The mindset powering AMD's Radeon: people and tech first, business 2nd

[Delicieuxz]

Forbes: How AMD's 'The Uprising' Ad Campaign Will Incite A Radeon Rebellion

 

Article excerpt:

Quote

... With a bold new logo and the notion of the pixel as their guiding light, AMD got to work creating a campaign that would be quite different than what we’ve seen in the past.

But before we look ahead, I had a burning question for Koduri: They forged forward based on what they wanted to be, but I was curious to learn what they didn’t want to be. “We’ve discussed this internally, and our goals are not based on how to be a $10 billion company,” Koduri answers. “They’re based on how do we create these pixels super fast and super smooth?”

Koduri went on to explain that although their explicit goal is not to be a great business brand, that could inevitably happen. Obviously they want to make money so they can in turn create more advanced chips, but Koduri and company believe that if they can capture pixels, the rest will fall into place.

Chris Hook interjects: “What was cool about Raja’s vision was that it was very much ‘if we focus on community first, and we listen carefully to what we get tweeted at everyday, and we make the customer first, it may take us some time to build the business but we’re going to be on the right track.”
...

While Radeon Technologies Group was conceptualizing Polaris, they kept their ear to the ground and learned what their community wanted right now, but also what their community desired 2 years from now. Hook says this resulted in five key takeaways:

1: Prestige. “They wanted the prestige of a $700 graphics card, but they didn’t want to have to pay for it,” Hook begins.

2: VR that just works. “They wanted the ability to have a great VR experience today or two years from now without worrying about upgrading power supplies and digging into their PC. They wanted to buy a headset at some point and just have it work.”

3: Respect their investment. “They wanted us to pay respect to the dollars they were giving us and do things in the architecture or transistors or APIs or ASync Compute, that provided a measure of ‘futureproofness.’ They wanted to be reassured that even if they’re only spending $200 they’d feel secure in their investment for a couple years.”

4: More overclocking control. “We brainstormed what kind of voltage control could be given to them to create a better experience.

5: Better drivers. “We feel we’ve made a great first step there, and we’re only going to be putting a heavier foot on the gas this year and next year to make those drivers better and better.”

Then Hook gets to the point and wraps everything in a bow. And I almost laughed outloud. “So it’s becoming less of a product launch and in our view more of a movement.” The moment this phrase leaves his lips, he actually acknowledges how unbelievable it sounds. “As a marketer, I almost don’t believe myself as I’m saying that. But this is where it became real to me…”

 

article continues...

 

I think that AMD has the right approach, and it brings to my recollection some things I've had to say about AMD recently. One of those things is this:

 

On 6/7/2016 at 8:56 PM, Delicieuxz said:

I don't think AMD is more open for the sake of staying competitive, and I'd guess that being as open as AMD is has lessened their market leverage, over time - but it has improved the state of technology for everyone, and every person running a home PC today is running a bit of "team red."

 

Innovation isn't pulled out of a hat in response to being under pressure. An ingenious mindset is consistent in its forward progress and prolific creativity - and that mindset often holds the idealism of free and open information [to be of] high value.

 

And another of those things has been noting that AMD pioneers a lot of good technology that becomes mainstream, in this thread, for example:

 

GDDR

HBM

APUs

GPU tessellation

Mantle (-> Vulkan / DX12)

multi-core x86 CPUs

on-die memory controllers

DisplayPort

64-bit x86 CPUs

async-computation in GPUs

 

.. and AMD is also committed to an open standard, which has allowed their tech to benefit the world, including their rival companies, like Intel, and Nvidia.

[Nena Trinity]

Awesome! :3

[niofalpha]

Obligatory Forbes comment about how the article just drops the format of a list of goals. Great writing there.

 

I always have disliked AMD's marketing... Just sounds childish, full of false promises, and just stupid. This continues the trend. Also whats with AMD "Reinventing" themselves so often? 7xxx to Rx( x = variable) 2xx to Fury to RX (X =/=  variable). It's just getting annoying at this point. I prefer RX to R9/R7/R5 but it's still annoying how they keep changing it, along with the placement of each card on the performance tree.

