[H]ardOCP Editor Claims Internal Problems for AMD - Polaris Underwhelming and RTG Seeking to Leave AMD For Intel

[Castdeath97]

Got this from the discussion thread of this article:

 

Kyle:

"Ya think?   There will be a few folks that call bullshit on this, but let's see how close I am to the mark when the smoke clears. That said, we do not publish articles such as this based on rumors, and never have."

 

I think he needs to revise his definition of rumour. I don't think he could have gotten a official or direct source, and if he did he should have linked it. Otherwise, It's a rumour! Check the dictionary, unless it's certain piece of news that can't be doubted (i.e you have a direct link or source when can check) it's a rumour.

[Misanthrope]

4 minutes ago, don_svetlio said:

Fiji? Flop? I beg to differ - check DX12 benches, it's above the 980 Ti/Titan X

Seriously? Check the 3 games out now for DX12 out of wich only Ashes of Singularity really outshines?

 

No, how about you check their sales instead since at the end, that's all that's really going to matter: AMD could be sitting on a fucking mind blowing piece of tech and it would be all for nothing if they go bankrupt before they manage to even get it out to the public, let alone optimize software enough that it fucking matters.

 

We're looking at least a year before DX12 is a decisive factor for buyers at all.

[-BirdiE-]

4 hours ago, don_svetlio said:

Seems like a PR stunt honestly. This should be under NDA aka this guy should get sued hard for leaking such info and since no action is being taken I'd imagine it's false.

NDA isn't something that you classify information as, and then no one can talk about it... It all depends how (if) he got the information. If he never signed a Non-Disclosure Agreement with and somehow found this out he's free to do with the information as he pleases.

 

However, if it turns out to all be false, he could be slapped with a slander suit.

[Derangel]

2 minutes ago, Castdeath97 said:

Okay this is gold, got this from the discussion thread of this article:

 

Kyle:

"Ya think?   There will be a few folks that call bullshit on this, but let's see how close I am to the mark when the smoke clears. That said, we do not publish articles such as this based on rumors, and never have."

 

I think he needs to revise his definition of rumour. I don't think he could have gotten a official or direct source, and if he did he should have linked it. Otherwise, It's a rumour! Check the dictionary, unless it's certain and doubtful (i.e you have a direct link or source when can check) it's a rumour.

Read what he said carefully. BASED on rumors. He's not saying the article itself doesn't fall into rumor territory right now, he is claiming that he trusts that the information from his sources is valid and they can be trusted. Remember, actual real journalists often don't reveal their sources either. This silly requirement to reveal confidential sources does not in anyway conform to what real journalists do. A good reporter tries to avoid revealing their sources so that those sources can continue to be used and so that those sources don't lose their jobs, get sued, or go to jail for revealing confidential information.

[Castdeath97]

8 minutes ago, Derangel said:

Read what he said carefully. BASED on rumors. He's not saying the article itself doesn't fall into rumor territory right now, he is claiming that he trusts that the information from his sources is valid and they can be trusted. Remember, actual real journalists often don't reveal their sources either. This silly requirement to reveal confidential sources does not in anyway conform to what real journalists do. A good reporter tries to avoid revealing their sources so that those sources can continue to be used and so that those sources don't lose their jobs, get sued, or go to jail for revealing confidential information.

If he can't reveal his sources then it's probably from a non-official indirect source. And being a non official indirect source means that it can't be taken for granted (and let's not forget his history with AMD).

 

So, it's probably from someone from AMD who may or may not have the full picture, may or may not have bias (i.e. from being fired) or may or may not have lied. Basically unless it's from a direct source who had access to the *final* card with the *final* software we can't take anything for granted now.

[-BirdiE-]

I really hope this turns out to be false. I'd hate to see Nvidia start to lose any competition.

 

Sorry if this has already been adressed, but some of it does seem to make sense:

I thought it was odd that the highest end card they were releasing with Polaris is the 480X, and that we have to wait for another architechture (albeit only until October) to see high end performance cards. HOPEFULLY that's just because they want to wait for HBM2 for those cards, not because Polaris wasn't what they wanted it to be.

I was also somewhat shocked when the benchmarks leaked that showed 2 480X cards in CF still being outperformed by a 1080, and are anticipated to be $300 USD. That doesn't sound competitive with the 1070 or 1080 at all. Again, hopefully all false or poor pre-release optimization.

