(GURU3D RUMOR) AMD Polaris 10 GPU To Offer Near 980 Ti Performance For 299 USD?

[Raising]

.
Hell, Fury X is now finally faster than 980 Ti and people are shrieking "see told you so!"

I always laugh when I see that quote, at default clocks the fury x might be faster in some cases over the vanilla blower 980ti but once you overclock it's not even close.

Picking any aftermarket 980ti at default clocks already destroys the fury x, overclock it more and leave poor fury in the dust.

 

[dzoni2k2]

I always laugh when I see that quote, at default clocks the fury x might be faster in some cases over the vanilla blower 980ti but once you overclock it's not even close.

Picking any aftermarket 980ti at default clocks already destroys the fury x, overclock it more and leave poor fury in the dust.

Not in DX12 though. You need a superclocked 980Ti to get close to FuryX in recent DX12 games.

 

[Phynaz]

prior to your quoted post there was no one fanboying it up, in fact, most were expressing quite a dose of skepticism that both (near 980ti performance) and ($299) could be true, or might only be true in *certain games*.

so, your post setting up your pejorative comments was nothing more than a strawman. that's not a contribution. and you're shading much closer to this



than you are to this

Are you posting this as a mod or a member?

 

[ElFenix]

Are you posting this as a mod or a member?

If I were posting as a moderator I'd be posting in bold and signing it as such. I'd rather informally take some of the edge off the tech forums than resort to harsh remedies.

 

[maddie]

Good point on higher refresh rates and more pixels... Normally I game at 2560 x 1440 with hopes trying keep as stable and above 144 refresh rate/frames.

It seems DX12 now may help the cause some if it really brings with it better support and latency with CF/SLi.

I see your point and take on it, this is not the platform to give up on performance, its more platform of change in a direction that should allow for real performance increases going forward.

Also makes much more sense why to release the Fury Pro Duo so as to fill the performance gap for a bit longer for those that spend bigger $$$.

This is something with which I have a problem. High frame rates as a solution to tearing.

Unless the frame is updated by the GPU exactly in time with the refresh update time of the monitor, there will exist tearing. In other words, you have to enable V-sync to eliminate tearing, even with high refresh rate monitors.

Having the GPU provide frames faster than the monitor's refresh rate is not a solution, as there will be a mismatch in timings and thus tearing will still occur.

Do you have V-sync enabled at 144 Hz?

 

[mohit9206]

I refuse to believe. $650 level of performance in 2015 for $300 in 2016?I don't think that's plausible. For $300 AMD will happily give you R9 390X performance but not 980Ti. That would be atleast $400 if you want 980ti performance probably $450. AMD obviously wants to profit from every card sold so they can't go around selling 980ti level card for $300.

 

[swilli89]

I refuse to believe. $650 level of performance in 2015 for $300 in 2016?I don't think that's plausible. For $300 AMD will happily give you R9 390X performance but not 980Ti. That would be atleast $400 if you want 980ti performance probably $450. AMD obviously wants to profit from every card sold so they can't go around selling 980ti level card for $300.

Another business analyst wanna be. You have NO idea about either company's corporate strategy for market position and it's insane to assume they won't charge x amount for y performance with ZERO knowledge of their cost structure or market strategy.

Leave the business analyst stuff alone and let these conversations be about being an enthusiast. Or maybe they should hire you as CMO to ensure they get huge margins and marketshare!

 

[tential]

I use DSR where I can. Some games take to it really well, some have odd UI element issues (ie it's rendered at 4K, but not scaled to my 1440p monitor, so it's meant for ants!)

I'd love to return to red team. Then I can get a freesync monitor and not have to pay the empire their tax. But current rumors are making me feel I'll have to wait for Vega if I want to return to red.

All in all, I paid less for my golden sample 980 Ti then I would have paid for the Fury X. So I'm not complaining about what team I ended up on But if AMD can make the right sales pitch, I got no qualms flip-flopping!

I still think gsync with the tax should be purchased. It'd worth it.

As for ui. I only ran into this problem with Witcher 2 but mods fixed it.

I don't care the camp you are in I just think it's insane to not be making gsync or freesync plans no matter the camp you're in.

Like we've shown with Poll on here people rarely even use dsr for some reason. No matter the camp people don't seem to be getting the best value out of their gpus.

 

[moonbogg]

Well, EYE think the rumor is complete BS. I can't imagine a scenario where you get $650 performance for anything near $300.00

Unless!!!

http://www.anandtech.com/show/2365/9

8800GT was $250 and nipped at the heels of the stupid expensive GTX. It came a year after the GTX.