Addressing the 5 keynotes

1. That "prestige" you want isn't gotten because you don't have the E-Peen of a 700$ GPU, but of a 200$ one. Lower the market price, lower the E-peen. 

2. VR that works heard that before. And AMD saying "No VR for the 1%" is just dumb. My 4 monitors + stand cost about 200$ less than the Occulus. GPU power isn't whats keeping it from the 1%, and unless AMD releases one themselves they can't really do this.

3.  Common sense aim of any Hardware manufacturer...

4. Heard that before, that ended up nicely...

5. Common sense aim of any Hardware manufacturer...

 

Then the posters

It's an improvement over the fixer, but I still think it's stupid. I feel as if AMD's focusing too much on the 12-15 year old group with those adds. They still feel too gamery, but a bit better than the fixer tho... "The VR for the 1%" poster is just stupid, since as I said earlier there's no point in making a GPU for VR and advertising it as cheap, if the actual VR headset costs 3x as much. If the performance benchmarks are true then it's on par with the 970/ 390, which puts it in the perfect position for a 1060 to role in around the same price point and give them competition (Can be ~67% the Size and outperform it if I did my math correctly). And that's gonna have issues running VR at a decent frame rate and higher settings. The rest seem like good goals to set, and I look forward to seeing them. Only other major complaint is it's pretty hypocritical. I'm not sure if I should direct this towards AMD, or Forbes as they say no "catchy slogans" then go on and say use Radeon Rebellion. That's not a catchy slogan, but it's still a slogan. 

[Beskamir]

I guess this is why AMD is losing money even though they are trying to be the good guys.

[Delicieuxz]

17 minutes ago, niofalpha said:

If the performance benchmarks are true then it's on par with the 970/ 390, which puts it in the perfect position for a 1060 to role in around the same price point and give them competition (Can be ~67% the Size and outperform it if I did my math correctly). And that's gonna have issues running VR at a decent frame rate and higher settings.

Recent stock-Mhz benchmarks suggest the RX 480 could be slightly faster than a GTX 980, and about as fast as an R9 Fury. It's being discussed in this thread:

 

[niofalpha]

4 minutes ago, Delicieuxz said:

Recent stock-Mhz benchmarks suggest the RX 480 is slightly faster than a GTX 980, and about as fast as an R9 Fury. It's being discussed in this thread:

 

About on par with the R9 Fury is just wrong... Slightly faster, but about on par with the 980 seems more accurate...

 

Saying as in that very thread you linked me, there's only a 5-6% difference between 980 and 480, yet a 11% between 480 and R9 Fury in those sketchy asian benchmarks. We'll see once we finally get legit benchmarks how it performs.

[Delicieuxz]

6 hours ago, niofalpha said:

About on par with the R9 Fury is just wrong... Slightly faster, but about on par with the 980 seems more accurate...

 

Saying as in that very thread you linked me, there's only a 5-6% difference between 980 and 480, yet a 11% between 480 and R9 Fury in those sketchy asian benchmarks. We'll see once we finally get legit benchmarks how it performs.

There are down-clocked (I think due to a BIOS issue) and stock Mhz benchmarks of the RX 480 being discussed in that thread. The Firestrike Ultra scores presented are these:

 

The GTX 980 gets a score of 3038

The down-clocked RX 480s get scores of 3194, 3268

The stock-clocked RX 480s get scores of 3460, 3513

The R9 Fury gets a score of 3587

 

So, it's, supposedly, notably faster than a GTX 980, while slightly slower than an R9 Fury... with current drivers.

[samcool55]

1 hour ago, Delicieuxz said:

Nope, there are down-clocked (due to bios issues) and stock Mhz benchmarks of the RX 480 being discussed in that thread. The Firestrike Ultra scores presented are these:

 

The GTX 980 gets a score of 3038

The down-clocked RX 480s get scores of 3194, 3268

The stock-clocked RX 480s get scores of 3460, 3513

The R9 Fury gets a score of 3587

 

So, it's, supposedly, notably faster than a GTX 980, while slightly slower than an R9 Fury... with current drivers.

