AMD Launching Polaris 10 400 Series GPU June 1st At Computex. $299 (rumor)

I'm seeing some optimistic speculation here, and as somewhat of an AMD fan I hope that it pans out. Because as of right now, the rumors aren't looking good.

390/390X performance means -10% to +20% performance compared to a GTX 970, depending on the game.
Up to 150W means same power draw as a GTX 1070, and 5W more than a reference-specced GTX 970
$299 means $30 cheaper than a GTX 970 (MSRP)

Essentially, we'd be getting a slightly faster, slightly cheaper, 2-year old Nvidia card. These rumors need to be wrong, or AMD will be failing hard.

It's AMD. I have not been excited to see an AMD product since the original 7970. I'm expecting a flop, but hoping for a home-run.
 
It's AMD. I have not been excited to see an AMD product since the original 7970. I'm expecting a flop, but hoping for a home-run.

I used a Radeon 9600xt then swapped for a 9800Pro. Ever since, I've used Nvidia exclusively, not out of fanboyism but out of perceived necessity. I settled for the 9600GT because the HD4850 was out of my budget and AMD had no viable alternative. I again settled for a GTX560 because the HD 7850 was just launched and out of my budget (I first bought the HD 6850 but returned it due to defects). And this past summer I chose the GTX 970 over the R9 390 because I own a Shield and thought GameStream would be nice. GS has been a disaster. So if Polaris is a hit, even a mild upgrade, I want to throw some support to the underdog. I normally upgrade GPUs when I can get double performance within my budget at the time. Now? I'd settle for less, but meet me halfway. I won't buy a sidegrade.
 
It seems an XFX rep has inadvertently confirmed the Polaris launch date is June 1 (check out the reddit screen shot in the comments):

AMD announces Computex 2016 Press Conference and Webcast, promises 'Polaris Updates' | VideoCardz.com

I do not think I have seen so much confusion caused by inconsistent messaging and lack of confidence in marketing a launch, even VideoCardz say they "think" it will launch rather than guaranteed due to the way this has been handled by AMD.
Anyway, the key is when they will give out hardware to reviewers IMO and not just details/verbal launch.
While the 1080 is technically a paper launch, at least the hardware was given out to a broad range of publications for review at the same time.
AMD needs to respond in kind IMO, although they do have some leeway on time as their target is not the same as the 1080 (and probably not the same as the 1070), importantly they still need to get their messaging and narrative on track and delivered with confidence.

Cheers
 
They will probably "paper launch" it at Computex, all the briefing to tech journalists (inviations would already be sent), and ~2 weeks later, GPU hits retail.

Launching entry and mainstream, you need a ton of volume because these markets are all about volume. It would do nobody any good by a premature early access low volume launch for these segments.
 
Must be a paper launch. No credible Polaris leaks recently, no specs, no PCB shots, no die shots. We don't even know the sp count ... 2048? 2304? 2560?
Plus we saw a rumor a week or two ago that no cards would be shown at Computex/E3 (forgot which) and AIBs were upset with AMD.

The whole gang incl. Raja will be at Computex so we will get something from them. Unless he is there to promote APUs?

I would be surprised if they announce the GPUs/prices at all. Probably happening at E3.
 
You guys keep forgetting 1 thing. AMD has talked about being efficient. Right now we have no idea how much IPC was gained with Polaris over Hawaii.

For all we know a 2048SP Polaris could match a 390x. Which has 2816? (im tired and don't want to look it up).

Another thing too, we know a 2048SP Polaris is cut down. So do not rule out AMD.

Sometimes IPC > GHZ.


Its not IPC that will change, each ALU can do how many ops at once? 2 that will not change.

What will change is throughput, the front end changes influence this greatly. And AMD already gave use what the split down of those changes do in a round about way......
 
AMD Takes More GPU Market Share From Nvidia In Q1 2016 - Builds Momentum Ahead Of Polaris Launch

Looks like it did grab more Marketshare....Wow AMD is shocking me.

I guess people do love those 390/390x's

"In desktops the company grew its share to 22.7% a 1.8 percentage points gain. In notebooks the company managed to take a much more impressive chunk of the market from its rival. it now sits at 38.7% a 7.3 percentage points improvement from the last quarter."

I bought one of those 390s...and I'm loving it. (Check sig. It's driving a 1440... very well.)
 
