The article I've linked is from January, I wanted to show that AMD has been hinting at mainstream for awhile. But since then they have clarified further: AMD’s Polaris will be a mainstream GPU, not high-end
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
From ATI to AMD back to ATI? A Journey in Futility @ [H]
- Thread starter FrgMstr
- Start date
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
If AMD were suddenly to gain "dominance" over NVIDIA, would AMD's dominance make you sick?
In a two person race, one person is going to win. If it's not NVIDIA that's winning, it'll be AMD that's winning. Are you going to constantly feel bad that one company does better than another?
Nenu
[H]ardened
- Joined
- Apr 28, 2007
- Messages
- 20,315
He said dominance.
As in one side winning for a long period.
Not innitially because AMD is well over due. They need big wins to stay relevant in this industry and I dont want to see them(especially the GPU section) dissapear. I wouldn't mind if Nv fell behind a few gens and market share was again evenly split between the 2 and prices come back down to a more reasonable lvl due to tough evenly split competition.
Again No. When its almost always Nvidia that dominates, that creates a monopoly and monopolies are usually not good for consumers. If Nvidia was the underdog, Ide probably feel the same for them.
AMD seems to release a bad card when transitioning to a smaller size, but none of that matters it's all happened before.
Point of article is graphics division trying to separate again will have huge consequences.
AMD motherboards don't have a chipset with pci-express 3 support.
Point of article is graphics division trying to separate again will have huge consequences.
AMD motherboards don't have a chipset with pci-express 3 support.
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
This doesn't make sense. Are you not aware of the fact that NVIDIA has a GP106 chip planned?
Read the original article from VentureBeat and not the interpretation from Ars.
Koduri: It’s an architecture codename. It’ll still be Radeon something something on the box. But we didn’t have a consistent architecture name like our competitors do. It was hard, because for people, including yourselves and some of the press and enthusiasts—This family of chips has this architecture and a similar class of features. You can group them easily together.
When we set to design this GPU, we set a completely different goal than for the usual way the PC road maps go. Those are driven by, the benchmark score this year is X. Next year we need to target 20 percent better at this cost and this power. We decided to do something exciting with this GPU. Let’s spike it so we can accomplish something we hadn’t accomplished before. The target we set was to do console-class gaming on a thin and light notebook. What does that take for the GPU in terms of power and configuration? I’m proud to say we’ve accomplished that goal with this GPU.
What did he mean by spike it? He explains it...
Koduri: Yes. We have two versions of these FinFET GPUs. Both are extremely power efficient. This is Polaris 10 and that’s Polaris 11. In terms of what we’ve done at the high level, it’s our most revolutionary jump in performance so far. We’ve redesigned many blocks in our cores. We’ve redesigned the main processor, a new geometry processor, a completely new fourth-generation Graphics Core Next with a very high increase in performance. We have new multimedia cores, a new display engine.
This is very early silicon, by the way. We have much more performance optimization to do in the coming months. But even in this early silicon, we’re seeing numbers versus the best class on the competition running at a heavy workload, like Star Wars—The competing system consumes 140 watts. This is 86 watts. We believe we’re several months ahead of this transition, especially for the notebook and the mainstream market. The competition is talking about chips for cars and stuff, but not the mainstream market.
So half the memory bandwidth (most likely GDDR5 and not GDDR5x) and fewer GCN cores than Hawaii yet set to compete with an R9 390x is a HUGE jump in performance.
Now push the amount of GCN cores to say, 4,096 and add HBM2... that’s potentially Big Vega.
So yeah... Polaris was always supposed to be a mainstream GPU. That much is evident. Vega has two models. Big Vega (mean’t to tackle the likes of a 1080Ti) and Small Vega (mean’t to compete with the likes of a 1070 and 1080).
AMD had to pick a GPU between the two to focus on first and they picked the mainstream part. Why? Because mainstream GPUs outsell highend parts by a landslide. So while AMD will hit the ground running with higher amounts of sales (which they need in order to grab market share) nVIDIA will be stuck selling fewer highend parts.
Then AMD will release Vega around October or November.
As for the article by Kyle...