I really want to know how the 480 will overclock.

Now AMD told us that they really kept overclocking in mind and want to give end-users a good OC experience i won't be suprised if the rx 480 beats the fury when overclocked.

Would be awesome  

[That Norwegian Guy]

I really hope the blower style cooler for the 480 will actually be possible to buy this time, because I'm sick of the heat getting to my CPU.

[Khvarrioiren]

3 minutes ago, That Norwegian Guy said:

I really hope the blower style cooler for the 480 will actually be possible to buy this time, because I'm sick of the heat getting to my CPU.

Watercooling is the answer

[That Norwegian Guy]

1 minute ago, Khvarrioiren said:

Watercooling is the answer

Sorry I hate watercooling aesthetically, and will never use it.

 

A beefy CPU air cooler looks so much better, kind of like an engine block 

[laminutederire]

AMD only issue is their image, in my opinion at least. On the gpu market they've always deliver good value product, but people always end up raping themselves buying overpriced products from our friends nvidia.

On the cpu market people have forgotten that they used to be the best a few years back, and that it's not because they are not so competitive right now that they can't make a really great cpu in the future.

But people always think poorly of them, and that's what hurts them the most.

That's probably why they try new marketing strategy to gain the market share they really deserve. It should be like 50/50 between them and nvidia when they deliver the same performance  and 75/25 for them when they make a better product. Instead it's more 25/75 on average and 50/50 when they make better products, which is for me the sign they have had bad marketing strategies for the masses 

[Trixanity]

I'm personally not a fan of this new campaign. Maybe I'm just not the target audience but it just doesn't appeal to me. 

 

I, sadly, think that their attempt to stir things up and try something new will fail. 

[Daiyus]

I find most of these "marketing campaigns" go over my head, as it does for most of us I'd guess. We already know that the product exists. We already know to wait for benchmarks and reviews. Anybody who isn't biased will buy whatever suits them best regardless of branding once they have all the raw information in front of them. Frankly as long as I know the thing exists then the marketing has done its job.

 

How this is going to impact the masses that don't follow on tech news though I have no idea; most people don't make sense to me.

[patrickjp93]

4 hours ago, Delicieuxz said:

Forbes: How AMD's 'The Uprising' Ad Campaign Will Incite A Radeon Rebellion

 

Article excerpt:

 

I think that AMD has the right approach, and it brings to my recollection some things I've had to say about AMD recently. One of those things is this:

 

 

And another of those things has been noting that AMD pioneers a lot of good technology that becomes mainstream, in this thread, for example:

 

GDDR

HBM

APUs

GPU tessellation

Mantle (-> Vulkan / DX12)

multi-core CPUs

on-die memory controllers

DisplayPort

64-bit x86 CPUs

async-computation in GPUs

 

.. and AMD is also committed to an open standard, which has allowed their tech to benefit the world, including their rival companies, like Intel, and Nvidia.

Intel made the first multi-core CPU. All AMD did is beat them to a monolithic die dual-core with inferior performance anyway. AMD sucks at tesselation. Nvidia has them beaten by a mile in that category. Intel beat them to APUs too and had the first chip with an iGPU several years before AMD.

[Delicieuxz]

47 minutes ago, patrickjp93 said:

Intel made the first multi-core CPU. All AMD did is beat them to a monolithic die dual-core with inferior performance anyway. AMD sucks at tesselation. Nvidia has them beaten by a mile in that category. Intel beat them to APUs too and had the first chip with an iGPU several years before AMD.

There's a list towards the bottom of this page that appears to be disagreeing some of the things you mention: http://www.amd.com/en-us/innovations/software-technologies/hbm

 

But if you've got specifics, please share em, I'd like to know.

 

It does list specifically multi-core x86 CPUs, and not just multi-core CPUs (I'll change the OP). Here's an article on the subject:

Tom's Hardware: AMD demos x86 dual-core processor

Quote

AMD has won another race against Intel by demonstrating the first dual-core processor.

...