Must be a paper launch. No credible Polaris leaks recently, no specs, no PCB shots, no die shots. We don't even know the sp count ... 2048? 2304? 2560?
Plus we saw a rumor a week or two ago that no cards would be shown at Computex/E3 (forgot which) and AIBs were upset with AMD.

The whole gang incl. Raja will be at Computex so we will get something from them. Unless he is there to promote APUs?

I would be surprised if they announce the GPUs/prices at all. Probably happening at E3.

There weren't many credible leaks of Pascal prior to it's briefing by NV either. This gen's launch is real tight.

They pretty much have to launch at Computex now, expectations are simply too high. Paper launch and retail later but at least deliver the hype to give time for a build up of mindshare.

I mean if all they did was throw around a few benchmark graphs, cherry picked or whatever like NV's charts were, and that's gonna immediately cause a lot of gamers who were wanting 1070/80 to think twice.
 
I'm seeing some optimistic speculation here, and as somewhat of an AMD fan I hope that it pans out. Because as of right now, the rumors aren't looking good.

390/390X performance means -10% to +20% performance compared to a GTX 970, depending on the game.
Up to 150W means same power draw as a GTX 1070, and 5W more than a reference-specced GTX 970
$299 means $30 cheaper than a GTX 970 (MSRP)

Essentially, we'd be getting a slightly faster, slightly cheaper, 2-year old Nvidia card. These rumors need to be wrong, or AMD will be failing hard.

Yes, there is a lot of optimistic speculation. However, if your analysis proves to be true, then there would be no real reason to buy a 970: AMD would have taken that space. ( Check my sig...I am agnostic and have both 970 and 390.) The 390 is better for 1440. After that, it comes down to freesync or gsync, or other similar discriminators. 970 may be approaching eol. A poor Polaris would kick it off the cliff. A good Polaris? Well, let's wait for the [H]. ;)
 
I do not think I have seen so much confusion caused by inconsistent messaging and lack of confidence in marketing a launch, even VideoCardz say they "think" it will launch rather than guaranteed due to the way this has been handled by AMD.
Anyway, the key is when they will give out hardware to reviewers IMO and not just details/verbal launch.
While the 1080 is technically a paper launch, at least the hardware was given out to a broad range of publications for review at the same time.
AMD needs to respond in kind IMO, although they do have some leeway on time as their target is not the same as the 1080 (and probably not the same as the 1070), importantly they still need to get their messaging and narrative on track and delivered with confidence.

Cheers
Funny thing is none of these rumors have come directly from amd. And has said nothing so I really don't thing they owe anyone n explanation. It's just people speculating. For once I think amd has done the right thing, they are overhyping their product and staying quiet.
 
IMO Not much choice for companies but to paper launch their products nowadays to limit leaks and avoid SEC issues. Since so many review sites would ignore the NDAs to get some clickbait revenue.
 
Funny thing is none of these rumors have come directly from amd. And has said nothing so I really don't thing they owe anyone n explanation. It's just people speculating. For once I think amd has done the right thing, they are overhyping their product and staying quiet.
True but they are fuelling these rumours with their management of narrative/communication.
They know if a message is not clear then it can be misconstrued, and when two competitors are about to release highly anticipated products then rumours and speculation go into overdrive.
Classic mismanagement was the tweet regarding the webinar, another classic is this invitation that has been allowed to be interpreted as a gauranteed launch of Polaris and so exacerbated with the watermark added saying launch but it is not clear, 6 months now showing Polaris against a 950, etc.
And more could be said.
This unfortunately is not how one ideally operates their product and business narrative-communication.

As I said in the past, AMD has some exceptional engineers shafted by marketing and business-project management.
Cheers
 
Yep silly season :p

yeah AMD's issues are not their engineering capabilities at all, its Marketing and Management's messages to the public and setting up expectations in a reasonable time frame. When things start getting stretched out the over hype is created by the rumor mill as well.
 
If AMD doesn't paper launch, it is feeding the rumor mills intentionally to get hype. If they paper launch they are carrot feeding the potential buyers. AMD just need to have a good product in a timely fashion that competes well against their only competitor - Nvidia for discrete graphic cards.

I do think AMD will have something (speculation) on improvements in CFX, be it hardware in Polaris and/or software. Besides on June 1 AMD could have a 1080 in one machine being compared to a 490x with DDR5x :D
 
If AMD doesn't paper launch, it is feeding the rumor mills intentionally to get hype. If they paper launch they are carrot feeding the potential buyers. AMD just need to have a good product in a timely fashion that competes well against their only competitor - Nvidia for discrete graphic cards.
May be more a matter of the market segment they're targeting. At <$300 the end user may just be purchasing a PC/laptop with the OEMs doing the actual decision making there. That is likely the bulk of the market and addressed separately. The other major chunk is the actual gamers who will be following benchmarks, and that's a whole other can of worms.