So, maybe EYE actually think the rumor is NOT BS. Yeah, probably not BS. I mean, maybe not BS. My official position is it could potentially not be BS. I'm solid.

 

[DaveSimmons]

More competition at the high end would be great. 980ti performance at even $400, hopefully with lower power use, would be a nice leap forward for AMD. Plus it might force nvidia to price the 1080ti at $500 instead of $650 again.

 

[maddie]

Well, EYE think the rumor is complete BS. I can't imagine a scenario where you get $650 performance for anything near $300.00

Unless!!!

http://www.anandtech.com/show/2365/9

8800GT was $250 and nipped at the heels of the stupid expensive GTX. It came a year after the GTX.

So, maybe EYE actually think the rumor is NOT BS. Yeah, probably not BS. I mean, maybe not BS. My official position is it could potentially not be BS. I'm solid.

Reinforcing your post is this reader comment from your linked article.

"My question is, "Why so cheap, and why now?" for injection into the market, NVIDIA could have raised the price at least $50 (which most retail shops have already done to capitalize on its popularity) and still have a product that sells like crazy"

 

[VulgarDisplay]

Why would they not sell a such a small chip at around $300? If it performs around 980ti that's a nice bonus, and will get a LOT of people to buy one.

Sent from my XT1575 using Tapatalk

 

[Adored]

There are a few reasons why AMD would sell this fairly cheap.

1) They could really use the "shock factor" of a ~980 Ti level card priced around the 970. That would blow the minds of many people, including those with 970's.

2) They similarly could use doing harm to Nvidia's high-end margins, as Nvidia is making all the money in gaming GPU there anyway. Basically 4870 vs 260/280 again.

3) As suggested, there isn't a whole lot of money in midrange GPU at any rate. The difference between $349 and $299 could be well under $100 million in income (after everybody else gets their share) by the end of a year even if the card sells spectacularly. It's a lot less likely to sell spectacularly at $349 and tbh if Zen flops it's not like another $100 million is gonna matter anyway.

4) VR. This one actually makes a lot of sense for AMD and it's why they are talking so much about it. Nvidia could have made an error here if GP106 falls short of the Oculus's VR min reqs.

I still think $349 and $249 for the salvage part though, which should also be VR min req capable. And also that the card will fall a little short of the 980 Ti, but we'll see.

 

[maddie]

People willing to post performance and price predictions for future reference?

Polaris 10 full Fury X to 980Ti class $299
Polaris 10 cut 390 -390X class $225

 

[Erithan13]

People willing to post performance and price predictions for future reference?

Wouldn't be a good forum without it

I see Polaris 10 being at the very worst equivalent to 390 level performance wise (obviously with much less power draw). I think it will average around Fury level and under games favourable to its strengths it will nip at the heels of the 980Ti. I couldn't discount it meeting or exceeding the 980Ti under very specific games/settings/resolutions but I don't buy it's going to meet it on average. All of this is subject to change over time as drivers improve, given what happened with 290/390/380 etc versus Kepler Polaris may have considerable room to grow in the future.

Pricewise? I have reason to be optimistic here. AMD have been very vocal about bringing VR performance down in price and they've been perfectly open about Polaris being midrange. I'll say $400 tops and whatever that ends up translating to this side of the pond. If Polaris turns out to be a monster at mining I'm calling shortages and price hikes right now (6700k all over again unless AMD have anticipated that and have supply ready). Great for AMD, not so much for people just trying to get their hands on one for gaming. The flood of used 390s will be something to behold though.

 

[casiofx]

I refuse to believe. $650 level of performance in 2015 for $300 in 2016?I don't think that's plausible. For $300 AMD will happily give you R9 390X performance but not 980Ti. That would be atleast $400 if you want 980ti performance probably $450. AMD obviously wants to profit from every card sold so they can't go around selling 980ti level card for $300.

R9 390 is only $329 and very near 390X's performance...

 

[moonbogg]

Based on what Nvidia did with the 8800GT, maybe we should be wondering if PASCAL will top 980ti for $300? BAHAHA! Just kidding. That sounds so incredibly stupid lol. Nvidia giving us a good deal? AHAHAHA!

OK, here are my predictions.

Polaris 10 = 980 for $250 or less. Why? Because AMD said they wanted to increase people with VR capable GPUs. The 970 is already around $300 and is said to be a VR card. It would have to be cheaper than a 970 in order to do a better job than the 970 is already doing. It simply must be cheaper than a 970 and perform better than it.

Polaris 10 will cost $250 or less and perform like a 980.

Take it to the bank. Bet the house on it (your house, not mine). This word is golden.

