AMD Officially Announces Computex Event on June 1st-To Release Bristol Ridge and Give More Information on Polaris

[Ze_Mako]

Don't worry guys I have the scoop on what AMD is up to

 

I present to you, codename "Flipper". 

 

[SteveGrabowski0]

54 minutes ago, ONOTech said:

Your links prove nothing. AMD will be mainstream, we understand that. But the 1070 replaces the 970, a mainstream card, yet boasts Titan X performance. Why exactly wouldn't the 390 replacement, which offers more perf/watt and uses a smaller node, present the same graphical capabilities? Especially with better DX12 support?

 

Yet, you spew all your claims like facts despite very little cold, hard facts of both the 1070 and P10 released. We don't know what the lineup for Polaris will be. It could be the 470 all the way through the 490X or just the 470 to the 480. Speculation is OK, but when you outright insist that an unreleased card WILL perform a certain way, you're already incorrect by default.

If Polaris 10 performs at the expected level of 1070 (eg slightly better than Titan X), it would overnight become AMD's new flagship: the Titan X outperforms the Fury X and Polaris 10 is undoubtably cheaper to manufacture than the 598mm^2 behemoth of a die in the Fury / Fury X, especially factoring in the money they'd save going GDDR5 instead of HBM1. I just don't see someone calling his company's flagship a mainstream card. The 1070 competition is likely either going to be Fury X with a hefty price cut or absolutely nothing if they can't make a profit selling Fury X under $400.

[TheMissxu]

22 minutes ago, SteveGrabowski0 said:

If Polaris 10 performs at the expected level of 1070

But that isn't the expected performance level... It's supposed to perform like a 390x/980/Fury. It's also expected to be priced between $200-$300, about $100 less than the 1070. It could very well perform at 75% of the 1070 (probably closer to 85-90%, but the price dictates as such).

[ivan134]

1 hour ago, Enderman said:

 

yup, both of you are wrong

 

uh, yeah, polaris is not going to be a high end card, the high end cards are Vega

thats what they're having issues with, if you read the article

makes sense because polaris is low-mid range and wont be performing above 390x

Vega is the one that is supposed to go above fury X level to compete with 1070 and 1080

and thats what the article talks about having issues

Wrong about what? This is straight from your link:

Quote

Okay before I start on this news-item, I really need to state that this is based on a vague rumor, nothing has been confirmed or denied otherwise. Here's the story, some reports say Polaris 10 can't hit 850 MHz reliably and that availability will be pushed back to October. I sincerely hope the rumor is not true.

What exactly am I wrong about?

[zMeul]

@DocSwag

might wanna edit the topic title, AMD aren't announcing Polaris, just details about it

the only launch will be, according to the press release, the Bristol Ridge series of APUs

 

even WCCF has a more tamed title

 

14 hours ago, SamStrecker said:

They aren't having problems that was fake.

said who? AMD .. oh, riight

 

by the looks of it, they aren't launching any Polaris based video cards anytime soon - why is that, I wonder .. no, I don't 

[DXMember]

3 hours ago, DeadEyePsycho said:

I saw that depending on the clock that Polaris could easily be on par with Pascal. At 1300MHz it's a bit slower and at 1600MHz it's faster than the 1080. I'll try to find the link.

 

EDIT: Found it

 

what like what? 3 mins spent on calculation in Excel and like 30 more mins on formatting the stuff...

also if you look closely then you will see that B11 and B13 cells as well as B12 and B14 cells should be equal, provided the formula...

[Prysin]

3 hours ago, Enderman said:

yes, polaris will be targeting the sub-$300 GPU market

efficiency is great, and performance should be about 390-390x based on the leaks

not a 1080/1070 competitior for sure

oh oh oh, but you've spent the last year or more arguing that Maxwell was better then any AMD product, because it saves you soooooo much power.

 

And now, when AMD may be releasing something truly efficient. efficiency doesnt matter.

 

GG. Integrity -10

[Enderman]

11 hours ago, Prysin said:

oh oh oh, but you've spent the last year or more arguing that Maxwell was better then any AMD product, because it saves you soooooo much power.

