AMD won't have partners showing Polaris cards at Computex (Nordic Hardware)

[MrTeal]

That's great news if there's quite a small gap between the $100 premium reference cards and the custom ones. Ideally this will end up like most other launches, where ever the "custom" designs just entail mounting the partner's open air cooler on the reference design and calling it a day. Hopefully some of them retail for the $600 MSRP if that's the case.

 

[railven]

In other words, we don't know when non reference 1000 series is coming. At least we got that out of the way.

I'd guess sometime in June.

As for maximising profits, AMD has a clearly defined market segment they're targeting with Polaris and AIB cards might simply not be what they think is a good use of the products they have at hand. It's no more a red flag than Nvidia putting a heavy premium their own limited offering.

NV doing what they've done since 2010 is surprising to people? Okay.

Oh well, guess we just wait and see. I hope you're right, because AMD could use a good launch.

 

[airfathaaaaa]

i really dont understand

many of amd partners are listed
amd has a computex event on their event list
computex arent listing amd
amd doesnt list that macau event
macau event exists as a rumor
partners say they wont show nothing
we dont know nothing as always

http://www.amd.com/en-us/who-we-are/corporate-information/events#2016/06

 

[Stuka87]

()()()

Yes, because obviously they would advertise an nVidia based card using AMD colors.

 

[MrTeal]

Yes, because obviously they would advertise an nVidia based card using AMD colors.

You mean Gigabyte Gaming colours?

 

[railven]

You mean Gigabyte Gaming colours?

No, it's the Intel colors!

 

[Sweepr]

Meanwhile, other sources of AMD's partner manufacturers that right now are frustrated. AMD is said to be moderate with concrete information even to their closest partners and one as yet been told is that it will be delivered a mid-range card that performs like a Radeon R9 390 and R9 390X. With the right price tag, which AMD is said to hit, this can provide a powerful advances in performance per penny - something AMD official pointed out, one of the main goals of Polaris architecture.

- We look at the entire spectrum of players, how many millions they are, what they buy and performance per dollar aspect. How do we make it optimal for them from the performance per dollar aspect, said AMD's graphics director Raja Koduri for PCPer in March .

In the wake of the launch of Nvidia's GTX and 1080 GTX 1070 begins frustration but growth of the red camp's partners. AMD is said not conveyed any information at all about a successor to the Radeon R9 Fury series or indeed disclose to plan how to counter-performance Nvidia's new card.

Even more frustrating is said to be the fact that AMD does not seem ready to show off some new graphics card public during Computex in early June. While the company is a closed press event in Macau on 26th May, the information conveyed there to be under embargo and not get published until later in June. AMD's partners hoped to showcase new graphics card models at the Computex trade show stands so far with the gag.

www.nordichardware.se/nyheter/grafi...osakerhet-hos-amd-partners-infor-polaris.html

Apparently closed press event in May 26 (NDA), embargo on all info till June and they're not ready to show off the new VGAs at Computex. No info about Fury/Fury X's successor yet.

 

[digitaldurandal]

To me reading through this thread I see:

2 or so people in the AMD camp with unrealistic expectations for Polaris that many people have already told them this is the case. With lots of reason not to expect 90% of Pascal performance from Polaris they continue to shout it simply because of the 2.5 perf/watt improvement statement and a demo in January that has no confirmed settings or fps.

At least one person in the Nvidia camp shouting there are problems with Polaris yields with no information but please make sure you document any claims about Pascal. They get called out on it, continue on their ranting. They get asked when non founder Pascal is launching and post an advertisement that has a shadow and a date on it.

I feel like it is very difficult to have any real conversation because at least half of the posts in every thread are just people shouting as loud as they can. They don't feel a need to provide any evidence for claims and even if the CEO of one of the companies made a youtube video addressing them and telling them they were mistaken they would continue on shouting.

I feel like if we all don't enter the thread with an open mind and attempting to actually learn something from new information, or maybe even from a different perspective on old information, that the entire thing is an exercise in time wasting. (Not that GPU releases are that important in the scheme of things.)

 

[flopper]

To me reading through this thread I see:

2 or so people in the AMD camp with unrealistic expectations for Polaris that many people have already told them this is the case. With lots of reason not to expect 90% of Pascal performance from Polaris they continue to shout it simply because of the 2.5 perf/watt improvement statement and a demo in January that has no confirmed settings or fps.

