Jump to content

AMD Radeon RX 680 Navi 10 GPU With GDDR6 Coming 2019?

AtlasWraith
1 hour ago, AluminiumTech said:

Why would it be too late?

 

If they want to target a performance level customers are familiar with at a mainstream price then that's good. Especially if that performance level previously cost $699 to get to.

 

Making a $299-399 GPU which performs on par with a 1080Ti and is coming in 2019 is good and AMD should be happy. I'd love to get 1080Ti performance for $299 in 2019. Honestly I'd be very happy with that.

 

Most people on the forum seem to forget that the people who buy GTX 1070s, GTX 1080s and GTX 1080Tis are the absolute minority less than 10% and the majority of GPU sales are in the $100-300 range. If it came out at $299 for 1080Ti performance then that'd be great. If it was instead GTX 1080 performance then obviously it would be less exciting but still good.

 

Because AMD isn't throwing down a direct 1080 Ti competitor, a lot of people forget that the actual, mass sales volume is in the 100-300USD range. There's a reason, when looking at broad computer usage stats, AMD actually has slightly larger GPU share than Nvidia. (dGPU + iGPU numbers.) People also forget that the #1 GPU making for nearly 20 years is... Intel.

 

The 1180 is going to come in at ~700USD. For about a 200USD price drop from the current 1080 Ti for similar performance. If AMD can bring ~1080 performance to their RX 580 replacement, they're going to sweep up a big chunk of the market on Price/Performance alone. Nvidia actually know this, which is why they're going to be weirdly aggressive on pricing & other details soon enough.

 

Other little detail, especially if AMD can get RX Vega 56/64 like performance out of their 580 replacement, is that FreeSync's ecosystem is getting pretty big. And it'll keep expanding over time. That's going to matter with each successive generation, especially as 144hz Freesync panels start getting pretty common.

Link to comment
Share on other sites

Link to post
Share on other sites

2 hours ago, mr moose said:

We can use whatever metric for competing we want but it's still the same thing, 2 years too late.

AMD is always late. However the point still stands. A card, as described in this thread, won't be competing in that bracket even if the performance is similar. It will only affect the used market in that comparison.

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, mr moose said:

Because we want AMD competing a year ago, not in a years time.

 

We know that, but while AMD has nothing to compete with nvidia in that  top 10% then they stand to loose sales in the bottom 90% and all the prices will scale accordingly.  Without competition across the board (and especially at the top) the rest of us who buy mid tier cards either end up paying more or have to settle for less than ideal.

 

 

It's a little more complex than that, given Nvidia & AMD's positions. With AMD competing on "value", it actually forces Nvidia's mid-range products to stay around their nominal value levels. Nvidia picks up the "First Mover" advantages, but the cost of the 1050/1060s matters more for Nvidia's bottom line than the entire 1080 range does (in Gaming). 

 

The actual issue, going forward, is that Gaming is a semi-premium task for the GPU, which is why it's starting to reflect those type of pricing effects. The Global DRAM pricing and Crypto Booms meant this generation was a wipeout for everyone, but the current price brackets are likely here to stay for a while.

Link to comment
Share on other sites

Link to post
Share on other sites

4 minutes ago, Trixanity said:

AMD is always late. However the point still stands. A card, as described in this thread, won't be competing in that bracket even if the performance is similar. It will only affect the used market in that comparison.

what point? you asked me why they are too late in my post. I don't care what Nvidia brings out next.

 

 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

16 minutes ago, Taf the Ghost said:

It's a little more complex than that, given Nvidia & AMD's positions. With AMD competing on "value", it actually forces Nvidia's mid-range products to stay around their nominal value levels. Nvidia picks up the "First Mover" advantages, but the cost of the 1050/1060s matters more for Nvidia's bottom line than the entire 1080 range does (in Gaming). 

