the G3258 is far from perfect, too.

 

[mysticjbyrd]

Hyper 212 EVO is rated at 180W so it could be used.

And as far as Wraith goes here is a really detailed review on it:

http://www.tomshardware.com/reviews/amd-wraith-cpu-cooler,4450.html

(AMD old 125W cooler on the left, Thermalright Macho Rev.B in the middle, Wraith on the far right)

Old AMD Stock CPU Cooler

AMD Wraith CPU Cooler

Thermalright Macho Rev.B CPU Cooler

Based on these results it appears to me Wraith could easily work with a 140W CPU by either accepting some temp increase or fan speed increase (or perhaps AMD could swap the Delta QFR0912H 92mm x 25mm PWM fan for a 92mm x 38mm PWM fan).

I think I linked that review a couple months ago when it came out. As I recall, they said the new wraith was roughly equivalent to a $30 aftermarket cooler. The hyper 212 evo is $30.


You can currently OC an 8xxx with the stock wraith cooler, so running 140w Zen at stock shouldn't be an issue.

I watched a video awhile back where someone compared the Hyper212 and noctua D15. He replaced the fan on the hyper212 with two high end noctua fans, but temperatures didn't decrease much.

 

[Dresdenboy]

Well its not official yet, it was mentioned here.

Perhaps its only for OC and we will not see more than 95W TDP SKUs.

I see. I asked because of the BL slide listing 95W. A 140W rated cooler would be even more quiet on 95W CPUs. OTOH aside from OC, there is also an effect on thermally limited turbo modes.

 

[William Gaatjes]

I think I linked that review a couple months ago when it came out. As I recall, they said the new wraith was roughly equivalent to a $30 aftermarket cooler. The hyper 212 evo is $30.


You can currently OC an 8xxx with the stock wraith cooler, so running 140w Zen at stock shouldn't be an issue.

I watched a video awhile back where someone compared the Hyper212 and noctua D15. He replaced the fan on the hyper212 with two high end noctua fans, but temperatures didn't decrease much.

They better watch out with that extreme heat. When the heat sink is 105C, the components may be a lot hotter. Inductors have a ferrite core made of a soft magnetic material. That core has also a Curie (Neel) temperature where the effect is that the ability to store energy in the core is reduced. In effect, the inductance of the inductor is reduced. This could cause instabilities and can damage the processor.
A motherboard that is good designed will have inductors designed for high temperatures.
A high switching frequency ( >> 1 MHz ) design is needed because the permeability and the maximum work temperature are related.

edit :
I should note that usually the curie temperature is high enough that it is never reached. Usually above 150C (but 100C ferrites exist too).
Most inductors are rated at 125C.

 

[majord]

I personally think official 140w CPU support for the platform as a whole is a shit idea. Ensuring the socket can support it (and well beyond) absolutely, but I would hope they wouldn't release any SKU's higher than the mainstream 95w standard.

The only way I think it could work successfully is if they partially segregate higher power SKU's onto something like an "AM4+" platform.. This would be a nice compromise between a 'one size fits all' platform (which doesn't work these days IMO as they become cost prohibitive for low end market - see AM3+ for example) , and the complete socket segregation intel go with. This way board partners can still keep a common platform, but with some key differences in regards to things like VRM spec, HS clearance, even certain PCB layout spec's.

 

[Shehriazad]

I personally think official 140w CPU support for the platform as a whole is a shit idea. Ensuring the socket can support it (and well beyond) absolutely, but I would hope they wouldn't release any SKU's higher than the mainstream 95w standard.

The only way I think it could work successfully is if they partially segregate higher power SKU's onto something like an "AM4+" platform.. This would be a nice compromise between a 'one size fits all' platform (which doesn't work these days IMO as they become cost prohibitive for low end market - see AM3+ for example) , and the complete socket segregation intel go with. This way board partners can still keep a common platform, but with some key differences in regards to things like VRM spec, HS clearance, even certain PCB layout spec's.

I'm almost certain that any half-decent mainboard brand is not gonna be retarded and make a 3 + 1 design for 140W chips.

If they really support 140W on the platform and PLAN TO USE IT...then Board makers will likely half multiple segments.

I could imagine 3 segments

AM4 A: 95W, mainly 3+1 phase designs....cheap boards for people not looking to OC and save some money

AM4 B: 140W, better designs but pretty much only designed to run the 140W stuff "as is"

AM4 C: 140W, best designs, want to overclock the living hell out of it? You will want one of those.


140W needs some good hardware, so I don't really think AMD wants to "force" boardmakers to even put expensive designs on the low end...raising the platform cost unnecessarily like this will only scare away potential customers in my eyes.


Also I can definitely see 140W CPUs from AMD. Just imagine FX 9000...but without all the suckage. 140W for 8cores/16 threads at 4 Ghz? That's gonna need a lot of juice...

 

[MiddleOfTheRoad]

AM4 the last gasp of AMD? Say it ain't so! Not sure how I feel about Bristol Ridge what with some of those 845 reviews showing up badly in games, but hey, we'll see what it can do soon enough.

Actually, the 845 reminds me of the Covington Intel Celeron where they totally eliminated the L2 Cache. The performance on those things was dismal. When you hack down the cache to be cheap for OEM's -- then performance is going to suffer.

 

[majord]

Well, unfortunately history has proved otherwise.

Asrock, Gigabyte, Asus have all at some point been guilty of marginal 140w, even 125w support on cheaper motherboards, usually in the hope that the majority of people won't load them to the brink in poorly ventilated cases, and/or not properly taking into account the higher current requirement of low Vcore, high TDP parts... (typically higher core count SKU's)

And it's not that they intentionally designed the board to be sub par from the outset, it's just that exuding new, higher TDP SKU's from their support list is going to lose them sales, and with no clear boundaries, it's a bit of a case of "monkey see monkey do" when a new SKU support shows up on a competitors equivalently priced motherboard.