At least one person in the Nvidia camp shouting there are problems with Polaris yields with no information but please make sure you document any claims about Pascal. They get called out on it, continue on their ranting. They get asked when non founder Pascal is launching and post an advertisement that has a shadow and a date on it.

I feel like it is very difficult to have any real conversation because at least half of the posts in every thread are just people shouting as loud as they can. They don't feel a need to provide any evidence for claims and even if the CEO of one of the companies made a youtube video addressing them and telling them they were mistaken they would continue on shouting.

I feel like if we all don't enter the thread with an open mind and attempting to actually learn something from new information, or maybe even from a different perspective on old information, that the entire thing is an exercise in time wasting. (Not that GPU releases are that important in the scheme of things.)

well 17 may you know more.
Likely amd have to do new clocks due to the way nvidia appraoched 16nm.
or its some other reason, cant expect more than speculation before launch, thats part of the fun

 

[Genx87]

To me reading through this thread I see:

2 or so people in the AMD camp with unrealistic expectations for Polaris that many people have already told them this is the case. With lots of reason not to expect 90% of Pascal performance from Polaris they continue to shout it simply because of the 2.5 perf/watt improvement statement and a demo in January that has no confirmed settings or fps.

At least one person in the Nvidia camp shouting there are problems with Polaris yields with no information but please make sure you document any claims about Pascal. They get called out on it, continue on their ranting. They get asked when non founder Pascal is launching and post an advertisement that has a shadow and a date on it.

I feel like it is very difficult to have any real conversation because at least half of the posts in every thread are just people shouting as loud as they can. They don't feel a need to provide any evidence for claims and even if the CEO of one of the companies made a youtube video addressing them and telling them they were mistaken they would continue on shouting.

I feel like if we all don't enter the thread with an open mind and attempting to actually learn something from new information, or maybe even from a different perspective on old information, that the entire thing is an exercise in time wasting. (Not that GPU releases are that important in the scheme of things.)

Welcome to VC&G since about 2003
A place where dreams and nightmares happen on a regular timetable around launches of products.

 

[ShintaiDK]

www.nordichardware.se/nyheter/grafi...osakerhet-hos-amd-partners-infor-polaris.html

Apparently closed press event in May 26 (NDA), embargo on all info till June and they're not ready to show off the new VGAs at Computex. No info about Fury/Fury X's successor yet.

The question starts to be, what comes first. GP106 or Polaris.

 

[tential]

AMD is going to screw this launch up, because that's what AMD always does. So nothing to sees here, just another bumbled launch by AMD.

 

[railven]

AMD is going to screw this launch up, because that's what AMD always does. So nothing to sees here, just another bumbled launch by AMD.

I've accepted this. And have moved on. RIP ATI.

 

[digitaldurandal]

well 17 may you know more.
Likely amd have to do new clocks due to the way nvidia appraoched 16nm.
or its some other reason, cant expect more than speculation before launch, thats part of the fun

Speculation is fine, and it can be fun to try to guess what reality will be in the end, but I would prefer people have at least a working theory with some evidence and not just post the same post a dozen times in four different threads.

My post was also influenced by a discussion I have had recently regarding how difficult it is for human beings to change their mind regarding something, even if they made an initial decision with little or no information. I think it is a flaw in the way our brains work and I want to fight against it.

 

[Arachnotronic]

Yields are fine on GP104, and I think those who are claiming that GTX 1080 will be in tight supply in a reasonable interval around the launch (1-2 weeks) will be proven incorrect.

 

[S.H.O.D.A.N.]

Yields are fine on GP104, and I think those who are claiming that GTX 1080 will be in tight supply in a reasonable interval around the launch (1-2 weeks) will be proven incorrect.

Those suggestions (that I've made as well) were based on DDR5X availability, not Nvidia's yields. Since we know Hynix is pushing mass production now, it's no longer a concern. Well, at least not a major one.

 

[Azix]

Does the translation actually say they won't have cards to show? it sounds more like the partners are in the dark. How do they manage to keep their own Partners in the dark though? How long does it take them to get GPUs together? Do they just have the PCB info ready to drop chips in and ship out or what?

Or are the partners just not talking? Tweaktown got "don't think so" and "as of now, no information".

 

[S.H.O.D.A.N.]

www.nordichardware.se/nyheter/grafikkort/frustration-och-osakerhet-hos-amd-partners-infor-polaris.html

Apparently closed press event in May 26 (NDA), embargo on all info till June and they're not ready to show off the new VGAs at Computex. No info about Fury/Fury X's successor yet.