 

And now, when AMD may be releasing something truly efficient. efficiency doesnt matter.

 

GG. Integrity -10

I said "efficiency is great" about polaris

[Prysin]

3 minutes ago, Enderman said:

I said "efficiency is great" about polaris

you say efficiency is great, but you bash the whole move by AMD. Aka focusing on efficiency rather then a "1080 contender".

 

[Tech_Dreamer]

Hope that event is insured against burn damage & fire hazards #AMD #ISpitHotFiYah #Popcorn

[Enderman]

1 minute ago, Prysin said:

you say efficiency is great, but you bash the whole move by AMD. Aka focusing on efficiency rather then a "1080 contender".

 

???

polaris is really good, their efficiency claims are awesome if they are true, and the cost for performance will be very good too based on rumours

 

all i said is that pascal is going to be midrange and low end (according to AMD) so it wont be competing with the 1070 and 1080

 

whats wrong with that?

they will be great budget cards

[ivan134]

25 minutes ago, zMeul said:

said who? AMD .. oh, riight

 

by the looks of it, they aren't launching any Polaris based video cards anytime soon - why is that, I wonder .. no, I don't 

The ability to keep one upping yourself in the dumb things you say is an amazing talent. You're insinuating that AMD is not a reliable source on whether their product is having issues and the baseless rumors are the ones who are correct? This is some single digit IQ shit. How do you even manage to tie your shoes without setting yourself on fire?

[Bouzoo]

3 hours ago, Enderman said:

polaris is the low end stuff

Not completely, AMD has been saying they will be low end and mainstream iirc, but according to them Polaris will even replace Fury cards (not necessarily architecture wise since it won't be built on HBM). My guess is they are miss using the word mainstream to appeal to masses. If not, I'd like a low end card that can run that Hitman demo as we've seen. 

Either way, it won't be only low end, not even close. I'm still positive they will compete with at least the 1070. 

[2FA]

20 minutes ago, DXMember said:

what like what? 3 mins spent on calculation in Excel and like 30 more mins on formatting the stuff...

also if you look closely then you will see that B11 and B13 cells as well as B12 and B14 cells should be equal, provided the formula...

I did notice that, it's called realizing the dude accidentally put 2304 for Vega. I just redid the calculation and it would 4096 SP. The G4*1300*4096=102.27 FPS. The calculations are accurate, just not the sample equation he put down.

[Enderman]

2 minutes ago, Bouzoo said:

Not completely, AMD has been saying they will be low end and mainstream iirc, but according to them Polaris will even replace Fury cards (not necessarily architecture wise since it won't be built on HBM). My guess is they are miss using the word mainstream to appeal to masses. If not, I'd like a low end card that can run that Hitman demo as we've seen. 

Either way, it won't be only low end, not even close. I'm still positive they will compete with at least the 1070. 

all of the news, leaks, and rumours, including stuff from AMD and leaked benchmarks show that it is about 390/390x performance level and below

so if they deliver that with 2.5x efficiency improvements and a lower price then great

but yeah, its not going to be furyx/1070/1080 level, and if you think it is you will surely be disappointed

[Master Disaster]

4 hours ago, ANewFace said:

Rumor has it that Polaris will not compete well with Pascal, some say AMD was anticipating the massive jump in performance Pascal gave

AMD probably got hands on with Pascal before anybody else in the entire world outside of Nvidia, straight up no joke. They weren't anticipating it, they knew it.

[Bouzoo]

36 minutes ago, Enderman said:

all of the news, leaks, and rumours, including stuff from AMD and leaked benchmarks show that it is about 390/390x performance level and below

so if they deliver that with 2.5x efficiency improvements and a lower price then great

but yeah, its not going to be furyx/1070/1080 level, and if you think it is you will surely be disappointed

Leaks showed 1080 to be quite worse than how it ended up so don't take them for granted. I'm open for everything. I hope it will be as good as Pascal, but don't expect anything. Whoever wins gets my money, that simple. 