 

[Shivansps]

Well, unfortunately history has proved otherwise.

Asrock, Gigabyte, Asus have all at some point been guilty of marginal 140w, even 125w support on cheaper motherboards, usually in the hope that the majority of people won't load them to the brink in poorly ventilated cases, and/or not properly taking into account the higher current requirement of low Vcore, high TDP parts... (typically higher core count SKU's)

And it's not that they intentionally designed the board to be sub par from the outset, it's just that exuding new, higher TDP SKU's from their support list is going to lose them sales, and with no clear boundaries, it's a bit of a case of "monkey see monkey do" when a new SKU support shows up on a competitors equivalently priced motherboard.

AMD did not even release a cheap AM3+ reference board, because they never launched a new cheap chipset, FM2, and then AM1 was the new low end, AM3 was not intended to be cheap. So OEMs did what the wanted, re-release GF7025 and 760G... ASRock 980DE3/U3S3 and Asus M5A78L-USB3 are actually not bad cheap 140W AM3+ boards.

But i whould use a 970 board instead, over the years the 970 chipset became cheap enoght.

 

[jpiniero]

I personally think official 140w CPU support for the platform as a whole is a shit idea. Ensuring the socket can support it (and well beyond) absolutely, but I would hope they wouldn't release any SKU's higher than the mainstream 95w standard.

65W is really the OEM standard now but for Summit Ridge 140W is not out of the question.

 

[dooon]

Thermalright Facebook

For all AMD fans, we have good news.
As we advance AMD has assured the new AM4 processors and motherboards are
put on the usual base-fixing, which is standard for AM2.
To follow all the Thermalright coolers are used on the Zen processors without additional accessories!

https://www.facebook.com/thermalright/posts/10153794924514102

 

[DrMrLordX]

Wat? How is that possible?

 

[.vodka]

Maybe AM4 is similar to their 940 and up sockets in layout and infrastructure and decided not to break years of compatibility at the last minute. The old rumors of different heatsink mounting specs for AM4 didn't state big differences with their old sockets anyway.

Good news if true, for those who already own a nice heatsink.

 

[LTC8K6]

Wat? How is that possible?

Tiny, closely spaced, delicate little pins?

 

[sm625]

AMD would be foolish to waste resources targetting > 95W for a 2016 design.

 

[The Stilt]

Wat? How is that possible?

Why wouldn't it be? :sneaky:
The AM4 package is physically exactly the same size as all AMD CPUs since Socket 754 Athlon 64s.

 

[DrMrLordX]

Well it just confuses me that they announce this now, after clearly stating that AM4 had a different mount pattern than AM3+/FM2+ and older.

 

[zentan]

Thermalright Facebook

https://www.facebook.com/thermalright/posts/10153794924514102

That particular Facebook post by Thermalright is not available anymore. Removed likely

 

[Redentor]

New facebook post: https://www.facebook.com/thermalright/posts/10153796912434102?__fns&hash=Ac1SLaIFUQYWMrwc

So B&C guys were right: http://www.bitsandchips.it/52-engli...holes-arranged-differently-than-older-sockets

 

[PG]

New facebook post: https://www.facebook.com/thermalright/posts/10153796912434102?__fns&hash=Ac1SLaIFUQYWMrwc

So B&C guys were right: http://www.bitsandchips.it/52-engli...holes-arranged-differently-than-older-sockets

The facebook link has an error:

 

[superstition]

Who the hell buys a $100 cooler for AMD these days? For that money you could step up to a far better Intel chip.

Good question!

But the question still remains: Who in their right mind would have bought a $100 cooler for an AMD system in the last decade (i.e. since C2D's 2006 debut) instead of funneling said funds into creating a much faster Intel system to begin with?

Interesting observation.

AMD Enthusiasts? They exist.

I agree that it would have to be irrational bias that would prompt someone to buy an AMD part.

They were already disqualified by the first premise in my question.

Definitely.

I know some people who flat out will never buy Intel -- because in their eyes they are the evil empire and they will lose their rebel status. Seriously.

Too bad for AMD stock that there aren't enough rebels.

 

[superstition]

But you don't even know how Zen will perform?

Can't be worse than Piledriver

It could be!

Bulldozer was utterly fantastic to read about.

ShintaiDK, I remember that Bulldozer fiasco well.

At this point that comment is the equivalent of pointing and laughing when a special needs kid falls down.

It was a joke alright, but hardly fantastic, almost like AMD was trolling us at first.

do we really need to be reminded in thread after thread that AMD hasn't done well over the last 10 years?

I think so. As long as AMD is around we're all in danger of having to deal with their terrible products.

Yes, and it was ridiculous to think they would release Bulldozer with worse IPC in a lot of workloads that Phenom, but they did it. And that was a "clean sheet" design. What I am saying is that it is really not that easy, and bad designs make it to market.

I agree we need competition, but all the hysteria and hand-wringing about the lack of competition is crazy as well.

Why do you have this obsessive compulsive need to post anti-amd rhetoric in every single thread?

It's called "keepin it real"...

Well, if one thinks a poster is biased or posts useless information, they are certainly free to use the ignore button instead of making personal attacks.

No one doubts Bulldozer was an epic fail.

I see.

Why don't you give us a detailed run down on why you are so optimistic on Zen based of the TECHNICAL details we know so far.

That's not why they stick with AMD, that is an arguing point used to try and explain away lack of performance.

Good to know.