 

[Kenmitch]

Based on what Nvidia did with the 8800GT, maybe we should be wondering if PASCAL will top 980ti for $300? BAHAHA! Just kidding. That sounds so incredibly stupid lol. Nvidia giving us a good deal? AHAHAHA!

Your all over the place.

Are you using fast travel?

 

[moonbogg]

Your all over the place.

Are you using fast travel?

I am a prophet and all prophets are a little weird. You know this. See my last ninja edited post for my prophetic vision which I prophesy will come to pass.

 

[SteveGrabowski]

People willing to post performance and price predictions for future reference?

I'll guess Polaris 10 will offer 390x performance for $270 based on them marketing it as VR spec for the mainstream.

 

[SimianR]

I prefer Nvidia (for now) but I think the most interesting and exiting products will be coming from AMD over the next few years, particularly because people will actually be able to buy them and a great gaming experience can be had by a lot of people and that makes it much more exciting.
No one wants to spend their damn rent money on an overpriced and soon to be underperforming Nvidia ripoff product. That's the picture that has been taking shape lately regarding their ridiculously over priced, short lifespan products. AMD is the brand to be excited about IMO.

Its not your fault that you fell for this trick as many people have since Kepler and somehow continue to be tricked. Nvidia used to sell high end products first and then release mid range products after. They recently started reversing that release schedule, so we get mid range cards first and high end ones later, but they STILL charge the same high end prices for the mid range cards.
They then release the high end cards and pretend like they were made by angels in the magical workshop of Santa Christ and charge you $1,000.00 EFFING DOLLARS for it. Don't do it. Its a trap.

I completely agree with this. I get that everyone wants to be blown away by the high performance parts, and having the absolute fastest card on the market is nice, but with mid range cards being released as "high end" products and then only 3-6 months later the "real" high end card is released to dethrone the last one, it starts to feel like NVIDIA is chasing every last dollar because they already have the market cornered. I've seen people go on and on here about how this is the new reality of GPU's and to simply accept it, but you can speak loud and clear with your wallet. The GPU market really needs a shake up.

 

[raghu78]

With 4 SKUs rumoured to be based on Polaris 10 , I can see a range of SKUs from USD 199/USD 249 - USD 399/USD449.

http://wccftech.com/amd-gcn-4-0-c99-flagship-polaris-rra-certification/

I think Polaris 10 will surprise everyone, even the most optimistic AMD supporter and the die hard Nvidia supporter. I can see a flagship SKU based on GDDR5X while the rest use GDDR5. I think given that yields on FINFET are not so great there will be room for lots of SKUs based on partially disabled Polaris 10 GPUs. I think the statements by Raja Koduri that Polaris is their most revolutionary performance jump in performance so far and that Polaris would cover the entire performance range

http://arstechnica.com/gaming/2016/01/amd-confirms-high-end-polaris-gpu-in-development-for-2016/

"We have two versions of these FinFET GPUs. Both are extremely power efficient," said Koduri. "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."

http://www.pcper.com/news/Graphics-...past-CrossFire-smaller-GPU-dies-HBM2-and-more

"There have been concerns that AMD was only going to go for the mainstream gaming market with Polaris but Raja promised me and our readers that we “would be really really pleased.” We expect to see Polaris-based GPUs across the entire performance stack."

and by Roy Taylor that Polaris will bring faster GPUs than R9 290X at lower prices and in higher volume

http://wccftech.com/amd-polaris-architecture-vr-minimum-spec/

all point to a massive architectural update which is going to be very competitive with Pascal in performance and aggressively priced. I would not be surprised if we see a 2304 sp GDDR5 based Polaris 10 match 980 Ti at USD 299 and a 2560 sp GDDR5X based SKU at USD 399 which is 15-20% faster. The lower price points of 199-249 can be filled by SKUs which have more sp disabled.

The idea here is mainstream pricing does not mean poor performance as is being deduced by a lot of people. It means that AMD is keen to price products very competitively in order to gain back the market share lost. Its not unheard of from AMD and Nvidia to price products very competitively when they are behind in market share or time to market in a specific generation. Eg: HD 4870 at USD 299, GTX 460 at USD 199. In both cases the companies aggressively priced their products to gain market share.

PS: A 232 sq mm GPU at 14nm FINFET is equivalent to 2.2x the die size in terms of a 28nm GPU since TSMC 16FF++ = 2x TSMC 28nm transistor density and 14LPP > 14LPE which is 1.1x the density of TSMC 16FF+ . I think Polaris 10 could end up with 7 billion transistors (I took the median of Pitcairn and Hawaii density scaling as 14LPP is better than a 14LPE which is 1st gen 14nm) while GP104 comes in at 8 billion transistors. I don't see why a fully enabled Polaris 10 SKU cannot end up quite close to a fully enabled GP104. GP104 would be faster but the gap need not be 30-40% and could very well be 10-20%.

