From ATI to AMD back to ATI? A Journey in Futility @ [H]

Roy refused to supply hardware to sites that didn’t review their other stuff in anything other than a positive light. This often resulted in good products not getting reviewed and being mostly over looked.
 
Roy refused to supply hardware to sites that didn’t review their other stuff in anything other than a positive light. This often resulted in good products not getting reviewed and being mostly over looked.


Yeah the Nano was the BEST card in the Fury line up and they didn't want to have reputable sites to review it because they thought they weren't going to get good reviews? That card actually competed well with the gtx 980, beating it and consuming a little bit more power. Think its perf/watt was right around the gtx 980 or just a bit better.
 
No one is interested in stopping you from buying AMD products. What you do with your money is none of my or others concern. What interests me is why you claim Roy 'saved' AMD and why you would care enough to post this in a thread that is replete with evidence to the contrary.
FWIW, he doesn't actually care to post evidence for anything he states, it's a recurring theme obviously, no surprise there.
 
I RESURRECT THIS FROM THE GRAVE looks like a enclave of ati has formed at the intel https://wccftech.com/darren-mcphee-...-director-joins-intel-discrete-graphics-team/

Interesting, but also possibly not when you think of the people that have left AMD and joined Intel have not had much to show for their time at AMD. These people are pretty much responsible for the demise of AMD graphics over the last 5 years, and responsible for the state the division is in now. Personally, I think these people leaving AMD for Intel might actually be better for AMD than it is for Intel.
 
Interesting, but also possibly not when you think of the people that have left AMD and joined Intel have not had much to show for their time at AMD. These people are pretty much responsible for the demise of AMD graphics over the last 5 years, and responsible for the state the division is in now. Personally, I think these people leaving AMD for Intel might actually be better for AMD than it is for Intel.
And don't overlook the fact that in that huge organisation these people are even more restrained than at AMD... Intel is a tough place to work.
 
Interesting, but also possibly not when you think of the people that have left AMD and joined Intel have not had much to show for their time at AMD. These people are pretty much responsible for the demise of AMD graphics over the last 5 years, and responsible for the state the division is in now. Personally, I think these people leaving AMD for Intel might actually be better for AMD than it is for Intel.
I agree. All these marketing folks made a joke out of every launch. The drums, the rebellion crap or what not. Fury the over locking king lol. Poor Volta. I mean never understood why you would talk shit about other company and hype yours when you have nothing to back it up with. If there is one thing that turned me off about RTG was their marketing and their hype tactics. To that I say good riddance, Raja is bringing his crew, good luck. Raja and team is already showing his hype videos about their GPU almost 2 years from release. Already shows they are going to be doing similar marketing at intel.
 
I agree. All these marketing folks made a joke out of every launch. The drums, the rebellion crap or what not. Fury the over locking king lol. Poor Volta. I mean never understood why you would talk shit about other company and hype yours when you have nothing to back it up with. If there is one thing that turned me off about RTG was their marketing and their hype tactics. To that I say good riddance, Raja is bringing his crew, good luck. Raja and team is already showing his hype videos about their GPU almost 2 years from release. Already shows they are going to be doing similar marketing at intel.

Yeah but did you follow AMD marketing closely ? I mean there trying stuff out of desperation to put a dent into Nvidia armour and that did not go well they even hired the Nvidia guy that more or less made things happen.

I would not doubt Raja Koduri ability to make something work in the gpu department especially if his hands are not bound to a constrained budget.
 
Yeah but did you follow AMD marketing closely ? I mean there trying stuff out of desperation to put a dent into Nvidia armour and that did not go well they even hired the Nvidia guy that more or less made things happen.

I would not doubt Raja Koduri ability to make something work in the gpu department especially if his hands are not bound to a constrained budget.

I have my doubts at his ability to lead, He is great at hyping a product tho.
 
Who knows, maybe Raja will give us an Intel GPU two years from now that will help knock nVidia down a notch. Maybe AMD will have something better than mid-range at the same time. I can dream can't I? Who knows, maybe 7/10nm will be the best we can do without radical changes in production processes. If everyone hits that wall, maybe we can get a little parity.
 
Yeah but did you follow AMD marketing closely ? I mean there trying stuff out of desperation to put a dent into Nvidia armour and that did not go well they even hired the Nvidia guy that more or less made things happen.

I would not doubt Raja Koduri ability to make something work in the gpu department especially if his hands are not bound to a constrained budget.

I think raja is a great engineer. But I honestly don’t believe he has the leadership skills that hold people accountable as well as coach them to succeed. I do think he probably is a boss that’s everyone loves to work for because he is chill and not likely to hold people accountable.

I mean Ryzen did decent without having the budget of intel. Rx 480 pci-e power debacle was clearly a fail and should not have happened and had nothing to do with budget. That Was a clear management failure.

