AMD Announces RX 5000 Series Graphics processors at Computex - Demos RX 5700

Pieter3dnow

Supreme [H]ardness
Joined
Jul 29, 2009
Messages
6,784
https://www.guru3d.com/news-story/a...ics-processors-at-computex-demos-rx-5700.html

As we all know NAVI is the new family of Radeon GPUs. NAVI is based on 7nm RDNA architecture (Radeon DNA). NAVI will be PCIe Express 4.0 compatible, has a new compute design, new cache-hierarchy and will be reducing latency, lower power. 1.25x perf per clock improvement and 1.5x better power consumption.

AMD considers Navi as a completely new architecture. No longer GCN, but RDNA. RDNA uses completely redesigned Compute Units, which should significantly increase per-clock performance and efficiency. This includes a new cache structure. The changes are expected to boost Vega's performance by at least 25 percent per cycle ("IPC") and increase efficiency (wattage) by at least 50 percent.
 
Slower than a 2+ year old 1080 Ti and will cost more. Performance between a 2060 and 2070 in most games.

Dammit AMD I wanted to believe. At least they stopped shooting themselves in the foot with the HBM nonsense and went GDDR6.
 
Last edited:
Slower than a 2+ year old 1080 Ti and will cost more. Performance between a 2060 and 2070 in most games.

Dammit AMD I wanted to believe. At least they stopped shooting themselves in the foot with the HBM nonsense and went GDDR6.
Where'd you get your pricing info? Plus, Nvidia just released several video chips that are slower than the 2+ year old 1080 Ti, so I don't see how that is relevant. AMD already has a GPU that competes on price and performance with the RTX 2080, so this "it's not as fast as the (insert however many years old) (insert some preferred NVIDIA flagship GPU)" argument is getting old.

The GPU used in the demonstration "appears" to be faster than the 2070. If it is priced at the same point or a little lower, I don't see why anyone should complain. If it's priced higher, then, yes, we should certainly pile on AMD for that screw up.
 
Slower than a 2+ year old 1080 Ti and will cost more. Performance between a 2060 and 2070 in most games.

Dammit AMD I wanted to believe. At least they stopped shooting themselves in the foot with the HBM nonsense and went GDDR6.

this was likely an es and it preformed right were it needs to 2070 for half the power and a smaller chip and pcie 4.0 amd wins
 
Slower than a 2+ year old 1080 Ti and will cost more. Performance between a 2060 and 2070 in most games.

Dammit AMD I wanted to believe. At least they stopped shooting themselves in the foot with the HBM nonsense and went GDDR6.

You mean you still did not get the memo that AMD poured all of their R&D resources into Zen and did not have the budget for graphics cards.

Maybe you know some people that can make great gpu out of thin air but in the real world they cost money and the relation to performance is there...
 
It's not nonsense. Fabrication of stacked chip with TSVs is difficult and expensive but has less latency and can pack more in less space. HBM is still in it's infancy and will be around on package for a long time.
And as soon as AMD can get better solutions working for graphics something as an APU with HBM on the IO die would be something special ;)
 
It's not nonsense. Fabrication of stacked chip with TSVs is difficult and expensive but has less latency and can pack more in less space. HBM is still in it's infancy and will be around on package for a long time.
None of that changes what I said. Like putting monster truck tires on a Ferrari, HBM was the wrong fit for AMDs GPUs. Nobody gives a shit about a superior component when it doesn't translate into superior performance for the intended application.
 
None of that changes what I said. Like putting monster truck tires on a Ferrari, HBM was the wrong fit for AMDs GPUs. Nobody gives a shit about a superior component when it doesn't translate into superior performance for the intended application.

Not really without HBM Vega would never have been viable as a consumer product due to GDDR5 requiring to much power.
There was some superior performance when you enabled HDR the hit from the bandwidth needed was very little compared with Nvidia ;) .
 
You mean you still did not get the memo that AMD poured all of their R&D resources into Zen and did not have the budget for graphics cards.

Maybe you know some people that can make great gpu out of thin air but in the real world they cost money and the relation to performance is there...
I don't care what the excuses are, we're dealing in what is. For selfish reasons I would've preferred more emphasis on GPU development because CPUs have blown past the point of "good enough" years ago. Whoopdy do, 15% faster than what was already overkill, with 4 more cores than was already overkill -- let's shoot our guns in the air and fire up our video transcoders just to have something to do.