 

[Tomsen]

1 hour ago, Enderman said:

all of the news, leaks, and rumours, including stuff from AMD and leaked benchmarks show that it is about 390/390x performance level and below

so if they deliver that with 2.5x efficiency improvements and a lower price then great

but yeah, its not going to be furyx/1070/1080 level, and if you think it is you will surely be disappointed

And all that was archived with a clockspeed of ~850Mhz? GCN's sweetspot in frequency/power is most likely around there, but that doesn't mean they can't crank up the frequency for performance. The new node should allow a much higher max frequency than 28nm (see pascal for reference), at a better frequency/power curve.

Just say they get around ~1200-1400Mhz, that is still a 50-75% increase in frequency.

[Spectrez]

Amazing the negativity in here

 

its going to be good to see what price to performance bracket they will occupying and if they will be worth considering before Vega launches.

 

my guess that the AMD board partners will push the cards Mhz with the unlocked option

 

other then that i guess im stuck with a gtx 1070  

[Prysin]

1 hour ago, Master Disaster said:

AMD probably got hands on with Pascal before anybody else in the entire world outside of Nvidia, straight up no joke. They weren't anticipating it, they knew it.

these companies know eachother products before the leaks even reach the internet.

Industrial espionage, reverse engineering (there is two companies in the US alone that specializes in taking images of operating silicon with electron microscopes to dissect the architecture down at a transistor level), computer simulations and projections. They know almost everything, and play their cards accordingly.

 

Nvidia knows that they cannot cut AMD any slack at all. Because GCN is quite possibly the best architecture, concept wise. It can do almost everything at exceptional performance levels, however drivers, APIs and developer support is missing (read: missing CUDA, performance issues in some APIs, AMD drivers) and as a result, you cannot really get the best out of it.

 

The fact that Nvidia has given up on double precision completely is probably a good indiactor of how much work Nvidia would have to do to compete with GCN at this point. Leaving markets that need double precision almost entirely to AMD. Albeit, not many markets require DP these days. Still, choosing to exit a market isn't always a good thing.

 

[captain cactus]

4 hours ago, SteveGrabowski0 said:

If Polaris 10 performs at the expected level of 1070 (eg slightly better than Titan X), it would overnight become AMD's new flagship: the Titan X outperforms the Fury X and Polaris 10 is undoubtably cheaper to manufacture than the 598mm^2 behemoth of a die in the Fury / Fury X, especially factoring in the money they'd save going GDDR5 instead of HBM1. I just don't see someone calling his company's flagship a mainstream card. The 1070 competition is likely either going to be Fury X with a hefty price cut or absolutely nothing if they can't make a profit selling Fury X under $400.

I want to point out to you that all the recent benchmarking videos of the 1080 also had a Fury X in there as well as a Titan X, and in like 4/5 games the Fury X was on par with/beat the Titan X with the one game where it didn't outperform a Titan X it was like 2 fps below it.

 

So no, the Titan X isn't beating the Fury X. It used to, but in the same way the 290x and 780ti were trading the top spot whereas now the 290x is on top of the charts like all the time.

[Tomsen]

4 minutes ago, Prysin said:

these companies know eachother products before the leaks even reach the internet.

Industrial espionage, reverse engineering (there is two companies in the US alone that specializes in taking images of operating silicon with electron microscopes to dissect the architecture down at a transistor level), computer simulations and projections. They know almost everything, and play their cards accordingly.

 

Nvidia knows that they cannot cut AMD any slack at all. Because GCN is quite possibly the best architecture, concept wise. It can do almost everything at exceptional performance levels, however drivers, APIs and developer support is missing (read: missing CUDA, performance issues in some APIs, AMD drivers) and as a result, you cannot really get the best out of it.

 

The fact that Nvidia has given up on double precision completely is probably a good indiactor of how much work Nvidia would have to do to compete with GCN at this point. Leaving markets that need double precision almost entirely to AMD. Albeit, not many markets require DP these days. Still, choosing to exit a market isn't always a good thing.