Also I highly doubt Vega has broken feature because of budget. It was leadership failure as well. Work culture can be chill and relaxed but If there is not much accountability shit will likely not go as planned.

I think if Raja as an engineer and as a project manager will be much better than a guy who handles the business side of things as well. I think if his only focus was Vega and he just handled that portion it might have turned out better. It felt like business side of things got the best of engineering side of things on the first go around. To much focus on internal politics and trying to get rtg to intel. When he went to intel I firmly believed the guy was madly in love with intel and was more interested in rtg being part of intel than actual products. I am not sure the rx 480 pci-e and Vega with broken features, may be it was just him being mad at Lisa lol.
 
To be honest that PCI-E thing was nothing worth mentioning it got blown out of proportion and it took software to fix it.

I would not know what was going on at AMD at that time some suggest that it was to sell off RTG to Intel all that I do know is that there are no others that came into the market when AMD was really weak.

That about sums it up if it was easy then others would happily cash the billions either from Nvidia side or from Intel and yet no one did.

Vega shows how bad things were at AMD their most innovative project which came short of delivering a comeback and were not financially a success(consumer) . But cancelling Vega would have been worse in the end they did use Vega in smaller forms and it shows that R&D was very limited. The Vega GPU could not push forward and needed HBM in order to match mid range cards. With more money that would not have happened.

And during Koduri's stay at AMD he was not able to do anything regarding Vega and Polaris were projects already ongoing.

And that is maybe why those rumours were ongoing while Koduri was there because what he saw could not be fixed by anything else but a better budget and that was not the case with AMD back then.
 
To be honest that PCI-E thing was nothing worth mentioning it got blown out of proportion and it took software to fix it.

I would not know what was going on at AMD at that time some suggest that it was to sell off RTG to Intel all that I do know is that there are no others that came into the market when AMD was really weak.

That about sums it up if it was easy then others would happily cash the billions either from Nvidia side or from Intel and yet no one did.

Vega shows how bad things were at AMD their most innovative project which came short of delivering a comeback and were not financially a success(consumer) . But cancelling Vega would have been worse in the end they did use Vega in smaller forms and it shows that R&D was very limited. The Vega GPU could not push forward and needed HBM in order to match mid range cards. With more money that would not have happened.

And during Koduri's stay at AMD he was not able to do anything regarding Vega and Polaris were projects already ongoing.

And that is maybe why those rumours were ongoing while Koduri was there because what he saw could not be fixed by anything else but a better budget and that was not the case with AMD back then.

I disagree with Vega not being a consumer success, sure it was not the 1080ti challenger many hoped for but they sold like mad due to mining. Which means AMD sold every Vega they could produce. The mistake made with Vega was by Raja hyping the thing up to something it was never going to be. I believe Raja to be a good engineer but he is no leader by what I have seen and heard, perhaps he will change my mind at Intel. Biggest issue for Vega is the margins were really slim so in the end selling all those cards did not make a huge profit for AMD. Personally I am looking forward to see what Navi will offer cause Nvidia lost their minds with the 2000 series and the price they think it's worth.
 
I disagree with Vega not being a consumer success, sure it was not the 1080ti challenger many hoped for but they sold like mad due to mining. Which means AMD sold every Vega they could produce. The mistake made with Vega was by Raja hyping the thing up to something it was never going to be. I believe Raja to be a good engineer but he is no leader by what I have seen and heard, perhaps he will change my mind at Intel. Biggest issue for Vega is the margins were really slim so in the end selling all those cards did not make a huge profit for AMD. Personally I am looking forward to see what Navi will offer cause Nvidia lost their minds with the 2000 series and the price they think it's worth.

I hope Navi has some architecture changes that makes AMD more efficient with their teraflops. As of right now nVidia gimped themselves by about 25% (or could have had 33% higher) CUDA performance since they incorporated RTX features that do nothing.

I haven’t personally seen anything solid on Navi although I don’t particularly look. I would do anything I could right now to capitalize on this if I was AMD.

On the flip side I believe I heard this arguement on the full nerd podcast. If I nVidia and wanted to push ray tracing tech now is the perfect time to do it (gimp your cards) considering the market.
 
I disagree with Vega not being a consumer success, sure it was not the 1080ti challenger many hoped for but they sold like mad due to mining. Which means AMD sold every Vega they could produce. The mistake made with Vega was by Raja hyping the thing up to something it was never going to be. I believe Raja to be a good engineer but he is no leader by what I have seen and heard, perhaps he will change my mind at Intel. Biggest issue for Vega is the margins were really slim so in the end selling all those cards did not make a huge profit for AMD. Personally I am looking forward to see what Navi will offer cause Nvidia lost their minds with the 2000 series and the price they think it's worth.