 

 

[maxrosciano]

All I see is a butthurt journalist spewing BS.

[Derangel]

1 minute ago, Castdeath97 said:

If he can't reveal his sources then it's probably from a non-official indirect source. And being a non official indirect source means that it can't be taken for granted (and let's not forget his history with AMD).

Well, of course, it's not an official source. That doesn't mean the source itself is a rumor though. It all depends on how much the person reporting trusts the source and how much that source actually knows. No matter the industry articles like this almost never reveal sources. Specifically for reasons I already mentioned. Not revealing sources does not invalidate an article. Even looking at Kyle potentially being mad at AMD for one reason or another doesn't. As of yet, [H]ardOCP has never done an article like this that has turned out to be full of lies. The history of the site and the connections Kyle has gained in the industry of over the last 18 years the site has been around lends some level of credibility to the article. Obviously it's worth looking at it critically and thinking it over, but you can't lambast them for actually following journalistic ethics in how it was written.

[don_svetlio]

8 minutes ago, Derangel said:

Well, of course, it's not an official source. That doesn't mean the source itself is a rumor though. It all depends on how much the person reporting trusts the source and how much that source actually knows. No matter the industry articles like this almost never reveal sources. Specifically for reasons I already mentioned. Not revealing sources does not invalidate an article. Even looking at Kyle potentially being mad at AMD for one reason or another doesn't. As of yet, [H]ardOCP has never done an article like this that has turned out to be full of lies. The history of the site and the connections Kyle has gained in the industry of over the last 18 years the site has been around lends some level of credibility to the article. Obviously it's worth looking at it critically and thinking it over, but you can't lambast them for actually following journalistic ethics in how it was written.

Thing is - the things he talks about go against what other outlets report. I'm sorry but PCper, Anandtech, TechPowerUp, Arstechnica are far more credible in this point of time

[Castdeath97]

1 minute ago, Derangel said:

Well, of course, it's not an official source. That doesn't mean the source itself is a rumor though. It all depends on how much the person reporting trusts the source and how much that source actually knows. No matter the industry articles like this almost never reveal sources. Specifically for reasons I already mentioned. Not revealing sources does not invalidate an article. Even looking at Kyle potentially being mad at AMD for one reason or another doesn't. As of yet, [H]ardOCP has never done an article like this that has turned out to be full of lies. The history of the site and the connections Kyle has gained in the industry of over the last 18 years the site has been around lends some level of credibility to the article. Obviously it's worth looking at it critically and thinking it over, but you can't lambast them for actually following journalistic ethics in how it was written.

But then again, does the source have direct access to final hardware and final software? Polaris is launching in July and there is no way his sources know how for certain Polaris will turn out from the get go.

 

I didn't say he should link the source if it isn't official, because I know it isn't, I'm not angry over them for doing this I'm just pointing out that there is no way the source of this news is certain as it's not official and direct (and therefore doesn't qualify as rumors).

[Thiback]

30 minutes ago, don_svetlio said:

Fiji? Flop? I beg to differ - check DX12 benches, it's above the 980 Ti/Titan X

Yeah but beside that Fiji was underwhelming, notably in DX11, in regards to the leaks prior to its release stating it would be over the 980ti. The opposite happened. I was following that closely, and was also disappointed by the fact that they only put 4gb of VRAM in there for 4k GPU. I know that they couldn't put more HBM due to low yields, but you gotta admit for a 4k card, the 980ti comes rather on point with its 6gb.

Anyway, I don't think the Fiji cards were a success, even though they had the advantage of HBM.

[Trixanity]

It seems like the author sees connections where there are none. This is some ratchet Illuminati type shit going on. It's like going 2 + 4 = 6. 6 x 3 = 18. 18 / 6 = 3. Half Life 3 = confirmed.

 

How on Earth would a subsidiary go about selling itself underneath the noses of the parent company without permission? How would Intel swoop in and buy RTG? How would they get all the necessary patents to go along? I mean this is quite the tale. There is no doubt that some of the stuff listed in the article is true but it's like the author sees this grand scheme on a chess board while we are in fact playing fucking checkers.

[Castdeath97]

1 minute ago, Thiback said:

Yeah but beside that Fiji was underwhelming, notably in DX11, in regards to the leaks prior to its release stating it would be over the 980ti. The opposite happened. I was following that closely, and was also disappointed by the fact that they only put 4gb of VRAM in there for 4k GPU. I know that they couldn't put more HBM due to low yields, but you gotta admit for a 4k card, the 980ti comes rather on point with its 6gb.