 

Corporate espionage is a real thing, but not far as you suggest. Nvidia, AMD and Intel is also getting better at containing information/leaks, than they were in the past.

It is a common thing to reverse engineer the competition chips, but this is almost always late in the process (competitions chip is already for public sale).

Sure corporations knows more than the general public about their competition, but there are still lots of uncertainty. By far, they don't know everything. Big corporations also don't like to gamble on uncertainties.

 

The best hardware only perform as good as the software it runs. The product they both are selling, is not hardware alone, but also software.

That is a big part about designing a finished product, sure you can make the most capable hardware in the industri, but if you have trouble developing the necessary software to utilize the hardware, you have an incomplete product.

Which is a big part of why I think AMD didn't rush out polaris. They want to do a proper launch, because reviewers aren't going to give them any slack (drivers, TDP, noise, CF, whatever you would think of). Also gives them the opportunity to have Nvidia finalize the specs for their SKUs, and adjust polaris to it.

 

Well, they haven't. Nvidia's P100 will do well enough with DP, efficiency should be great.

I don't see Nvidia leaving DP market, quite the opposite to be honest.

 

[Daiyus]

I really hope AMD Zen, Polaris and Vega kick ass in their respective market segments just so we can stop with the insane negativity and slander being thrown at AMD. I'm no fanboy of either side; I've had an AMD/AMD gaming rig which is now an Intel/AMD rig which will become whatever the hell offers me the best performance when it's time to upgrade to 4K next year; it might be a GTX 1080 Ti, or it could be an AMD Vega card.If Zen turns out well I'd even consider ditching my Intel CPU.

 

Do people really not realise we NEED both companies to do well? Without competition the market simply stagnates and it's the consumer that gets screwed. Even if you do prefer Intel and nVidia it's suicide to want AMD to fail; have fun with your overpriced monopolies.

[Valentyn]

Just yesterday Polaris 10 Samples were spotted on SiSoft Sandra's database running at 1.27Ghz, yet again debunking that false rumor of 850Mhz issues.

This also shows that the recent rumor of P10 having 2048 Stream Processors is false, as here it shows 2304 Stream Processors.

 

Also worth noting SiSoftware Sandra 2015 is n OpenGL benchmark, and AMD's OpenGL performance is meh to say the least.

 

We also don't know if this is P10 XT or Pro, although 2304 SP's looks more like the Pro aka 480 ( cut down ) than the XT/480X.

 

The P10 card shown there also has 36 Compute Units, while the GTX 1080 only has 20.

http://www.guru3d.com/news_story/amd_polaris_10_samples_work_at_1_27_ghz.html

 

Also if you the SISoft Sandra scores mean anything the P10 card on it scored more than a GTX 980Ti. Mostly because it does significantly better in Double and Quad float GP Compute.

 

Polaris 10 = 1674.64 Mpix/s

http://ranker.sisoftware.net/show_run.php?q=c2ffcdf4d2b3d2efdce4dde4d5e5c3b18cbc9aff9aa797b1c2ffc7&l=en

 

GTX 980Ti = 1256.49 Mpix/s

http://ranker.sisoftware.net/show_run.php?q=c2ffcdf4d2b3d2efdce5d5e6d4e7c1b38ebe98fd98a595b3c0fdc5&l=en

[Prysin]

8 minutes ago, Valentyn said:

Just yesterday Polaris 10 Samples were spotted on SiSoft Sandra's database running at 1.27Ghz, yet again debunking that false rumor of 850Mhz issues.

This also shows that the recent rumor of P10 having 20148 Stream Processors is false, as here it shows 2304 Stream Processors.

 

Also worth noting SiSoftware Sandra 2015 is n OpenGL benchmark, and AMD's OpenGL performance is meh to say the least.

 

We also don't know if this is P10 XT or Pro, although 2304 SP's looks more like the Pro aka 480 ( cut down ) than the XT/480X.

http://www.guru3d.com/news_story/amd_polaris_10_samples_work_at_1_27_ghz.html

so if it matched 390/390X at 850 MHz, then 1.3GHz will be quite the beast.