AMD paid $160 for the memory alone , do the math ...
And consumer that pre ordered got the launch price the rest did not ...
 
AMD paid $160 for the memory alone , do the math ...
And consumer that pre ordered got the launch price the rest did not ...

Unless you can find a signed contract and pricing then your just using analyst speculation but no doubt using it hurt the bottom line. You would be surprised what a big manufacturer pays compared to what a analyst thinks they pay, more often then not the analyst is a fair amount higher.
 
Unless you can find a signed contract and pricing then your just using analyst speculation but no doubt using it hurt the bottom line. You would be surprised what a big manufacturer pays compared to what a analyst thinks they pay, more often then not the analyst is a fair amount higher.

In my experience it’s 30-50% of “list”.
 
I disagree with Vega not being a consumer success, sure it was not the 1080ti challenger many hoped for but they sold like mad due to mining.

Which means that AMD moved units and processed revenue all while losing market share. Money is good, but as you'll note from Nvidia's and Intel's successes in the desktop market, if you don't have market share you get to play from behind- doubly so if your products aren't competing with your competitors' best.
 
Which means that AMD moved units and processed revenue all while losing market share. Money is good, but as you'll note from Nvidia's and Intel's successes in the desktop market, if you don't have market share you get to play from behind- doubly so if your products aren't competing with your competitors' best.

It's been pretty flat since the 1000 series launched. But your kind of wrong where it's been heading since Vega.
Q3-2017-Discrete-GPU-Market-Share-NVIDIA-AMD_1.png


AIB-001-1.png
 
The mistake made with Vega was by Raja hyping the thing up to something it was never going to be. I believe Raja to be a good engineer but he is no leader by what I have seen and heard, perhaps he will change my mind at Intel.

He got bent over a bit with the R&D budget at AMD. That said, after this push with Intel there will be no more excuses. And the approach at AMD sure seems to have been the right one. Raja and Vega were unfortunate collateral damage.
 
It's been pretty flat since the 1000 series launched. But your kind of wrong where it's been heading since Vega.
View attachment 116234

View attachment 116235

I think selling decent midrange parts works. AMD just needs to make Navi work well. If it hits home run and gets close to 1080-1080ti performance and stays below 300-350. They should win some market share for sure.
 
It's been pretty flat since the 1000 series launched. But your kind of wrong where it's been heading since Vega.
View attachment 116234

View attachment 116235

Wonder if add in mining market... another interesting market is workstation numbers - Vega 64 like 1080 Ti, top cards sell few in units even though they tend to pave way for the technology behind them and often carry higher profits - also, COD BO4 as well as BF 5, some benchmarks (updated ones with current drivers etc...) the Vega 64 goes head to head with the beloved 1080 Ti. Its when you add-in the Titan Xp that AMD had nothing (near) to compete, they (AMD) have had cards to mix it up with the stronger selling nvidia parts (1060/70 with Vega 56/64 and down, availability, price, and consumer choice hurt (in this market noted above in unit/market sales) AMD, its as simple as that.
 
He got bent over a bit with the R&D budget at AMD. That said, after this push with Intel there will be no more excuses. And the approach at AMD sure seems to have been the right one. Raja and Vega were unfortunate collateral damage.

Yeah the excuse train will die for him there. I just dont like that suddenly Intel started bragging about what is coming graphic wise years in advance. AMD seems to be executing well now so I am hopeful this means things will get better on all fronts and not just the cpu race. Biggest thing people forget tho is the mid range card or what we consider lower end is what really sells well and AMD is quite competitive with Nvidia there.
 
From the looks of that chart, AMD must be due to hit a home run. It's slowly moving further apart.
 
Yeah the excuse train will die for him there. I just dont like that suddenly Intel started bragging about what is coming graphic wise years in advance. AMD seems to be executing well now so I am hopeful this means things will get better on all fronts and not just the cpu race. Biggest thing people forget tho is the mid range card or what we consider lower end is what really sells well and AMD is quite competitive with Nvidia there.

Competitive on paper, but not sales. People buy Nvidia Midrange because Nvidia owns the top-range. People would rather buy a 1050Ti for more money than an RX570 for less, because Nvidia's name is on the top of the charts. The fastest cards, the most badass systems, those "Dream Systems" all have Nvidia's logo plastered all over them in big RGB LED letters. They have the mindshare because of this.

The high-end video card market is worth WAAAAAY more than the direct profit it makes from sales. The high-end cards are what create the brand. Yeah, Nvidia makes MOST of their money from the midrange, but nobody would consider their mid-range if Nvidia's high-end wasn't holding the carrot on a stick.
 