Anyway, I don't think the Fiji cards were a success, even though they had the advantage of HBM.

The R9 Fury non-x was a good card.

[Trixanity]

1 minute ago, Thiback said:

Yeah but beside that Fiji was underwhelming, notably in DX11, in regards to the leaks prior to its release stating it would be over the 980ti. The opposite happened. I was following that closely, and was also disappointed by the fact that they only put 4gb of VRAM in there for 4k GPU. I know that they couldn't put more HBM due to low yields, but you gotta admit for a 4k card, the 980ti comes rather on point with its 6gb.

Anyway, I don't think the Fiji cards were a success, even though they had the advantage of HBM.

Not yields. It was a HBM gen 1 limitation. 1 GB per stack. Fiji could only fit 4 stacks hence 4 GB HBM (not sure if there is a stack limitation).

[don_svetlio]

1 minute ago, Trixanity said:

Not yields. It was a HBM gen 1 limitation. 1 GB per stack. Fiji could only fit 4 stacks hence 4 GB HBM (not sure if there is a stack limitation).

There is. HBM1 COULD go up to 8GB but not at launch. HBM2 is more practical

[Thiback]

2 minutes ago, Castdeath97 said:

The R9 Fury non-x was a good card.

Granted, since it fills a price gap that no other card does. And the Nano has the advantage of being really small. But I find that's all there is about it. Fury X wasn't a real great card either.

[don_svetlio]

1 minute ago, Thiback said:

Granted, since it fills a price gap that no other card does. And the Nano has the advantage of being really small. But I find that's all there is about it. Fury X wasn't a real great card either.

Unless you need compute power - where even a 390X beats the 980 TI

[Valentyn]

1 hour ago, Misanthrope said:

I am not ignoring it I am implying that is late for a reason: engineering problems. Nvidia just jumped ahead of them in tech again, even if it's a few months it's significant since by the time Vega rolls around they'll be ready with full pascal so again, AMD will launch behind the times by months.

Wait, does that also mean NVIDIA are having issues getting GP100 and 102 out? So they can't compete at the Enthusiast level, and only the mid-high end; Nevermind having nothing for Mainstream cards.

[Thiback]

1 minute ago, don_svetlio said:

Unless you need compute power - where even a 390X beats the 980 TI

Compute power, yes. Not gaming power.
390x is Hawaii though. To me the shining card in their current lineup is the 390 with its price/perf ratio. Cons : it doesn't OC well, since it's already a rebranded/factory OC'ed Hawaii. But still, I would take one over the GTX 970.

[don_svetlio]

1 minute ago, Thiback said:

Compute power, yes. Not gaming power.
390x is Hawaii though. To me the shining card in their current lineup is the 390 with its price/perf ratio. Cons : it doesn't OC well, since it's already a rebranded/factory OC'ed Hawaii. But still, I would take one over the GTX 970.

The 390 has a reworked power delivery thus most 390s hit 1150-1200 whereas most 290s couldn't go past 1125

[Derangel]

13 minutes ago, don_svetlio said:

Thing is - the things he talks about go against what other outlets report. I'm sorry but PCper, Anandtech, TechPowerUp, Arstechnica are far more credible in this point of time

We'll see. I haven't used Anand as a trusted source since the owner left. Like I said earlier on the page the truth is probably somewhere in the middle. Nothing Kyle states would really be all the surprising if it were true. AMD claims they don't want to target high-end, but there is no more reason to fully believe their statements then there is to take Kyle's article without a least a few grains of salt. There is no reason AMD would willingly back away from competing in the high end GPU market unless they are simply unable to or that planned high-end parts were unable to deliver the performance they wanted. The market share they've pissed away is entirely the fault of horrible mismanagement from the top, so RTG wanting to bail out of the sinking should would not be at all surprising.

[don_svetlio]

2 minutes ago, Derangel said:

We'll see. I haven't used Anand as a trusted source since the owner left. Like I said earlier on the page the truth is probably somewhere in the middle. Nothing Kyle states would really be all the surprising if it were true. AMD claims they don't want to target high-end, but there is no more reason to fully believe their statements then there is to take Kyle's article without a least a few grains of salt. There is no reason AMD would willingly back away from competing in the high end GPU market unless they are simply unable to or that planned high-end parts were unable to deliver the performance they wanted. The market share they've pissed away is entirely the fault of horrible mismanagement from the top, so RTG wanting to bail out of the sinking should would not be at all surprising.

