No AMD Radeon "Navi" Before October: Report

[ArbitraryAffection]

Please no. I can't wait that long. I have the itch right now.

 

[64K]

I understand where AMD stands financially, BUT they should not lie to their own fans. I have been postponing GPU buy based on AMD's GPU roadmap for more than a year and when bloody Navi should at least get release date AMD goes silent and offers us abysmal overpriced and nowhere to be found R7 - or to be truthful quality failed and crippled Instinct MI150 instead. WTF, what kind of PR is this? They're doing marvelous job if they want to lose remaining 28 % share of PC gaming market. I always bought AMD for my gaming rig just to support an underdog and partially because of ATI nostalgia (I was forced to go green for my workstations because of Adobe) but I'm not willing to go red anymore. They should admit it if they don't have competitive architecture, close the doors and go back to their drawing boards until they have it, like they did with Ryzen. Saying to fans that they will compete and then offer us bullshit after bullshit is the worst thing they can do. Someone should get fired in their marketing department. Their last good release was Polaris 30 months back for god sake.

I understand your frustration and I agree with you that AMD needs to stop over promising and under delivering.

That may improve in the future. AMD lost their Marketing Director Chris Hook to Intel last year after 20 years with AMD.

Intel is serious about discrete GPUs imo.

 

[bug]

I understand your frustration and I agree with you that AMD needs to stop over promising and under delivering.

That may improve in the future. AMD lost their Marketing Director Chris Hook to Intel last year after 20 years with AMD.

Intel is serious about discrete GPUs imo.

So... AMD will hype us up about Intel's launches next?

 

[Gasaraki]

We don't know if Navi is GCN or isn't. All we know for sure is that it was designed for PS5, just as Polaris was designed for Xbox One X (has the full 40 compute units instead of the gimped, but higher clocked, 36 of the desktop cards).

Also, GCN has been evolving with each generation. Vega, for example, significantly changed the compute units (longer pipes allowing for higher clocks).

GCN is still a jack-of-all-trades and master-of-none architecture. I'm not convinced AMD wants to fork Radeon architectures from Instinct and Pro cards. They effectively need two engineering staffs to do that.

What GCN is really lacking is tiled rasterization. If Navi brings that, it will more align it with GeForce performance in gaming.

AMD did not design Polaris for XboxOneX. Xbox used the Polaris architechture and added some custom items that it wanted.

 

[Manu_PT]

AMD doesnt care about PC market, at least not high end. They care about what gives them profits, console. Navi being developed with sony. They know PS5 sales gonna obliterate pc hardware sales, like it is doing now.

All they care about is launching 2 mid range competitive cards for the battle royale or wannabe esports kids and their usual games, and thats it. Thats where 95% of the PC gaming revenue is anyway.

 

[Punkenjoy]

I wouldn't be surprised if AMD choose to do Ray Tracing with Brute force (just tons of extra computes units) instead of using a dedicated engine. That could mean much lower DXR/RTRT performance initially, but much less silicon wasted when game don't use these features and better performance in non raytracing games. It would also be more flexible than the current Nvidia solution. The developper could use the extra computes power to do something else than RTRT when not needed

That might explain why AMD isn't more vocal about having dedicated Raytracing core. That remind me a bit the switch from hardware T&L to software T&L using Vertex Shaders Engine (that got merged into Shaders Engine/Compute core eventually).

 

[ShurikN]

Don't need to be Jim Keller to realize (form this rumor), that AMD is prioritizing Ryzen, Epyc and Instinct. Now that GPU mining is dead, Radeon probably brings the least amount of money to AMD.
PS5 mass production will most likely happen early/mid next year for a Q4 2020 (holiday season) release date. So at least that's not eating into it.

GCN is still a jack-of-all-trades and master-of-none architecture.

Spot on. Maybe with money from CPUs they'll finally split the two in the future. Definitely not in time for Navi IMO, but hopefully for Arcturus (wishful thinking).

Please no. I can't wait that long. I have the itch right now.

Haha I feel you completely, but luckily for me I won't be returning to my home country until Q2 2020... Just in time for Ryzen 3000 + Navi combo.

 

[unikin]

How could Navi compete with Turing if it is still based on somewhat tweaked GCN? NVidia will surely move Turing to 7nm by 1H2020 and lower prices if needed while maintaining profit margin due to die shrink. AMD has missed the only time gape when they could hurt NVidia by offering mid range price/performance kings to the market and NVidia couldn't just lower their prices. All NVidia has to offer is RTX 2060 on 7 nm die for $300 and Navi will be DOA. It really looks like AMD doesn't give a flying f... about PC gaming community anymore and will serve us only slightly modified GPU console ports, which R&D was financed by Sony's capital anyway.