"Wait for Navi, wait for Navi" is the cry we've been hearing for eons. And its a mousefart, right on cue.

Wait until Intel rolls into town swinging it's dGPU dick around. AMD will end up just a pure CPU company at that point. Can't blame them though and I understand the business reasons.
 
Last edited:
I don't care what the excuses are, we're dealing in what is. For selfish reasons I would've preferred more emphasis on GPU development because CPUs have blown past the point of "good enough" years ago. Whoopdy do, 15% faster than what was already overkill, with 4 more cores than was already overkill -- let's shoot our guns in the air and fire up our video transcoders just to have something to do.

"Wait for Navi, wait for Navi" is what we've been hearing for eons, and its a mousefart, right on cue.

Wait until Intel rolls into town with dGPUs swinging it's dick around, clearing the chessboard of the GPU landscape with it's swinging dick. AMD will end up just a pure CPU company at that point. Can't blame them though and I understand the business reasons.
Actually the thing we been waiting for is after Navi ;). New people in RTG new leaders for that gpu team and a design that supposedly finally fixes problems as scaling ....
But I got to say that if you do not buy Intel cpu there might be some spare cash for the gpu :) .
 
  • Like
Reactions: N4CR
like this
It's not nonsense. Fabrication of stacked chip with TSVs is difficult and expensive but has less latency and can pack more in less space. HBM is still in it's infancy and will be around on package for a long time.
HBM is also really great in systems applications where real time processing is needed. Even if it doesn’t take off in GPU’s it will have a good life outside it for real time image recognition and processing.
 
So little information was given.

If AMD going to do a full unveiling at E3, is the release date even later than that?
 
Navi isn't a huge win and AMD knows it. It will be similar performance to Vega 64, for less money and less power. They know this, that's why the vast majority of the presentation was about Zen2, where they know they have created a truly competitive offering.
 
Would have been hilarious if after she announced the price of the 3900x she took her headset off and dropped it as she exited the stage.
 
Navi isn't a huge win and AMD knows it. It will be similar performance to Vega 64, for less money and less power. They know this, that's why the vast majority of the presentation was about Zen2, where they know they have created a truly competitive offering.


I know it's not a homerun like everyone wanted, but it's still a step in the right direction, right?

If it's similar to 2070 at similar or less pricing, at least they are somewhat competitive.
 
Obviously it will depend on price and how performance turns out but the whole Navi thing, at least for desktops, just left me underwhelmed. I think the real interesting bit for Navi is going to be what it can do in the upcoming Sony and, presumably, MS consoles. With both Zen 2 and Navi powering next generation consoles we might actually see Sony and MS finally put out something with solid CPU and GPU performance.
 
I know it's not a homerun like everyone wanted, but it's still a step in the right direction, right?

If it's similar to 2070 at similar or less pricing, at least they are somewhat competitive.

Yes, it should be better but it isn't. This could be teething a new architecture on a new process node that could improve over time. But only time will tell. I have high hopes for the future. AMD has been silent for entirely too long and Intel and Nvidia are going to take notice.
 
Yes, it should be better but it isn't. This could be teething a new architecture on a new process node that could improve over time. But only time will tell. I have high hopes for the future. AMD has been silent for entirely too long and Intel and Nvidia are going to take notice.

In what way? What is supossedly wrong with it?
 
Wow, arm chair engineer to the rescue!

So the memory that forced Fury to be limited to 4GB, caused Fury and Vega to be way too expensive, and in all but a handful of situations did fuck all for consumer and even prosumer applications was the right choice?
 
So the memory that forced Fury to be limited to 4GB, caused Fury and Vega to be way too expensive, and in all but a handful of situations did fuck all for consumer and even prosumer applications was the right choice?

You are wrong but that is fine, being all emotional about it does not make you correct.

Edit: Or even more to the point, you were going to be one of these guys on the forefront bemoaning the power consumption and size of the cards because.........
 
July 7 for everything to hit the shelves.
That is for the cpu , Navi has not been announced other then July.
So the memory that forced Fury to be limited to 4GB, caused Fury and Vega to be way too expensive, and in all but a handful of situations did fuck all for consumer and even prosumer applications was the right choice?