I do think AMD will have something (speculation) on improvements in CFX, be it hardware in Polaris and/or software. Besides on June 1 AMD could have a 1080 in one machine being compared to a 490x with DDR5x
With "Scalability" on their roadmap in 2 years they better have started something in that direction. That's why I can't help but think targeting lower tier chips and adding an ability to pair them up, at least pairs for now, makes sense.
 
True but they are fuelling these rumours with their management of narrative/communication.
They know if a message is not clear then it can be misconstrued, and when two competitors are about to release highly anticipated products then rumours and speculation go into overdrive.
Classic mismanagement was the tweet regarding the webinar, another classic is this invitation that has been allowed to be interpreted as a gauranteed launch of Polaris and so exacerbated with the watermark added saying launch but it is not clear, 6 months now showing Polaris against a 950, etc.
And more could be said.
This unfortunately is not how one ideally operates their product and business narrative-communication.

As I said in the past, AMD has some exceptional engineers shafted by marketing and business-project management.
Cheers

Thats the thing. AMD has not said shit about specs on polaris. They have been quiet, heck they have even lowered expectations by calling it mainstream gaming card. So at what point did amd lie or try to deceive anyone? I just think due to their track record it seems that way but this time they have literally shut the hell up and let rumor mill roll and everyone know polaris 10 to be 32, 36 or 40 CU's but every has failed to realize that it probably is all three in different cards. I don't know why everyone here has assumed that polaris is only one of those but not all three. I highly highly doubt you are going to see one polaris 10 based card. I bet there are atleast 4 of them.
 
Thats the thing. AMD has not said shit about specs on polaris. They have been quiet, heck they have even lowered expectations by calling it mainstream gaming card. So at what point did amd lie or try to deceive anyone? I just think due to their track record it seems that way but this time they have literally shut the hell up and let rumor mill roll and everyone know polaris 10 to be 32, 36 or 40 CU's but every has failed to realize that it probably is all three in different cards. I don't know why everyone here has assumed that polaris is only one of those but not all three. I highly highly doubt you are going to see one polaris 10 based card. I bet there are atleast 4 of them.
that will be very cool if true:) More choices the better
 
Thats the thing. AMD has not said shit about specs on polaris. They have been quiet, heck they have even lowered expectations by calling it mainstream gaming card. So at what point did amd lie or try to deceive anyone? I just think due to their track record it seems that way but this time they have literally shut the hell up and let rumor mill roll and everyone know polaris 10 to be 32, 36 or 40 CU's but every has failed to realize that it probably is all three in different cards. I don't know why everyone here has assumed that polaris is only one of those but not all three. I highly highly doubt you are going to see one polaris 10 based card. I bet there are atleast 4 of them.

AMD has not lied or deceived anyone, i don't think anyone stated that. There is a time to be quiet and there is a time to talk. Since the P10 won't compete with the pascals that have been released, they can still keep quite, but by showing "how much they were ahead" with 14nm chips it has fueled speculation and the rumor mills have been ablaze. Its fine if its a small amount of time, that way things won't get blown out of proportion either way (good or bad) but the longer the time is for launch or at least reviews, the worse it will get. Timing is everything. And right now, if June 1st isn't a launch or launch imminent (2 or 3 weeks even 4 weeks would be ok), there will be negative press around it as it will be compared to what happened last year, even though it might not be as bad when we start looking at performance and metrics vs cost.

p10 2 for desktop and 2 for notebook, I don't see why they would need more than 2 for desktop. And for notebook, they may only need 1 or the same chip for desktop in the notebook.