 

[notb]

Still, it matters little if Navi launches two-three months early or late. What matters is if it can take the fight to Nvidia. So far, it looks like it can't. To make things worse, in Q4'19 means it will go up against 7nm Turing (or whatever it will be called).

I wouldn't worry so much about 7nm, if it gives Turing a similar "boost" to what we see with Vega.
What I would worry about is 2nd gen DLSS and RTX.

Look how many games actually use DXR. Then look at the primary customer Navi is for: Sony who doesn't even use DirectX in their consoles. I just don't see tensor cores nor DXR as being a priority for AMD, especially not something to derail product timelines for. As the OP says, I think the delay is because of 7nm issues moreso than Navi itself. The fact there's limited availability of Radeon VII cards also hints at 7nm issues.

RTX is just an interface to RTRT. And RTRT itself can be run on any hardware that has an efficient RNG and can do basic arithmetics.
Both Nvidia and AMD have their own libraries for RTRT.

As for RTRT popularity, I think you're looking at this in the wrong way.
First of all: there are already few dozen games that are confirmed to use RTX or DLSS in future releases. This also includes titles that are very popular on consoles: Assetto Corsa, Hitman and Final Fantasy.
Second: RTRT is a feature that should be a total hit in consoles, IMO. Because while PC gamers care about fps, console people are very much into visual effects and they wish for better image quality. It's hard to believe console manufacturers will ignore RTRT.

They are a very, very small company compared to Intel and they are small compared to Nvidia as well.

People use this argument over and over again - like if the the whole Nvidia workforce was designing GPUs.
This job is done by a relatively small team of highly trained people. A team that AMD has as well, because clearly they do have some designs of their own. And they're pretty good from time to time.

Because AMD is small, we can expect bad customer support, less products, poor marketing and so on. But it shouldn't have such a huge impact on GPU designs.
Also, with Radeons you always have the sense that the engineers could have make them better. That they're held back by bad management decisions.

 

[64K]

Also, with Radeons you always have the sense that the engineers could have make them better. That they're held back by bad management decisions.

I agree with you about decisions to limit R&D in their GPU business a management issue and that was the main reason I think Koduri left AMD for Intel. He wanted to do more but Lisa Su was focused on Ryzen. It wasn't a bad decision on her part though. AMD is showing a profit for a while now and bear in mind when AMD was primarily focused on their GPU business and letting their CPU business mostly stagnate AMD was up to their necks in red ink. They had all 3 console contracts but were posting losses in the hundreds of millions of dollars year after year.

 

[bug]

I wouldn't worry so much about 7nm, if it gives Turing a similar "boost" to what we see with Vega.
What I would worry about is 2nd gen DLSS and RTX.

Disregarding 2nd gen DLSS and/or RTX, given no performance boost at all, 7nm will make the die smaller, which would mean sane prices for GPUs. I would worry about that

AMD doesnt care about PC market, at least not high end. They care about what gives them profits, console. Navi being developed with sony. They know PS5 sales gonna obliterate pc hardware sales, like it is doing now.

All they care about is launching 2 mid range competitive cards for the battle royale or wannabe esports kids and their usual games, and thats it. Thats where 95% of the PC gaming revenue is anyway.

Yeah, no, that's not it. It goes up and down, but their income from "computing and graphics" (which includes GPUs) is about the same as their revenue from "enterprise, embedded and semi-custom" (which includes consoles.
http://ir.amd.com/news-releases/new...-reports-third-quarter-2018-financial-results <- scroll down to "Segment and Category Information".
Granted, it's not that granular, but I wouldn't say they will be ignoring either segment.

 

[Divide Overflow]

Navi will be late and only targeted at low and mid-range performance tiers?
If Radeon VII is AMD's excuse at a performance part this year, competition from them for enthusiast gamers is almost all but dead.

 

[efikkan]

AMD got their engineering samples back in October 2018, and while they were initially reported to be looking "good", they quickly became silent and rushed in a Vega20 product for CES. Navi is still lacking from launch schedules, and there is likely something wrong with the chip, at least the one intending to compete with RTX 2070/2080.

I wouldn't be surprised if "big" (consumer) Navi slips into very late 2019 or early 2020, and at this pace it will be facing the next generation from Nvidia.

 

[unikin]

I wouldn't be surprised if "big" (consumer) Navi slips into very late 2019 or early 2020, and at this pace it will be facing the next generation from Nvidia.