I have a hard time believing AMD would show up to Computex with zero Polaris presence, unless they sold every single chip they made and can make in the next two or three months to mobile OEMs. And that just ain't happening.

 

[airfathaaaaa]

I have a hard time believing AMD would show up to Computex with zero Polaris presence, unless they sold every single chip they made and can make in the next two or three months to mobile OEMs. And that just ain't happening.

its the usual BS article

if only someone would believe that aib's wouldnt know if amd had or not and they caught "by suprise" and they are "mad" or whatever

that pretty much sums it up really

 

[Glo.]

The "article" came out day after "leak" about pushing Vega for October release.

For me there is absolutely no surprise in this.

 

[R0H1T]

The "article" came out day after "leak" about pushing Vega for October release.

For me there is absolutely no surprise in this.

So you're saying Nvidia leaked this article just a day after AMD leaked the other article, about Vega, & that's just a day or two after the 1070 was leaked, no final specs so leak is just about appropriate

 

[Glo.]

I am not saying this, at all. It is all AMD marketing team fault. Complete silence on the matter of P10 GPUs makes creates space for... things like this. I know that AMD wants their hardware to defend itself, but...

What concept people have about R9 390X? Being slower than GTX 980 Ti. Where reference GPU is on the same level of performance, both graphical, and per watt. It is only however in a situation where both GPUs are not bottlenecked by software. Look at Killer Instinct benchmarks, that have been published lately. R9 290X is 5-10% slower than GTX 980 Ti. Why? It has 96 ROPs and higher core clocks, compared to 64 ROPs. But core count and compute power is very similar for both GPUs: 6.1 TFLOPs, vs 5.6 TFLOPs. Makes sense?
The same goes for R9 380X and GTX 980: GTX is faster due to 64 ROPS, and 4.6 TFLOPs of compute power, and R9 380X - 4.2 TFLOPs, and 32 ROPs. And it is within 10% of GTX 980. What is more, both GPUs have similar power draw.

However absolutely NOBODY will consider Polaris 10 interesting because it will be on the level of R9 390X. Even if in practice - it might be on par, or faster than GTX 980 Ti.

All of this is AMD's fault. They are not able to be loud and change peoples mindshare in fair way. And what they do? They are COMPLETELY silent. And that creates room for speculation, like above.

I do apologize in advance to mods, and everyone who will feel uncomfortable with these words, but their marketing team sucks D***.

 

[airfathaaaaa]

I am not saying this, at all. It is all AMD marketing team fault. Complete silence on the matter of P10 GPUs makes creates space for... things like this. I know that AMD wants their hardware to defend itself, but...

What concept people have about R9 390X? Being slower than GTX 980 Ti. Where reference GPU is on the same level of performance, both graphical, and per watt. It is only however in a situation where both GPUs are not bottlenecked by software. Look at Killer Instinct benchmarks, that have been published lately. R9 290X is 5-10% slower than GTX 980 Ti. Why? It has 96 ROPs and higher core clocks, compared to 64 ROPs. But core count and compute power is very similar for both GPUs: 6.1 TFLOPs, vs 5.6 TFLOPs. Makes sense?
The same goes for R9 380X and GTX 980: GTX is faster due to 64 ROPS, and 4.6 TFLOPs of compute power, and R9 380X - 4.2 TFLOPs, and 32 ROPs. And it is within 10% of GTX 980. What is more, both GPUs have similar power draw.

However absolutely NOBODY will consider Polaris 10 interesting because it will be on the level of R9 390X. Even if in practice - it might be on par, or faster than GTX 980 Ti.

All of this is AMD's fault. They are not able to be loud and change peoples mindshare in fair way. And what they do? They are COMPLETELY silent. And that creates room for speculation, like above.

I do apologize in advance to mods, and everyone who will feel uncomfortable with these words, but their marketing team sucks D***.

man i would keep all of those quotes just to laugh in 2 weeks so many assumption there coming out of nothing lol

 

[raghu78]

Yields are fine on GP104, and I think those who are claiming that GTX 1080 will be in tight supply in a reasonable interval around the launch (1-2 weeks) will be proven incorrect.

Nobody other than TSMC and Nvidia know as to the actual yields on GP104. So stop making any claims. If yields are great why does GTX 1070 require a whopping 20% of cuda cores/SMs disabled. Compare that to GTX 670 which had only 12.5% cc disabled.

 

[Shmee]

This thread is getting out of hand. Too much mindless speculation, flaimbaiting, and snide remarks, as well as way off topic. Closing.