Competitive on paper, but not sales. People buy Nvidia Midrange because Nvidia owns the top-range. People would rather buy a 1050Ti for more money than an RX570 for less, because Nvidia's name is on the top of the charts. The fastest cards, the most badass systems, those "Dream Systems" all have Nvidia's logo plastered all over them in big RGB LED letters. They have the mindshare because of this.

The high-end video card market is worth WAAAAAY more than the direct profit it makes from sales. The high-end cards are what create the brand. Yeah, Nvidia makes MOST of their money from the midrange, but nobody would consider their mid-range if Nvidia's high-end wasn't holding the carrot on a stick.

The reality is that may make people take a look at your stuff but most people are looking at value over absolute performance, kind of how the Japanese rode to power in cars over the American companies and now how Hyundai has risen to be a force despite being the low end selection.
 
Competitive on paper, but not sales. People buy Nvidia Midrange because Nvidia owns the top-range. People would rather buy a 1050Ti for more money than an RX570 for less, because Nvidia's name is on the top of the charts. The fastest cards, the most badass systems, those "Dream Systems" all have Nvidia's logo plastered all over them in big RGB LED letters. They have the mindshare because of this.

The high-end video card market is worth WAAAAAY more than the direct profit it makes from sales. The high-end cards are what create the brand. Yeah, Nvidia makes MOST of their money from the midrange, but nobody would consider their mid-range if Nvidia's high-end wasn't holding the carrot on a stick.

I believe what you are saying about what high end cards do towards branding. But honestly I think AMD is selling mid range cards quiet well. If they weren’t they should be hurting a ton in Market share right now. I do think rx series is doing quiet well where it’s priced much better than GTX mid range cards. Sure nvidia might be selling more on midrange but I do believe AMD is doing well with rx 580 series with the promotions they have had going on.
 
The reality is that may make people take a look at your stuff but most people are looking at value over absolute performance, kind of how the Japanese rode to power in cars over the American companies and now how Hyundai has risen to be a force despite being the low end selection.

That's half the story, Japanese and Korean cars rose to popularity after proving beyond a shadow of any doubt that they are more reliable and safe. That's not a super big priority with video cards, as Nvidia has had TONS of card-killing issues and people still love them.
 
Unless you can find a signed contract and pricing then your just using analyst speculation but no doubt using it hurt the bottom line. You would be surprised what a big manufacturer pays compared to what a analyst thinks they pay, more often then not the analyst is a fair amount higher.

Well it is simple it is expensive if it was not every card would have had HBM2 at that time and guess which other products had HBM2 that was Nvidia in the professional segment plus there was a shortage at the time as well which was posted about as well, until SK Hynix came with their offer which was also pretty limited Samsung was the only provider for HBM2.

In my experience it’s 30-50% of “list”.

Unless the both of you produce a signed contract I will stick with my assessment of too fucking expensive and if that is $160 or more who cares it is just something which did not help Vega or AMD. The HBM was a necessity rather then a better option for performance. And since the supply was short the analysts prolly got a conservative estimate.
 
Last edited:
Well it is simple it is expensive if it was not every card would have had HBM2 at that time and guess which other products had HBM2 that was Nvidia in the professional segment plus there was a shortage at the time as well which was posted about as well, until SK Hynix came with their offer which was also pretty limited Samsung was the only provider for HBM2.



Unless the both of you produce a signed contract I will stick with my assessment of too fucking expensive and if that is $160 or more who cares it is just something which did not help Vega or AMD. The HBM was a necessity rather then a better option for performance. And since the supply was short the analysts prolly got a conservative estimate.

He said mfgs get large discounts. I agreed and said what was typical. Let me go back a few jobs and ask procurement for a signed contract. You’re hilarious.

AMD never needed HBM. They completed fucked that up and the vast majority here called out multiple reasons why it was a dumb decision months before launch.
 
Last edited:
He said mfgs get large discounts. I agreed and said what was typical. Let me go back a few jobs and ask procurement for a signed contract. You’re hilarious.

AMD never needed HBM. They completed fucked that up and the vast majority here called out multiple reasons why it was a dumb decision months before launch.

For Vega they needed HBM otherwise that card would have used over 500 Watt with GDDR5 and still not be able to run high clocks.
Checkout Buildzoid he tested Vega and explained why it would not have worked with GDDR5.

Since when do you get large discounts when there is scarcity?
 
For Vega they needed HBM otherwise that card would have used over 500 Watt with GDDR5 and still not be able to run high clocks.
Checkout Buildzoid he tested Vega and explained why it would not have worked with GDDR5.

Since when do you get large discounts when there is scarcity?

Those contracts are generally written up way before scarcity is known and the sales/procurement teams are optimistic.

There’s only ~25W difference between GDDR and HBM.
 
Back
Top