 

 

AMD has never really competed price wise with Nvidia here in Australia.  They may have good offerings in the US, but I would argue (especially recently) that when accounting for ram and mining etc, that nvidia have charged more because AMD has been less of a threat.  IF your card performs better you can charge more.  Which is why I would like to see AMD actually release a product at the same time as Nvidia that goes toe for toe with Nvidia. None of this:-  wow check out our new card, it performs as well as a 1080 did 2 years ago.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

7 minutes ago, mr moose said:

AMD has never really competed price wise with Nvidia here in Australia.  They may have good offerings in the US, but I would argue (especially recently) that when accounting for ram and mining etc, that nvidia have charged more because AMD has been less of a threat.  IF your card performs better you can charge more.  Which is why I would like to see AMD actually release a product at the same time as Nvidia that goes toe for toe with Nvidia. None of this:-  wow check out our new card, it performs as well as a 1080 did 2 years ago.

For reasons I don't know specifically, AMD products are really expensive in Oz, which would explain the perspective difference. Though most of the recent mining boom profit actually went to the wholesalers, as neither Nvidia nor AMD actually increased production to meet demand.

 

As for the release cycles, I don't think it works to either Nvidia or AMD's interests to release at the exact same time. First Mover will have an advantage, but nearly identical release time frames would hurt the sales for both sides. It's a downside for the consumer, but they aren't being stupid with their release time frames. Though it would be nice if AMD wasn't behind schedule, though everyone has to deal with Node issues.

 

And if the RX 680 performs like the 1080 for 299USD, that's taking Titan X performance from 3 years ago at over 1000USD to under 300USD, which is where the vast, vast majority of the market is. If the RX 680 really does show up at that performance level, Nvidia will be cutting prices regularly for months.

Link to comment
Share on other sites

Link to post
Share on other sites

19 minutes ago, mr moose said:

what point? you asked me why they are too late in my post. I don't care what Nvidia brings out next.

 

 

I did no such thing so either you're thinking of someone else or you're gaslighting me. 

 

I said 1080 ti would not compete with this card. That's a very valid statement. The only place it would compete is on review sites and therefore the used market. It will compete against Nvidia's next mid-range offering unless you mean Nvidia's next mid-range offering will also compete against the 1080 ti which I'd still disagree with. 

To put it simply: competition is not based on performance alone. If that was the case the landscape would not be as it appears now.

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Trixanity said:

I did no such thing so either you're thinking of someone else or you're gaslighting me. 

 

I said 1080 ti would not compete with this card. That's a very valid statement. The only place it would compete is on review sites and therefore the used market. It will compete against Nvidia's next mid-range offering unless you mean Nvidia's next mid-range offering will also compete against the 1080 ti which I'd still disagree with. 

To put it simply: competition is not based on performance alone. If that was the case the landscape would not be as it appears now.

Well it may not be Nvidias midrange. And given their price trends, 2070 at the price of a 1070 with a perf of a 1080ti (personally I'd guess 1080 instead, but even then the point still stands), it would still cost more than 300$ quite a bit. Right now it sells nowhere near the 350$ mark it should be at, and they first released it at 380$ from memory. Navi at 1080ti level at less than 300$ is necessarily bad for Nvidia in all cases. Even 1080 level can be bad for Nvidia because that puts them still on equal footing against amd with a more budget friendly ecosystem on Amd's side. All in all, Nvidias next gen won't be as competitive as people think it'll be if Nvidia do not reduce its margins more than it usually does.

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, Trixanity said:

I did no such thing so either you're thinking of someone else or you're gaslighting me. 

 

 

Our conversation as it went:

 

my first post:

6 hours ago, mr moose said:

How does this look good for AMD?  If it's right, AMD will be finally releasing something to compete with the 1080ti  2 years later.  If they're wrong they have nothing to compete with it for more than 2 years.  

 

 

you talked about not competing on performance alone:

4 hours ago, Trixanity said:

Thing is you don't compete on performance alone so they wouldn't compete against a 1080 Ti. Pascal will have been replaced by then anyway.

I said it doesn't matter which metric you use it's still two years too late:

3 hours ago, mr moose said:

We can use whatever metric for competing we want but it's still the same thing, 2 years too late.

