Overclock.net banner

1 - 20 of 108 Posts

·
Registered
Joined
·
800 Posts
Discussion Starter #1
Quote:
Normally GPU manufacturers release the flagship or 'high-end' products first to get all the attention and then release the mid-range chips for the great unwashed a lot later once they have sorted out yields.

But AMD's cunning plan suggests that it is going to do the opposite. It is risky, but it could mean that the outfit could make more money quickly. This is because mainstream GPUs account for the majority of GPU sales.
Source

I did think the same way but I waited that a major(sic) site picked it up.
 

·
Not a Fan
Joined
·
2,648 Posts
AMDs cunning plan.

nVidia does the same thing (launch cutdown x80 first, full die x80TI/Titan later) and gets thousands of posts on how greedy their release plan is.

Should work fine for AMD, though, cunning genius.
 

·
Registered
Joined
·
616 Posts
Quote:
Originally Posted by Robenger View Post

We'll see if it pays off, but people forget how strong the Nvidia hivemind is.
Well to be fair with GPUs it's pretty clear cut, given a specific price range see which card does best in benchmarks and you have your winner.

If AMD's release is cheaper than the 1070, which I assume it will be, it won't really have any competition in that market segment.

How much cheaper, and how close it is in terms of power, will probably be the only thing to look for.
 

·
Irrigator of Souls
Joined
·
1,173 Posts
  • Rep+
Reactions: LAKEINTEL

·
Registered
Joined
·
616 Posts
Quote:
Originally Posted by GnarlyCharlie View Post

AMDs cunning plan.

nVidia does the same thing (launch cutdown x80 first, full die x80TI/Titan later) and gets thousands of posts on how greedy their release plan is.

Should work fine for AMD, though, cunning genius.
Isn't the full die on Pascal supposed to be using HBM2?

So this generation at least the wait is somewhat justified, although I suspect they'd have followed the same strategy regardless.
 

·
Registered
Joined
·
800 Posts
Discussion Starter #8
Quote:
Originally Posted by GnarlyCharlie View Post

AMDs cunning plan.

nVidia does the same thing (launch cutdown x80 first, full die x80TI/Titan later) and gets thousands of posts on how greedy their release plan is.

Should work fine for AMD, though, cunning genius.
We are talking about sub 400$ cards. Nvidia has nothing in site yet.
 

·
Simpleton
Joined
·
1,894 Posts
So this time people are setting themselves for something striking from amd even though they(amd) are pretty clear not to? this just proves whatever they do, they cannot meet expectations because idiots people expect too much whether it is tech media or just the consumers. they are bound to under deliver. because people over-hype by themselves.
 
  • Rep+
Reactions: LAKEINTEL

·
Premium Member
Joined
·
2,813 Posts
Everyone will repeatedly forget this. I'll be like a broken record, from here on, reminding people that die size is relevant if you want to compare architectures.

Polaris 10 is a 232mm2 chip while GP104 is 317mm2. They're in entirely different classes, no doubt.

I don't know how "genius" this is for AMD. It's long overdue. AMD has been trying to beat or match NV since the dawn of time, and it's a fruitless endeavor. If AMD is much less expensive with a smaller die and slightly less expensive, they are considered to have lost. If all things are equal, they are considered to have lost. Even when AMD has produced better performers for a better price, they are considered to have lost.

All things equal, marketing will decide who is better. All things unequal, marketing still decides who's better.

If AMD wants to win, it needs to retake the budget market by storm to make new fans, or they need to outmaneuver NV in the high end market to a point of near fantasy. They only have the financial means to do one of these things; they're making the right choice.
 

·
Premium Member
Joined
·
10,885 Posts
Quote:
Originally Posted by Mookster View Post

Polaris 10 is a 232mm2 chip while GP104 is 317mm2. They're in entirely different classes, no doubt.
That 232 mm size has gotten a lot of traction considering the only source (as far as I know) was some guys LinkedIn profile, where none of the other chip information matched up with any released chips.
 

·
Professional Proletariat
Joined
·
5,671 Posts
Quote:
Originally Posted by Mookster View Post

Everyone will repeatedly forget this. I'll be like a broken record, from here on, reminding people that die size is relevant if you want to compare architectures.

Polaris 10 is a 232mm2 chip while GP104 is 317mm2. They're in entirely different classes, no doubt.

I don't know how "genius" this is for AMD. It's long overdue. AMD has been trying to beat or match NV since the dawn of time, and it's a fruitless endeavor. If AMD is much less expensive with a smaller die and slightly less expensive, they are considered to have lost. If all things are equal, they are considered to have lost. Even when AMD has produced better performers for a better price, they are considered to have lost.