He’s correct that Raja muscled his way to an RTG position but not for the reasons he has cited. I can’t speak at great lengths about this but it has to do with Mark Papermaster and AMD higher ups focussing on marketing rather than substance (and being terrible at marketing as we all know).
Raja took over in order to steer RTG towards a return to competition. His first task was to grab more marketshare in the upcoming VR market (said to surpass the Smartphone market by 2020 by none other than Goldman Sachs). In order to do this, AMD needs to seed the market with a higher performance part to serve as a testbed for VR titles (particularly for developers). AMD did this with the Radeon Pro card (not targetted towards gamers but VR devs). Next up is to upgrade the consoles towards VR capability (upcoming console upgrades) and then bring down the cost of VR gaming (to offset the cost of a headset (Polaris).
This grants RTG a head start on VR (ahead of nVIDIA).
That’s what’s happening. As for RTG spinning off or being sold to Intel... nope. Instead RTG will sign a Graphics IP agreement with Intel which should be announced as completed soon enough. This will grant Intel what they’re lacking... HSA.
nVIDIA are also heading towards HSA after having picked off AMDs head of HSA around 9 months ago or so.
I doubt that Polaris will be loud or hot but you never know. The die is tiny and this could pose a problem in terms of surface area contact with its cooler.
Nenu
[H]ardened
- Joined
- Apr 28, 2007
- Messages
- 20,315
AMD built the small chip to take on the high end but clock speeds didnt pan out.
So they repositioned it for the mid range.
NVidia will be using a chip that is half of a 1080 to compete.
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
I don't NVIDIA has a "monopoly." AMD has about 30% of the dGPU market right now, so clearly they are a player and a viable one at that. If you are truly concerned about NVIDIA becoming a monopoly then I would encourage you to buy AMD GPUs even when they are the inferior choice. Shouldn't you take one for the team for the good of the GPU market?
Realistically, nobody is going to do that. I'm going to buy the product that'll serve my needs best, it's not my job to "support" any vendor. It's up to THEM to earn MY business. If NVIDIA consistently does a good job of earning our business, why complain? Why wish and hope that they "fall behind"? That doesn't make sense to me.
I remember AMD or I should say ATI having better luck with node transitions than Nvidia over the history. 5870 was a flawless transition to 40nm for instance. It was preceded by a trial run with a 4870 die shrink though, but they were back to back quick releases to get the yields up.
- Joined
- May 18, 1997
- Messages
- 55,652
Koduri: "These are not the GPUs you are looking for."
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
Ten years is a lifetime in technology. ATI/AMD have gotten weaker in the dGPU space as NVIDIA has gotten stronger.
They also have far lesser R&D/manpower compared to Nvidia. Its a surprise what they can do with so little. Unfortunately I think the days of Radeon being able to substantially take the single GPU crown for anything longer than a nmonth or 2 due to launching first is past us unless they come out with a miracle arch like the R300 which is looking a lot less likely given their money problems.
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
The problem is that AMD's total R&D is now lower than NVIDIA's. If you think the architectures put out today show NVIDIA in the lead, wait until the results of what is being put into the pipeline today show up over the next 2-3 years. AMD seems to be betting it all on Zen and the server market and are also trying to duke it out with NVIDIA in GPUs. It's tough.
It's 7 years and.. I think AMD still has strong process implementation engineers. They need them for both the CPU and GPU side of things.
- Joined
- May 18, 1997
- Messages
- 55,652
LOL! You could take that either way. Keep in mind, those were EXACTLY the droids they were looking for. You guys are too easy sometimes.
Your right ofcourse. All of these negative posts or lack of positive ones concerning AMD are having an effect I guess. LOL
Again that is IF its alone, and we know it won't be fairly soon. And that is the problem, when nV grabbed all that marketshare from AMD, it had no competition, but in this case AMD will have competition.
EuphoricRage470
[H]ard|Gawd
- Joined
- Apr 15, 2011
- Messages
- 1,105
The fact that they are pricing them at $199 tells you everything you need to know about how competitive they'll ultimately be. Out of that $199 bill of materials, after the retailers get their cut, AIBs get their cut, all the component makers (think VRMs, DRAM, etc.) get their cut, how much do you think is really left for AMD?