True, without new architecture Navi will be DOA. GCN can't even compete with pascal, Turing is generation ahead and tweaked 7nm Turing with higher clock boosts will be 2 generations ahead in 2020 if Navi doesn't have completely new die arch. I really hope AMD gets serious with R&D or they will lose all PC gaming market, especially with blue giant insight to compete with both.

 

[M2B]

We don't know if Navi is GCN or isn't. All we know for sure is that it was designed for PS5, just as Polaris was designed for Xbox One X (has the full 40 compute units instead of the gimped, but higher clocked, 36 of the desktop cards).

Polaris was't designed for Xbox One X.
Polaris was in development way before development of Xbox One X.

 

[eidairaman1]

How could Navi compete with Turing if it is still based on somewhat tweaked GCN? NVidia will surely move Turing to 7nm by 1H2020 and lower prices if needed while maintaining profit margin due to die shrink. AMD has missed the only time gape when they could hurt NVidia by offering mid range price/performance kings to the market and NVidia couldn't just lower their prices. All NVidia has to offer is RTX 2060 on 7 nm die for $300 and Navi will be DOA. It really looks like AMD doesn't give a flying f... about PC gaming community anymore and will serve us only slightly modified GPU console ports, which R&D was financed by Sony's capital anyway.

Stop making assumptions, none of us know what Navi Brings, we are not AMD engineers here

Navi will be late and only targeted at low and mid-range performance tiers?
If Radeon VII is AMD's excuse at a performance part this year, competition from them for enthusiast gamers is almost all but dead.

Stop gap card, it's a RI card that had yield problems to not be a full RI card. It will suffice for them until Q4 2019+

Pretty much a Radeon Pro

 

[efikkan]

True, without new architecture Navi will be DOA. GCN can't even compete with pascal, Turing is generation ahead and tweaked 7nm Turing with higher clock boosts will be 2 generations ahead in 2020 if Navi doesn't have completely new die arch. I really hope AMD gets serious with R&D or they will lose all PC gaming market, especially with blue giant insight to compete with both.

Development of Navi is sponsored by Sony and is primarily intended for the next upcoming PlayStation. PC gamers will get products derived from the same architecture, and as AMD have earlier stated it will offer "Vega level performance". Since the beginning, Navi has been referred to as the final iteration of GCN. There is nothing indicating that Navi will be a major improvement over Vega, but rather a more efficient successor of Vega.

 

[Super XP]

With Navi still being GCN it is probably wont be much of a challenger to current nvidia anyway.

Navi will be a new microarchitecture (in other words the first AMD Radeon uArch to NOT be based on GCN).Nov 10, 2018

So... AMD will hype us up about Intel's launches next?

Development of Navi is sponsored by Sony and is primarily intended for the next upcoming PlayStation. PC gamers will get products derived from the same architecture, and as AMD have earlier stated it will offer "Vega level performance". Since the beginning, Navi has been referred to as the final iteration of GCN. There is nothing indicating that Navi will be a major improvement over Vega, but rather a more efficient successor of Vega.

There will be a version of Navi for both PS5 and XBox 2 but it's not going to be gcn based. Last I heard Navi is a New GPU Design.

 

[efikkan]

There will be a version of Navi for both PS5 and XBox 2 but it's not going to be gcn based. Last I heard Navi is a New GPU Design.

I wish that were true, but I haven't seen solid evidence to support that.
Navi is not what AMD refers to as "next gen" though.

 

[medi01]

I think the delay is because of 7nm issues moreso than Navi itself. The fact there's limited availability of Radeon VII cards also hints at 7nm issues

But have there been shortages of MI150? VII uses helluva expensive HBM2 and that might be the major reason of the relavitvely low availability.
Besides, as far as rumors go, VII problem was close to MSRP manufacturing price, not production issues, they were ramping it up, fearing Turing, then saw Turing was RT lol for brainwashed.

Still, MI150 chip, having size close to what Navi is expected to be, plus Navi first tested for perf in Oct 2018, coming only in Oct 2019 looks weird. How come you could successfully port Vega, but need a year to fix chip of similar size? Especially given AMD's extensive experience in embracing new nodes.

r competition with the 2080 Ti

Why would 980Ti ($600-ish card, eh?) care about competition in 1k+ card area?
For starters, Navi is expected to top 1080, max 1080Ti (unlikely), but be small and cheap to produce.

NVidia is opening bottles right now.

Yeah. That 20% revenue drop must have been very inspiring.