You claim your point still stands and go one to talk about not competing again:

50 minutes ago, Trixanity said:

AMD is always late. However the point still stands. A card, as described in this thread, won't be competing in that bracket even if the performance is similar. It will only affect the used market in that comparison.

hence, I ask what point? because I never made one and only ever answered the question you asked, I did not dispute your opinions on what constitutes competing.

30 minutes ago, mr moose said:

what point? you asked me why they are too late in my post. I don't care what Nvidia brings out next.

 

 

 

If you ask me why I said they are too late, then I tell you they are too late because I think they should have brought this out to compete (in any metric you like) against Nvidias cards of the time. 

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

>tweaktown

>1080ti performance for 400 dollars

 

Yeah its fake

Link to comment
Share on other sites

Link to post
Share on other sites

21 minutes ago, mr moose said:

Our conversation as it went:

 

my first post:

you talked about not competing on performance alone:

I said it doesn't matter which metric you use it's still two years too late:

You claim your point still stands and go one to talk about not competing again:

hence, I ask what point? because I never made one and only ever answered the question you asked, I did not dispute your opinions on what constitutes competing.

 

If you ask me why I said they are too late, then I tell you they are too late because I think they should have brought this out to compete (in any metric you like) against Nvidias cards of the time. 

You said I asked you why they're late. I did not do that. In fact I said they're always late because quite frankly they are. However moving last gens top tier performance  to the mid-range mitigates that and is why I don't think using this alleged card in the same sentence as 1080 ti other than to compare performance is valid.

The problem AMD has is they can't make a halo product to save their life and in that scenario they're late.

 

Overall that's the point I made and the one I said still stands; that they're not competing. By extension that means most of what you said isn't valid. If they brought this card out as a $600+ halo card it would be an embarrassment and you could very easily claim they're late.

 

Basically what you're saying is that the performance of flagships is all that matters. Because that's how you're evaluating it. If this is a $300-400 card then the performance is acceptable if not great even. That's the natural progression. Yesterday's performance at a lower price point. That's how it's worked for the last decade I believe.

 

Tl;dr: they're late but not for the reasons you say.

Link to comment
Share on other sites

Link to post
Share on other sites

Well for price it can really be good. But, still need a new top tier flagship eventually. I know that mid range is the most popular segment and that it's very important so good start from there. Just when the flagship Navi will release is the question.

| Ryzen 7 7800X3D | AM5 B650 Aorus Elite AX | G.Skill Trident Z5 Neo RGB DDR5 32GB 6000MHz C30 | Sapphire PULSE Radeon RX 7900 XTX | Samsung 990 PRO 1TB with heatsink | Arctic Liquid Freezer II 360 | Seasonic Focus GX-850 | Lian Li Lanccool III | Mousepad: Skypad 3.0 XL / Zowie GTF-X | Mouse: Zowie S1-C | Keyboard: Corsair K63 Cherry MX red | Beyerdynamic MMX 300 (2nd Gen) | Acer XV272U | OS: Windows 11 |

Link to comment
Share on other sites

Link to post
Share on other sites

28 minutes ago, Doobeedoo said:

Well for price it can really be good. But, still need a new top tier flagship eventually. I know that mid range is the most popular segment and that it's very important so good start from there. Just when the flagship Navi will release is the question.

there is a good chance there will not be one as for it to happen gcn would have to go over their current soft limit of 64 cus, which considering amd is working on a new architecture for just after navi it would be a bad investment to fix gcn just before trashing it 

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, AluminiumTech said:

Most people on the forum seem to forget that the people who buy GTX 1070s, GTX 1080s and GTX 1080Tis are the absolute minority less than 10% and the majority of GPU sales are in the $100-300 range. If it came out at $299 for 1080Ti performance then that'd be great. If it was instead GTX 1080 performance then obviously it would be less exciting but still good.

 

Gtx 970 die size 398 mm² MSRP $314 Widely considered to be one of Nvidia's most successful gpu ever so I'm sure it has a healthy margin already.