Its like meeting a girl at a bar, you don't just go drop your pants lol, you talk and get her confidence and then drop your pants but do it at the right time cause you don't want too much talking going on......;)
 
Last edited:
Thats the thing. AMD has not said shit about specs on polaris. They have been quiet, heck they have even lowered expectations by calling it mainstream gaming card. So at what point did amd lie or try to deceive anyone? I just think due to their track record it seems that way but this time they have literally shut the hell up and let rumor mill roll and everyone know polaris 10 to be 32, 36 or 40 CU's but every has failed to realize that it probably is all three in different cards. I don't know why everyone here has assumed that polaris is only one of those but not all three. I highly highly doubt you are going to see one polaris 10 based card. I bet there are atleast 4 of them.
Err.
Where did I say they lied?
I think you misunderstand my post.
Cheers
 
AMD has not lied or deceived anyone, i don't think anyone stated that. There is a time to be quiet and there is a time to talk. Since the P10 won't compete with the pascals that have been released, they can still keep quite, but by showing "how much they were ahead" with 14nm chips it has fueled speculation and the rumor mills have been ablaze. Its fine if its a small amount of time, that way things won't get blown out of proportion either way (good or bad) but the longer the time is for launch or at least reviews, the worse it will get. Timing is everything. And right now, if June 1st isn't a launch or launch imminent (2 or 3 weeks even 4 weeks would be ok), there will be negative press around it as it will be compared to what happened last year, even though it might not be as bad when we start looking at performance and metrics vs cost.

p10 2 for desktop and 2 for notebook, I don't see why they would need more than 2 for desktop. And for notebook, they may only need 1 or the same chip for desktop in the notebook.

Its like meeting a girl at a bar, you don't just go drop your pants lol, you talk and get her confidence and then drop your pants but do it at the right time cause you don't want too much talking going on......;)

LOL the reason I think they will have cards based on 40, 36, 32 CUs and below is because every launch they have made I have seen them replace the x80 and x90 series cards (x being the generation number) now they are going to release 400 series card it its odd they won't replace the 390/390 x everything below it. I am skeptical they will release just a 480x and call it a day and there is no market for 390/390x since no one is going to buy them. I expect them to have 480x amd 490 series cards. Leave fury alone. They have done this almost every launch last few generations and its odd they will only release one polaris 10 desktop card. I may be wrong but the logic doesn't make sense.

When was the last time they only announced 1 desktop card? I don't remember ever. Fury series were two cards, 390 series were two, 290 series were two. It's something they have done for some time now, and they never make notebook chips a major announcement and I don't know why you would count that honestly. That is a separate segment and always has been. Then 380, 280 and so on. They won't replace this with one card. They have always replaced from top down.
 
Last edited:
LOL the reason I think they will have cards based on 40, 36, 32 CUs and below is because every launch they have made I have seen them replace the x80 and x90 series cards (x being the generation number) now they are going to release 400 series card it its odd they won't replace the 390/390 x everything below it. I am skeptical they will release just a 480x and call it a day and there is no market for 390/390x since no one is going to buy them. I expect them to have 480x amd 490 series cards. Leave fury alone. They have done this almost every launch last few generations and its odd they will only release one polaris 10 desktop card. I may be wrong but the logic doesn't make sense.
hmm for each bracket they have only done 2 GPU's usually, well, later on they have come out with other ones for OEM's and such possible just different SKU's due to yields.
 
I dont think we will see a full fledged Polaris 10 card. I think those are going to be saved for Apple again.
 
hmm for each bracket they have only done 2 GPU's usually, well, later on they have come out with other ones for OEM's and such possible just different SKU's due to yields.

True, but you have to admit they have never replaced an old generation with less than two card. I am pretty sure you will see replacement of 390/390x at the very least, because those are gone either way after polaris 10 launches so two cards will probably named 490 and 490x.

Rest is up in the air, may be more coming down the road such as 480/470 and stuff. But I expect 390 series to be replaced.
 
I dont think we will see a full fledged Polaris 10 card. I think those are going to be saved for Apple again.

say what? I don't think apple every got something as fast as 390x desktop. So I believe you were being sarcastic.
 
True, but you have to admit they have never replaced an old generation with less than two card. I am pretty sure you will see replacement of 390/390x at the very least, because those are gone either way after polaris 10 launches so two cards will probably named 490 and 490x.

Rest is up in the air, may be more coming down the road such as 480/470 and stuff. But I expect 390 series to be replaced.

well its a mid range and low end update, but it makes he 390/390x irrelevent, so they will be dropped just not replaced with anything yet, unless they drop Fury prices a 100 bucks
 
say what? I don't think apple every got something as fast as 390x desktop. So I believe you were being sarcastic.

No I am not. That is why we never got a full fledged Tonga card
 
No I am not. That is why we never got a full fledged Tonga card

tonga was never really a replacement of current gen. So I doubt apple Full polaris and rest of the people get cut down versions. If it is replacing the 300 series I highly doubt thats the case. Tonga was like a quiet launch on the side.
 
well its a mid range and low end update, but it makes he 390/390x irrelevent, so they will be dropped just not replaced with anything yet, unless they drop Fury prices a 100 bucks

yea but then you have 390 series irrelevant and not ever replaced? because they have vega as fury replacements. So still no 490 cards, its just odd.
 