They are a very, very small company compared to Intel and they are small compared to Nvidia as well.

They are small compared to nVidia only if you compare market capitalization, but not revenue.
nvidia stock price is irrational, given what Intel costs, much higher revenue, same margins.

 

[64K]

They are small compared to nVidia only if you compare market capitalization, but not revenue.
nvidia stock price is irrational, given what Intel costs, much higher revenue, same margins.

From NASDAQ I posted in #21 on this thread:

2017 Annual Financial Reports

Intel 71 billion dollars revenue and 21 billion dollars profit
Nvidia 9.7 billion dollars revenue and 3 billion dollars profit
AMD 5.3 billion dollars revenue and 43 million dollars profit (this includes both CPU and GPU businesses)

 

[eidairaman1]

VII Card should of came out during the 1080Ti Era. Too late.

AMD unfortunately is 1 step behind the 2080/Ti.

Navi needs to be something considerable, perhaps not component limited like the 480/580 were in compare to the 290-390X...

 

[TheoneandonlyMrK]

Adding tensor cores shouldn't be a problem. It's a separate module. Even if Navi wasn't MCM, this could literally be put in a separate chip (granted the interconnector is fast enough).
Nvidia sells a tensor accelerator and it can work in tandem with a traditional GPGPU device.

Also, how do you know they haven't started earlier? They most likely knew a lot about Nvidia's RTRT hardware all along. They may have been surprised by performance, so their solution needed more work to be competitive.

By principle you can run RTRT (DXR) even on a x86 CPU. Actually CPUs are quite competitive in this (ray tracing is a painfully sequential mathematical problem).
Using normal GPGPU you may be able to run a 500x500px 30fps render (the "preview" window in 3D software). Purpose built hardware is orders of magnitude faster.

Your getting confused clearly, Amd said ,one game is not enough dev support yet for Amd to bother.
And Amd have Rspid packed math for quadratic equations and such , different approach ,lower cost , Intrinsically (sic)clever design imho.

Sticking purpose built maths hardware on the side of an asic you Were selling For advanced maths use to prosumers is strange to me.

And you're so wrong on the difficulty of changing a asic design That much when you're virtually At validation, naive bull.

 

[Casecutter]

Just me, it sounds like AMD has to allocate 7nm wafer starts to the markets and products that give them best bang-for-buck.

They want to keep filling the severer, HPC, deep learning, cloud computing with EPYC (Gen2), Ryzen (Gen3) along with huge volumes of profitable Instinct compute cards.

I mean wouldn't we think there's many clients (IBM/Apple) that want to increase or start their production, and bet TSMC is fully booked... didn't I read TSMC is build another 7nm production facility?

 

[Aerpoweron]

Hi everyone.

Some people don't seem to be very happy with the delay and even the Radeon 7. Which i can understand, given all the arguments here.
But i think we live in a time, where you need to know what you want to do and buy the CPUs or GPUs which suits you best.

If you need AVX instructions a lot then go intel. If you need a lot of cores for multitasking go AMD. And so on... It really depends on your requirements.

AMD now has the professional market for the high paying customers. Navi will be a midrange card from what you can expect. And that is where you can earn your money.
In this forum we are most likely enthusiats so we want the fastest and best. But sadly this only makes a small part of the GPU/CPU market.

I really hope AMD brings something good to the table sooner or later concerning the GPUs. Because Nvidia needs competition there. Personally i think Nvidia is good for Gaming. But they have fooled me long enough. PhysiX was dead as long as it was properitary. And Nvidia stated at one point that it won't work (any longer) when you have an AMD GPU as primary and Nvidia as PhysiX accelerator, when it was working fine a month before. Then you can remember the GTX 970 memory issues. Just pick you year and you will find one or two things Nvidia has fooled with us. Sadly AMD is tending the same way, and has learned that from Nvidia, which i really don't like.

The RTX cards are well priced in my opinion. With all the specialized instruction units they have, they are quite huge and complex chips. Raytracing / DXR at least has an official API, so we can expect that it won't be dead soon. Raytracing is a nice feature so far, and very good for people who want to quick render a scene for a Movie. But in games? High performance impact. And moving/distorted water can't be done well right now.

AMD likes to adopt a new fabrication process as seen in earlier cards. But it always comes with a risk. Nvidia usually goes the save way and optimizes more on the current process.
With Navi we will see how well the 7nm works. And we will see if AMD will introduce special Raytracing units. The question will be how much Raytracing hardware you built in, since it seems to be unusable for other workloads.