Gtx 980 die size 398 mm² MSRP $549

Gtx 1070 die size 314 mm² MSRP $449

Gtx 1080 die size 314 mm² MSRP $699

 

You can see that 1070 and 1080 are 22% smaller (hence cheaper to make) than 970/980 but price have significantly increased. They may have more vram than before ($30-40 more bom cost maybe?) but consume less power so need fewer beefy components. Each of these 1070 1080 easily earn $100-150 more profit than 970/980 (which were already profitable) and they have consistently sold out for a year now thanks to miners. 

 

970 and 980 are very close 10% apart in performance but 1070 and 1080 are 30% apart so they can harvest a lot more partially defective gpus to make 1070. 1070ti later on is what 1070 should have been. 

 

1070/1080 may account for fewer % sale but their profit margins are insane. We as customers have already felt the effects of a lack of competition from AMD through Nvidia pricing and product segmentation. 

 

For our sake, I hope AMD come up with something good....

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, cj09beira said:

there is a good chance there will not be one as for it to happen gcn would have to go over their current soft limit of 64 cus, which considering amd is working on a new architecture for just after navi it would be a bad investment to fix gcn just before trashing it 

So this one is supposed to be 64 CU then? The yeah I can see that. Then that "Next-Gen" that is called is the next flagship successor. 

| Ryzen 7 7800X3D | AM5 B650 Aorus Elite AX | G.Skill Trident Z5 Neo RGB DDR5 32GB 6000MHz C30 | Sapphire PULSE Radeon RX 7900 XTX | Samsung 990 PRO 1TB with heatsink | Arctic Liquid Freezer II 360 | Seasonic Focus GX-850 | Lian Li Lanccool III | Mousepad: Skypad 3.0 XL / Zowie GTF-X | Mouse: Zowie S1-C | Keyboard: Corsair K63 Cherry MX red | Beyerdynamic MMX 300 (2nd Gen) | Acer XV272U | OS: Windows 11 |

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, Doobeedoo said:

So this one is supposed to be 64 CU then? The yeah I can see that. Then that "Next-Gen" that is called is the next flagship successor. 

i am expecting 64 or close to it, as on 7nm a 64 cu gpu will be around the size of the rx 480 

Link to comment
Share on other sites

Link to post
Share on other sites

9 hours ago, Trixanity said:

You said I asked you why they're late. I did not do that. In fact I said they're always late because quite frankly they are. However moving last gens top tier performance  to the mid-range mitigates that and is why I don't think using this alleged card in the same sentence as 1080 ti other than to compare performance is valid.

The problem AMD has is they can't make a halo product to save their life and in that scenario they're late.

 

Overall that's the point I made and the one I said still stands; that they're not competing. By extension that means most of what you said isn't valid. If they brought this card out as a $600+ halo card it would be an embarrassment and you could very easily claim they're late.

 

Basically what you're saying is that the performance of flagships is all that matters. Because that's how you're evaluating it. If this is a $300-400 card then the performance is acceptable if not great even. That's the natural progression. Yesterday's performance at a lower price point. That's how it's worked for the last decade I believe.

 

Tl;dr: they're late but not for the reasons you say.

My god, last night was such a bad night for me, I did that in the other thread too.   My apologizes in many forms, You clearly did not ask me why they are late.

 

Yes,  I think we all have a slightly different idea of what AMD being late to the game means for the market.  I don't think performance for the flagship is all that matters, but it does play a roll in sales and price structure.   My experience with pricing ( when there is no competition) is that a company will charge as much as the market will pay.  Which means  pricing structure is top down.  AMD want to undercut nvidia, but they aren't going to undercut further than they have too. If nvidia have the top 3 cards then they charge what the market pays and the rest of the line up from both AMD and nvidia will compete with each other scaling down from that.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

7 hours ago, mrthuvi said:

Gtx 970 die size 398 mm² MSRP $314 Widely considered to be one of Nvidia's most successful gpu ever so I'm sure it has a healthy margin already.