The original announcement of dual-core processors dates back to October 1999, when AMD's CTO Fred Weber was talking about "two x86 engines on a single chip" in a presentation at the Microprocessor Forum. In an ongoing battle with Intel, which is reminiscent to the famous Gigahertz race between the two manufacturers, AMD once again has a brief lead - with Intel also working on dual-core processors.

 

Here's some background on the AMD record for the other things you mention:

 

https://en.wikipedia.org/wiki/AMD_Accelerated_Processing_Unit

Quote

The AMD Fusion project started in 2006 with the aim of developing a system on a chip that combined a CPU with a GPU on a single die. AMD took a key step toward realising such a vision when it acquired the graphics chipset manufacturer ATI[2] in 2006.
...
The first generation desktop and laptop APU, codenamed Llano, was announced on January 4, 2011 at the 2011 CES show in Las Vegas and released shortly after.

 

https://en.wikipedia.org/wiki/ATI_TruForm

Quote

ATI TruForm was a brand by ATI (now AMD) for a SIP block capable of doing a graphics procedure called tessellation in computer hardware. ATI TruForm was included into Radeon 8500 (available from August 2001 on) and newer products.[1]

The successor of the SIP block branded "ATI TruForm" was included into Radeon HD 2000 series (available from June 2007 on) and newer products: hardware tessellation with TeraScale.

Support for hardware tessellation only became mandatory in Direct3D 11 and OpenGL 4. Tesselation as defined in those APIs is only supported by even newer GCN-based products (available from January 2012 on).

 

I'm not debating which graphics tech developer does tessellation better, I only mentioned it as an example of tech becoming mainstream after AMD pioneered it.

[patrickjp93]

34 minutes ago, Delicieuxz said:

There's a list towards the bottom of this page that appears to be disagreeing some of the things you mention: http://www.amd.com/en-us/innovations/software-technologies/hbm

 

But if you've got specifics, please share em, I'd like to know.

 

It does list specifically multi-core x86 CPUs, and not just multi-core CPUs (I'll change the OP). Here's an article on the subject:

Tom's Hardware: AMD demos x86 dual-core processor

 

Here's some background on the AMD record for the other things you mention:

 

https://en.wikipedia.org/wiki/AMD_Accelerated_Processing_Unit

 

https://en.wikipedia.org/wiki/ATI_TruForm

 

I'm not debating which graphics tech developer does tessellation better, I only mentioned it as an example of tech becoming mainstream after AMD pioneered it.

I forget if it was Core or Nehalem that had the first iGPU, but regardless Intel beat AMD to market with the first fusion product. The first programmable was Sandy Bridge, so at best you could label that a tie if you want to swing for AMD.

 

And AMD didn't pioneer tesselation. At best ATI did, but it was simultaneously being developed by Nvidia, so it's disingenuous to credit either one with inventing it unless you credit both of them.

 

Sorry but multi-core Pentium D beat AMD to market, regardless of announcement dates. All AMD did was achieve a monolithic die first, and it didn't matter much.

 

GDDR5 was Nvidia and iirc Elpida (GDDR4 was a fiasco a la AMD).

[Sakkura]

2 hours ago, huilun02 said:

I hope RX 480 will be the 7950 reincarnate. Also a $200 card, tremendous OC potential, asic efficiency champion of its day.

The 7950 was a $450 card...

 

The RX 480 is more like the 7790 or 7850 of this generation (not 7870, that was a lot more expensive).

[Sakkura]

1 minute ago, ONOTech said:

480 is more like the 4850 from years back. 490 will be like the 4870 (assuming it's P10 as well).

Yeah philosophically it looks like AMD is going for a repeat of the 4000 series.

[patrickjp93]

8 minutes ago, ONOTech said:

480 is reminiscent of the 4850 from years back. 490 will be like the 4870 (assuming it's P10 as well).

490 is Vega. AMD already said P10 tops out at 2304 and the Rx 480 has 36 CUs which is the max possible spec assuming the SPs/CU didn't decrease from 64.

[patrickjp93]

1 minute ago, ONOTech said:

They also mentioned a $300 card at Computex I believe they'll unveil another card.

The smaller Vega chip tops out at 3072, so that's still consistent.