HBM was the only solution all of the other memory would require a great deal of extra power to where it was no longer anything useful for consumers...
 
You are wrong but that is fine, being all emotional about it does not make you correct.

Edit: Or even more to the point, you were going to be one of these guys on the forefront bemoaning the power consumption and size of the cards because.........

You know what they say about making assumptions, right? I don't complain about power consumption. My biggest concern with hardware is one thing: Performance. If one card uses 300w and the other uses 250w but the 300w one has better performance than that is the one I'd want.https://hardforum.com/forums/hardforum-tech-news.116/
 
That is for the cpu , Navi has not been announced other then July.


HBM was the only solution all of the other memory would require a great deal of extra power to where it was no longer anything useful for consumers...

So which one makes more sense? The cheaper option that keeps it in parity or the more expensive option that lowers it's available memory and performance?

I'll tell you what, people may complain about power consumption here, but if the Radeon card was faster people would buy it anyway.. I mean, look at how many changed back to AMD CPUs even after bulldozer... They had a mega flop for AGES and then release a competitive cpu and it blew up.

The bigger issue for AMD GPUs was that they took more power AND were slower. That right there doesn't sell cards. Had they had the speed but used more, they would have sold.
 
So which one makes more sense? The cheaper option that keeps it in parity or the more expensive option that lowers it's available memory and performance?

I'll tell you what, people may complain about power consumption here, but if the Radeon card was faster people would buy it anyway.. I mean, look at how many changed back to AMD CPUs even after bulldozer... They had a mega flop for AGES and then release a competitive cpu and it blew up.

The bigger issue for AMD GPUs was that they took more power AND were slower. That right there doesn't sell cards. Had they had the speed but used more, they would have sold.

You mean that something uses near or over 500 Watt is something people would happily use in their computer case, because that is what Vega would have used(GDDR5) ?


RX 570 is a good example cheaper faster and yet does not outsell Nvidia counterpart.
 
In what way? What is supossedly wrong with it?

AMD knows it needs to meet or exceed it's competition. That's why they are so elated with Zen2. They know they haven't done this with Navi. With time, (as I stated already) the new architecture could produce a competitive product.

When their newest architecture can barely beat Vega64 (which I have so don't call me an Nvidia shill) it's not a *win*.
 
AMD knows it needs to meet or exceed it's competition. That's why they are so elated with Zen2. They know they haven't done this with Navi. With time, (as I stated already) the new architecture could produce a competitive product.

When their newest architecture can barely beat Vega64 (which I have so don't call me an Nvidia shill) it's not a *win*.

It offers a 1.25x IPC uplift. Which is beating Vega by a LOT. Between architecture and process it has 1.5x perf/watt uplift, which is again beating Vega by a LOT.
 
Slower than a 2+ year old 1080 Ti and will cost more. Performance between a 2060 and 2070 in most games.

Dammit AMD I wanted to believe. At least they stopped shooting themselves in the foot with the HBM nonsense and went GDDR6.

And they're gonna charge $500 for it. Should have been a $349 part if they want any marketshare. Maybe they like less than 10% of the market? The entire gpu price spectrum is a complete shit show right now. We're paying top dollar for 4 year old tech numbers right now.
 
That is for the cpu , Navi has not been announced other then July.

My bad, thought I had seen it somewhere that everything was coming out 7/7. Doesn't make much sense for them to release on different days imo. If I were them I'd want people benching my 2 new products together rather than benches being released with my cpu in tandem with the competitors gpus. We'll see what happens.
 
Last edited:
AMD knows it needs to meet or exceed it's competition.

In a wider context, they need to make money- that's done with volume and margins. Navi may not be faster, but it does bring the possibility of better margins in the mid-range segment with the possibility of some pricing flexibility.

Their biggest concern would be Nvidia dropping a similarly priced '1680' with ~2070/1080-level performance, or just dropping 2070 prices.
 
Consumers often don't like slimy corporations. Stuff like GPP puts a sour taste in their mouths. Prefer the ethical.

AMD is playing the 'ethical' card because their deck is that thin... to pretend than any corporation is ethical from the perspective of consumers is to live in a fantasy world.

Do note that AMD has lost GPU marketshare since the GPP was exposed.
 
Back
Top