 

[swilli89]

OK, here are my predictions.

Polaris 10 will cost $250 or less and perform like a 980.

Take it to the bank. Bet the house on it (your house, not mine). This word is golden.

390x already equals a 980. AMD will best 390x performance.

Polaris 10 $299 at 980 +20% Performance.

 

[SimianR]

I think it's better for everyone to keep their expectations in check and wait and see though. It's always better to under-promise and over deliver then hype something like crazy and be disappointed. If their high end polaris 10 sku does reach 980 Ti performance levels for $300 - then great. I'm guessing that we'll see something closer to 390-390X peformance, but if it's more than that then I will be pleasantly surprised

If they are purposely keeping clock speeds low to keep power use in check, I wonder if they'll bin some chips and have a higher end polaris 10 sku that is clocked much higher with more VRAM. I can see something like that as a decent hold over GPU until VEGA. A polaris 10 GHZ edition so to speak

 

[maddie]

With 4 SKUs rumoured to be based on Polaris 10 , I can see a range of SKUs from USD 199/USD 249 - USD 399/USD449.

http://wccftech.com/amd-gcn-4-0-c99-flagship-polaris-rra-certification/

I think Polaris 10 will surprise everyone, even the most optimistic AMD supporter and the die hard Nvidia supporter. I can see a flagship SKU based on GDDR5X while the rest use GDDR5. I think given that yields on FINFET are not so great there will be room for lots of SKUs based on partially disabled Polaris 10 GPUs. I think the statements by Raja Koduri that Polaris is their most revolutionary performance jump in performance so far and that Polaris would cover the entire performance range

http://arstechnica.com/gaming/2016/01/amd-confirms-high-end-polaris-gpu-in-development-for-2016/

"We have two versions of these FinFET GPUs. Both are extremely power efficient," said Koduri. "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."

http://www.pcper.com/news/Graphics-...past-CrossFire-smaller-GPU-dies-HBM2-and-more

"There have been concerns that AMD was only going to go for the mainstream gaming market with Polaris but Raja promised me and our readers that we “would be really really pleased.” We expect to see Polaris-based GPUs across the entire performance stack."

and by Roy Taylor that Polaris will bring faster GPUs than R9 290X at lower prices and in higher volume

http://wccftech.com/amd-polaris-architecture-vr-minimum-spec/

all point to a massive architectural update which is going to be very competitive with Pascal in performance and aggressively priced. I would not be surprised if we see a 2304 sp GDDR5 based Polaris 10 match 980 Ti at USD 299 and a 2560 sp GDDR5X based SKU at USD 399 which is 15-20% faster. The lower price points of 199-249 can be filled by SKUs which have more sp disabled.

The idea here is mainstream pricing does not mean poor performance as is being deduced by a lot of people. It means that AMD is keen to price products very competitively in order to gain back the market share lost. Its not unheard of from AMD and Nvidia to price products very competitively when they are behind in market share or time to market in a specific generation. Eg: HD 4870 at USD 299, GTX 460 at USD 199. In both cases the companies aggressively priced their products to gain market share.

PS: A 232 sq mm GPU at 14nm FINFET is equivalent to 2.2x the die size in terms of a 28nm GPU since TSMC 16FF++ = 2x TSMC 28nm transistor density and 14LPP > 14LPE which is 1.1x the density of TSMC 16FF+ . I think Polaris 10 could end up with 7 billion transistors (I took the median of Pitcairn and Hawaii density scaling as 14LPP is better than a 14LPE which is 1st gen 14nm) while GP104 comes in at 8 billion transistors. I don't see why a fully enabled Polaris 10 SKU cannot end up quite close to a fully enabled GP104. GP104 would be faster but the gap need not be 30-40% and could very well be 10-20%.

One problem here. Taking Samsung's claim of <0.2 defects/cm^2 at face value.

Using a die of 15mm x 15.5mm with a 300mm wafer and a defect density of 0.19/cm^2 gives 155 good full die out of 238 max. This does not leave a lot for cut down models which should have higher sales volumes.

By the way, if a wafer costs $4000, we have fab costs as $26/die assuming we throw all the defective ones away. With harvesting, I'll bet sub $20/die.

We could have low power high perf/W full dies with DDR5 and high power, high performance, lower perf/W full die models with GDDR5X. This would be exactly as the 4850 and 4870 line.

So cut version and several full versions.