Thing is - AMD have never said that they aren't competing - they've been hyping Vega as the answer to high end Pascal, Polaris has always been intended as a low-power performance option.

[Carclis]

3 minutes ago, don_svetlio said:

Thing is - AMD have never said that they aren't competing - they've been hyping Vega as the answer to high end Pascal, Polaris has always been intended as a low-power performance option.

In addition to this the 1080 is only a 300mm chip whereas the 380/x is reported to be 230mm. Both of them are really low/mid end cards in that regard which is probably telling that yields are far too low on the new process node for it to be economical to produce high end chips. At some point next year we should hopefully see something that is 400-500mm.

[Trixanity]

10 minutes ago, Derangel said:

We'll see. I haven't used Anand as a trusted source since the owner left. Like I said earlier on the page the truth is probably somewhere in the middle. Nothing Kyle states would really be all the surprising if it were true. AMD claims they don't want to target high-end, but there is no more reason to fully believe their statements then there is to take Kyle's article without a least a few grains of salt. There is no reason AMD would willingly back away from competing in the high end GPU market unless they are simply unable to or that planned high-end parts were unable to deliver the performance they wanted. The market share they've pissed away is entirely the fault of horrible mismanagement from the top, so RTG wanting to bail out of the sinking should would not be at all surprising.

Well, problem is RTG can't do shit without approval and how likely is it for AMD to sign off on their currently (probably) best asset?

 

Vega is the high end. They're not backing away. HBM is what they want for the high end and that's not ready yet. Even GDDR5X isn't ready. Nvidia went with pre-mature slow (ish) chips because they didn't want to wait. It's not like Polaris somehow failed and that they'll try again within 6 months. Chip design doesn't work like that. This has been planned at least a year in advance. Polaris was meant to be the first because the chips are small and would work better with a brand new process node. The yields will be much better for a small chip while the process matures and the assembly lines are running at full speed. Nvidia is shooting itself in the foot by trying to go for bigger chips when they can barely meet demand.

 

Polaris has always been about good performance at low power which is good for the high volume mobile market. That's where they want to be because that's where the big bucks are. And that's what they need. Selling 200k $600-700 cards is nothing if they could be selling millions upon millions of mobile chips to Apple, Dell, HP etc. and boosting their market share, getting design wins left and right and ensuring partnerships with large OEMs. Being able to use Polaris in mid range desktop is the cherry on top.

[Derangel]

3 minutes ago, Trixanity said:

Well, problem is RTG can't do shit without approval and how likely is it for AMD to sign off on their currently (probably) best asset?

 

Vega is the high end. They're not backing away. HBM is what they want for the high end and that's not ready yet. Even GDDR5X isn't ready. Nvidia went with pre-mature slow (ish) chips because they didn't want to wait. It's not like Polaris somehow failed and that they'll try again within 6 months. Chip design doesn't work like that. This has been planned at least a year in advance. Polaris was meant to be the first because the chips are small and would work better with a brand new process node. The yields will be much better for a small chip while the process matures and the assembly lines are running at full speed. Nvidia is shooting itself in the foot by trying to go for bigger chips when they can barely meet demand.

 

Polaris has always been about good performance at low power which is good for the high volume mobile market. That's where they want to be because that's where the big bucks are. And that's what they need. Selling 200k $600-700 cards is nothing if they could be selling millions upon millions of mobile chips to Apple, Dell, HP etc. and boosting their market share, getting design wins left and right and ensuring partnerships with large OEMs. Being able to use Polaris in mid range desktop is the cherry on top.

 

GDDR5X supply issues will be sorted out eventually. I'd argue that AMD is also shooting themselves in the foot by having nothing new on the high-end this year. The earliest we can expect Vega is probably Q2 of next year, if not later. Leaving Nvidia nearly a year to sort out supply issues and get something else (like a 1080ti and potentially a price drop) ready to counter AMD as soon as Vega launches. The mid-range is very important, but even with kind of having the best mid-range options (especially in the $300 and under segments) for a few years AMD has still taken a drastic loss in market share. AMD's mid-range and mobile options are all over the place, but it's not helping. If AMD can capture that $300-$350 market that Nvidia currently dominates then they will probably see their market share return. It'll be interesting to see how Polaris 10 does against the 1070.