Vega is the 390/390x and Fury replacement, there are two Vega chips one small and large.

Actually they have to drop the Fury more than that, or drop it entirely because the 1070 at its price, they need Fury to be less than 380 bucks.
 
Vega is the 390/390x and Fury replacement, there are two Vega chips one small and large.

Actually they have to drop the Fury more than that, or drop it entirely because the 1070 at its price, they need Fury to be less than 380 bucks.

they have never done this before. Vega is a fury replacement how does it replace fury/390/390x? fury already has 3 cards based on it, so they are going to have at least 4 or 5 different vega cards? To drop fury to 380? They rather stop making it because I highly doubt they are making any margins on that. So call me stubborn but I am thinkging 490x and 490 are launched and 390/390x series is cleared out and fury is phased in the next few months to be replaced by vega. That looks like a much smoother transition to me.
 
to get up to something that would compete in the 390/x replacement range, they would need around 200 watts, maybe they can get away with 175 watts, on Polaris, that won't be competing because that won't get up to the 1080 only the 1070 performance wise and then you still have to worry about that extra power. It would be better to get a chip that uses HBM with comparable performance and power usage to a 1070/1080.

Well as it is right now, the Fury line is in trouble they have to drop prices anyways, I'm not sure they have margins to cut it below 380 but if they don't have a choice its better to get something from the left over chips than nothing.

Now they haven't done this before, but things change, you need to shift strategies to compete that is what you do.
 
Why would Nvidia even need the 1070/1080 if the cards was only 390x fast as it only makes 980Ti buyers feel cheated .. I think we will see an Hawaii core under Polaris 10 using 150 watts as we all have seen the recent work AMD has been doing with Hawaii in getting it ready for an upgade..

Hawaii at 1600- 2000Mhz could handle the 1080 because we all know AMD scales better then Nvidia clock for clock and I see no real reason the clocks would not be insane for AMD also as they are all 14nm.
 
Why would Nvidia even need the 1070/1080 if the cards was only 390x fast as it only make 980Ti buyers feel cheated.. I think we will see an Hawaii core under Polaris 10 using 150 watts as we all have seen the recent work AMD has been doing will Hawaii in getting it ready for an upgade..

Hawaii at 1600- 2000Mhz could handle the 1080 because we all know AMD scales better then Nvidia clock for clock and I see no real reason the clocks would be insane for AMD also as they are all 14nm.


Its scales better clock for clock because of more units..... And then not even really because the front end bottlenecks get in the way of the ALU's when the application isn't ALU bound.
 
tonga was never really a replacement of current gen. So I doubt apple Full polaris and rest of the people get cut down versions. If it is replacing the 300 series I highly doubt thats the case. Tonga was like a quiet launch on the side.

Tonga was the 7970 replacement. Suppose to be same GCN cores and 384bit memory.

We got a cut down version with 256bit memory (380x).
 
to get up to something that would compete in the 390/x replacement range, they would need around 200 watts, maybe they can get away with 175 watts, on Polaris, that won't be competing because that won't get up to the 1080 only the 1070 performance wise and then you still have to worry about that extra power. It would be better to get a chip that uses HBM with comparable performance and power usage to a 1070/1080.

Well as it is right now, the Fury line is in trouble they have to drop prices anyways, I'm not sure they have margins to cut it below 380 but if they don't have a choice its better to get something from the left over chips than nothing.

Now they haven't done this before, but things change, you need to shift strategies to compete that is what you do.

Still there is more rumors that support there will be atleast two cards. Polaris has been rumored up to 40CUs, there is more for me to believe that they will have 2 cards atleast than to say they some how need to change strategy. I don't think you are seeing what I am trying to say. you are firm on saying that AMD will have just one desktop part and one laptop top part, which does not make sense what so ever when we are seeing so many parts floating around. May be we can agree to disagree? lol
 
Tonga was the 7970 replacement. Suppose to be same GCN cores and 384bit memory.

We got a cut down version with 256bit memory (380x).
I know, why dont you understand that it was something specific designed for OEM, I am not arguing you on that. What you don't understand it that Polaris is a launch of a new architecture they don't need to worry about special parts for apple. They will all be low wattage so no need to make anything special for apple. Not happening this time.
 
Back
Top