All things equal, marketing will decide who is better. All things unequal, marketing still decides who's better.

If AMD wants to win, it needs to retake the budget market by storm to make new fans, or they need to outmaneuver NV in the high end market to a point of near fantasy. They only have the financial means to do one of these things; they're making the right choice.
They should've ditched the small die strategy when nVidia brought out G80. I imagine things would've played out very differently then. Plus ATi was doing quite well back in the days, so they likely had enough resources to launch a big die if they so desired. And they also held the node advantage too until 40nm. I can't imagine what node advantage + big die would've done to nVidia.

inb4 ATi would've became nVidia
 

·
Banned
Joined
·
5,181 Posts
Editorial, not news.

Also, AMD already stated their plan from the beginning. These guys are just rephrasing it. Woaw. Such news, much excite.
 

·
Sunday League Jibber
Joined
·
4,254 Posts
Quote:
Originally Posted by GnarlyCharlie View Post

AMDs cunning plan.

nVidia does the same thing (launch cutdown x80 first, full die x80TI/Titan later) and gets thousands of posts on how greedy their release plan is.

Should work fine for AMD, though, cunning genius.
Something tells me that poor, beleaguered Nvidia will survive those of us questioning whether the Founder's Edition was an entirely necessary development.

Quote:
Originally Posted by Mookster View Post

Everyone will repeatedly forget this. I'll be like a broken record, from here on, reminding people that die size is relevant if you want to compare architectures.

Polaris 10 is a 232mm2 chip while GP104 is 317mm2. They're in entirely different classes, no doubt.

I don't know how "genius" this is for AMD. It's long overdue. AMD has been trying to beat or match NV since the dawn of time, and it's a fruitless endeavor. If AMD is much less expensive with a smaller die and slightly less expensive, they are considered to have lost. If all things are equal, they are considered to have lost. Even when AMD has produced better performers for a better price, they are considered to have lost.

All things equal, marketing will decide who is better. All things unequal, marketing still decides who's better.

If AMD wants to win, it needs to retake the budget market by storm to make new fans, or they need to outmaneuver NV in the high end market to a point of near fantasy. They only have the financial means to do one of these things; they're making the right choice.

100% this.
 

·
Registered
Joined
·
731 Posts
AMD did the small die strategy from 3870 to 7970, and look how cunningly it worked. They simply need to make better chips at any size, and also compete with software support. Given their resources atm, it's unlikely to happen. So they'll continue to languish with 20% market share while Nvidia and Intel have free reign to price their high end with ludicrous margins.
 

·
Premium Member
Joined
·
2,813 Posts
Quote:
Originally Posted by magnek View Post

They should've ditched the small die strategy when nVidia brought out G80. I imagine things would've played out very differently then. Plus ATi was doing quite well back in the days, so they likely had enough resources to launch a big die if they so desired. And they also held the node advantage too until 40nm. I can't imagine what node advantage + big die would've done to nVidia.

inb4 ATi would've became nVidia
Yeah probably. After the success of the 1950XTX, ATi sort of fizzled out for a while. They had another chance when VLIW4 completely dominated the first Fermi chip despite being a much smaller die, but then still didn't prepare anything to compete with the next generation of Fermi.

Tahiti was a great chip, and many NV users actually switched, but an unbelievable number of them went and sold their Tahiti cards for the inferior 680 once it was released. After that we had Titan, and major financial floundering / management issues inside of AMD.

Nothing new really. More than a decade since I bought my first 1950 XTX, and AMD has had nothing but magnificently bad luck.

I think Raja could have the talent to turn things around. His influence has probably had a greater impact on Polaris than Hawaii or Fiji. We could see a return of the "pre-2007" ATi, if he was as much of a contributor as I think he was.
 

·
Professional Proletariat
Joined
·
5,671 Posts
Even a man as brilliant as Raja is hamstrung by the resources at his disposal and bureaucratic redtape. Thankfully the establishment of RTG means the latter should be a non-issue.
 

·
Registered
Joined
·
800 Posts
Discussion Starter #19
I think the "small die first " is risky but at these difficult times AMD is going "all in". Winning is the only option.
Anything else is destined to fail and I personal hate monopolies.
 

·
Registered
Joined
·
56 Posts
I'll be most interested to see how AMD handle the Vega card release against the Ti. Presumably they won't match it for power, but if Vega is good value and comes out early enough it could take the top spot from the 1080 and sit there for a while tempting impatient potential Ti buyers. If they release at the same time as the Ti it's going to need to be really great value in comparison.

I look forward to finding out what happens.
 
1 - 20 of 108 Posts
Top