[Tomsen]

15 minutes ago, patrickjp93 said:

ISorry but multi-core Pentium D beat AMD to market, regardless of announcement dates. All AMD did was achieve a monolithic die first, and it didn't matter much.

 

GDDR5 was Nvidia and iirc Elpida (GDDR4 was a fiasco a la AMD).

"In April 2005, Intel's biggest rival, AMD, had x86 dual-core microprocessors intended for workstations and servers on the market, and was poised to launch a comparable product intended for desktop computers. As a response, Intel developed Smithfield, the first x86 dual-core microprocessor intended for desktop computers, beating AMD's Athlon 64 X2 by a few weeks"

 

Please list the launch date of Nvidias first GDDR5 product.

GDDR4 failure has nothing to do with AMD, unlike Nvidias history of transitioning for a new memory standards..

[Trixanity]

36 minutes ago, patrickjp93 said:

I forget if it was Core or Nehalem that had the first iGPU, but regardless Intel beat AMD to market with the first fusion product. The first programmable was Sandy Bridge, so at best you could label that a tie if you want to swing for AMD.

 

And AMD didn't pioneer tesselation. At best ATI did, but it was simultaneously being developed by Nvidia, so it's disingenuous to credit either one with inventing it unless you credit both of them.

 

Sorry but multi-core Pentium D beat AMD to market, regardless of announcement dates. All AMD did was achieve a monolithic die first, and it didn't matter much.

 

GDDR5 was Nvidia and iirc Elpida (GDDR4 was a fiasco a la AMD).

What I'm reading here is "AMD did it first, but others did it better, so it doesn't matter" and "AMD didn't do it first, but was better, so it doesn't matter". 

[Delicieuxz]

57 minutes ago, patrickjp93 said:

I forget if it was Core or Nehalem that had the first iGPU, but regardless Intel beat AMD to market with the first fusion product. The first programmable was Sandy Bridge, so at best you could label that a tie if you want to swing for AMD.

 

And AMD didn't pioneer tesselation. At best ATI did, but it was simultaneously being developed by Nvidia, so it's disingenuous to credit either one with inventing it unless you credit both of them.

 

Sorry but multi-core Pentium D beat AMD to market, regardless of announcement dates. All AMD did was achieve a monolithic die first, and it didn't matter much.

 

GDDR5 was Nvidia and iirc Elpida (GDDR4 was a fiasco a la AMD).

I think that's not a very legit disagreement.

 

AMD pioneered GDDR... which GDDR4/5 are developments of. You post looks like you're trying to occlude a recognition of GDDR, the foundation, by shifting focus to derivatives of the tech that AMD pioneered.

 

AMD bought ATi, and and today's AMD includes that which was ATi, which is a very big part of AMD, but you're right to point out that ATi did tessellation before AMD acquired them. But I think today's AMD is as much ATi as it is the original AMD, and that to use the AMD name is to recognize the merger of two intact groups.

 

AMD pioneered x86 multi-core CPUs... which is what I've said, and which isn't affected by Intel's release date. Although, Tomsen's post above clarifies that AMD had the first x86 multi-core CPUs on the market, regardless.

 

I think the only aspect of your post that has merit is that ATi was developing tessellation before they were bought by AMD.

[patrickjp93]

39 minutes ago, Tomsen said:

"In April 2005, Intel's biggest rival, AMD, had x86 dual-core microprocessors intended for workstations and servers on the market, and was poised to launch a comparable product intended for desktop computers. As a response, Intel developed Smithfield, the first x86 dual-core microprocessor intended for desktop computers, beating AMD's Athlon 64 X2 by a few weeks"

 

Please list the launch date of Nvidias first GDDR5 product.

GDDR4 failure has nothing to do with AMD, unlike Nvidias history of transitioning for a new memory standards..

Intel had them for servers before AMD as well.

 

The GDDR4 failure had everything to do with AMD. And Nvidia's first GDDR5 product was purchasable February 2010. The memory technology was developed with Infineon, Elpida, and Nvidia.

 

Just because Nvidia has historically had bumpy transitions doesn't mean it failed to assist in developing technologies.