Gtx 980 die size 398 mm² MSRP $549

Gtx 1070 die size 314 mm² MSRP $449

Gtx 1080 die size 314 mm² MSRP $699

 

You can see that 1070 and 1080 are 22% smaller (hence cheaper to make) than 970/980 but price have significantly increased. They may have more vram than before ($30-40 more bom cost maybe?) but consume less power so need fewer beefy components. Each of these 1070 1080 easily earn $100-150 more profit than 970/980 (which were already profitable) and they have consistently sold out for a year now thanks to miners. 

 

970 and 980 are very close 10% apart in performance but 1070 and 1080 are 30% apart so they can harvest a lot more partially defective gpus to make 1070. 1070ti later on is what 1070 should have been. 

 

1070/1080 may account for fewer % sale but their profit margins are insane. We as customers have already felt the effects of a lack of competition from AMD through Nvidia pricing and product segmentation. 

 

For our sake, I hope AMD come up with something good....

I don't think the smaller chips are actually cheaper to make per se, you just get more performance per sqmm of wafer.  Over time the price per chip goes down, just like with any product.  Yes they use less silicon (more dies per wafer) but there is the expense of upgrading the tooling and R+D to get the new shrink feasible. 

 

I wouldn't mind seeing some sort of in depth look at the fab process in this regard actually.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

 

4 hours ago, mr moose said:

I don't think the smaller chips are actually cheaper to make per se, you just get more performance per sqmm of wafer.  Over time the price per chip goes down, just like with any product.  Yes they use less silicon (more dies per wafer) but there is the expense of upgrading the tooling and R+D to get the new shrink feasible. 

 

I wouldn't mind seeing some sort of in depth look at the fab process in this regard actually.

You said it yourself smaller = more die per wafer. Also smaller = more tolerant to fault = better yield. So higher upfront investment cost (yeah R&D is getting more and more expensive) but higher profit margin later on. 

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, mrthuvi said:

 

You said it yourself smaller = more die per wafer. Also smaller = more tolerant to fault = better yield. So higher upfront investment cost (yeah R&D is getting more and more expensive) but higher profit margin later on. 

I think the smaller the lithography the less tolerant to fault. And the increased cost to manufacture offsets the volume advantage for at least the first few runs.

Grammar and spelling is not indicative of intelligence/knowledge.  Not having the same opinion does not always mean lack of understanding.  

Link to comment
Share on other sites

Link to post
Share on other sites

41 minutes ago, mr moose said:

I think the smaller the lithography the less tolerant to fault. And the increased cost to manufacture offsets the volume advantage for at least the first few runs.

the smaller the lithography the more steps it takes to create the product. the more steps the higher chance of a foulty product. smaller nodes are best for small chips, like say a CCX. which is why Ryzen can make relativly cheap 7nm chips. idk how AMD is going to make larger, more monolithic dies at 7nm and then sell the product at a relativly low price. this is what makes me kinda sceptical about 7nm on GPUs. Vega is supposed to be a "pipeline cleaner" for future chips, it better work as otherwise i think we can forget Navi at a low pricepoint. 

Link to comment
Share on other sites

Link to post
Share on other sites

On 6/13/2018 at 9:18 PM, ApolloFury said:

I could barely run Dragon Age Inquisition, Rise of The Tomb Raider and Battlefield 1 with the VRAM of the R9 290. Uses around - 3.9GB at 1080p and hitches whenever its VRAM is at that high. Lowering texture quality reduces VRAM usage considerably.

Never had a problem on my 290x......

I refuse to read threads whose author does not know how to remove the caps lock! 

— Grumpy old man

Link to comment
Share on other sites

Link to post
Share on other sites

11 hours ago, miagisan said:

Never had a problem on my 290x......

 

8 hours ago, Humbug said:

I never had a problem on my R9 290 either.

 

At 1080p he is taking as much VRAM as Guru3D's testing did at 4K.

 

http://www.guru3d.com/articles_pages/battlefield_1_pc_graphics_benchmark_review,11.html

 

Also no problems for me either on my 290X's and I even run 2560x1600.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×