AMD would love to be able to sell their GPUs for more (see: Fury X; see: Radeon 300-series launch which was an attempt to boost selling prices of 200-series cards that were seeing massive price cuts), but they simply can't. That's what Kyle seems to be talking about.
At that point they weren't, that article was done specifically when they were showing off P11 with starwars battlefront. That is the chip that will go into thin and light notebooks.
1060 will be a competition only because it has Nvidia's brand recognition. But when it comes to value proposition I doubt it. FreeSync is a huge selling point that price conscious folks when informed aren't going to pass up.
Lets not get into secondary sales stuff man, because that can go many different ways.
Card for Card, what I'm seeing, the gp106 will have a perf/watt advantage if P10 comes in around 130 watts. And yeah it looks to be doing just that. Die size is pretty much the same. We don't know anything about the VR features of Polaris, but if they didn't add any features like nV did, that is going to be a problem for the VR sales side of things, I'm sure they did expand somethings so it might be a wash too.
Card for Card, what I'm seeing, the gp106 will have a perf/watt advantage if P10 comes in around 130 watts. And yeah it looks to be doing just that. Die size is pretty much the same. We don't know anything about the VR features of Polaris, but if they didn't add any features like nV did, that is going to be a problem for the VR sales side of things, I'm sure they did expand somethings so it might be a wash too.
GPU is a secondary sale by definition.
Except the mainstream market doesn't buy GPUs for $300+ which has been beaten to death already. Vega 10 @ $350 - $500 SKUs, Vega 11 $650+ SKU.
AMD is waiting on VEGA I believe for HBM2.
The primary reason to buy a GPU isn't for Freesync or Gsync ok? Its for gaming. You don't go into a store and think, I want to get a freesync card unless you already have the monitor to begin with. Added to that you look at performance and price and power usage what not.
Obviously I can't read your mind or history and which articles you've read. But for as long as I've been tracking AMD 14nm GPU transition the community at large has pieced the following together:
Polaris 11: Notebook, low power, and low end
Polaris 10: Mainstream
Vega 10: high end
Vega 11: enthusiast
This has been expected for awhile now.
It was only expected after April 22nd or 27th, when AMD specifically stated that.
many journalists up till that point though P10 was performance, yeah you can look it up in google just like you did for that specific article. AMD set no expectations till that April date.
many journalists up till that point though P10 was performance, yeah you can look it up in google just like you did for that specific article. AMD set no expectations till that April date.
If you're not doing any kind of research sure, but I've stated for an informed shopper in $200 GPU bracket, FreeSync is a no brainer. Not to mention a lot of popular value monitors come with FreeSync out of the box.
The early leaks the journalists thought Polaris was a monster die with HBM 2.
So suppose it's April 22/27th when we found out, it's still plenty of time that at this point everyone should be aware.
yeah that was like a year ago dude. it was well know by end of last year what Polaris, Vega were, as AMD clarified them in their road maps.
FreeSync and Gsync are both pretty cool technologies. If I am deciding between two GPUs in the $200 range, similarly priced and similarly performant, I can definitely see FreeSync being a trump card for AMD as the competing solution costs $100-200 more.
Ok I misunderstood your post I thought we were talking about Polaris. I never mentioned Pascal.
yes that is your sales pitch, as should AMD's be, but its not everyone's cup of tea. How many people already have regular monitors without freesync or Gsync
Thats the whole thing Kyle was getting at, AMD might have thought they could have P10 in performance which actually they could have if nV did what they did last year with Maxwell 2, where the gtx 980 was just a little above the 780ti performance, but the 1080 is way above the titan x this time.
At some point you're going to upgrade your monitor. It's just how it is. There is also an HDR monitor invasion looming. For an "informed" shopper shopping in the $200 GPU range, FreeSync GPU is a no brainer. If for anything but to be future proof.
I keep my monitors for close to 3 or 4 generations of cards. And I think that is the typical norm as they tend to be a big purchase. How much are HDR monitors going to cost? You think a $200 card is the same person that will buy a HDR monitor ?