AMD: Navi Speculation, Rumours and Discussion [2019-2020]

[BoMbY]

66AF is Radeon VII.

 

[Deleted member 91098]

66AF is Radeon VII.

Thanks, now that you mention it I remember that there was a post here on B3D because of that, since 66AF was only listed with 20 CUs on gfxbench and as gfx1010 which was claimed to be Navi: https://compubench.com/device.jsp?benchmark=compu15d&os=Windows&api=cl&D=AMD+66AF:F1&testgroup=info

EDIT: found it: https://forum.beyond3d.com/posts/2061655/

 

[BoMbY]

GFXbench and CompuBench are mixing up results sometimes. At least I'm very confident about the ID:

 

[Cat Merc]

Thanks for posting. Interesting information. Nvidias GTX 1080 FE with it's 9.1 TFLOPs (usual boost is 1866 MHz so actually ~9.6 TFLOPs) beats Vega 64 in games. That would mean AMDs TFLOPs in games are better than Nvidias now, which would be nice for a change.

EDIT: I remember that there was a leak from gfxbench (or something) that suggested turned down compute capabilities but much increased game performance for Navi. Does someone remember it? I can't find the bench result but I believe it was with 20 or 24 CUs while performaning around RX 580 levels.

Note that Turing is better than Pascal in that regard. Navi looks to be at parity with Turing, above Pascal.

 

[Jawed]

"Game clock" seems like a new thing. Boost clock is higher.

I wonder if boost clock is for compute only?...

 

[Cat Merc]

"Game clock" seems like a new thing. Boost clock is higher.

I wonder if boost clock is for compute only?...

Boost clock is probably peak clock, while game clock is typical clockspeed.

 

[anexanhume]

Of course it is based on an average, so that with 1750 Mhz it will still be on average 11% short in performance of a Vega 64.

Sorry, I have no idea what you're getting at. All the numbers are consistent with the claims.

 

[Love_In_Rio]

Sorry, I have no idea what you're getting at. All the numbers are consistent with the claims.

To sum it up, RX 5700XT will be slower than Vega 64.

 

[Kaotik]

Alright, got some info about Navi, don't ask about the source, but it's reliable as hell, and I trust it implicitly.

The highest SKU launching will be named RX 5700 XT, 40CU, 9.5TFLOPS, 1900MHz max clocks, with 1750MHz being the typical gaming clock. Power delivery is through 2X 6pin connectors.

Clocks are slightly wrong though

 

[w0lfram]

To sum it up, RX 5700XT will be slower than Vega 64.

How/why do you keep saying/suggesting this..?

Dr Su said otherwise about Navi's performance over Vega 14nm. So how are you surmising that little Navi has less gaming performance than V64?

 

[DieH@rd]

Game Clock will be higher on AIB cards. Navi will launch only with blower cards. :-/

 

[Deleted member 13524]

Of course it is based on an average, so that with 1750 Mhz it will still be on average 11% short in performance of a Vega 64.

11% lower FP32 throughput. Gaming performance should be equal or higher.


The screenshot that matters:

Also seems like a confirmation that the reference cooler is a blower.
Shouldn't be too bad if power consumption under load is 180W.

Let's hope they'll have 3rd parties with open air coolers.

 

[BRiT]

So that's the 5700 XT, there's going to be a 5800 XT and a 5900 XT, right?

 

[Kaotik]

So that's the 5700 XT, there's going to be a 5800 XT and a 5900 XT, right?

Unlikely.

 

[Deleted member 13524]

So that's the 5700 XT, there's going to be a 5800 XT and a 5900 XT, right?

Rumors say there'll be a 5700 Pro. I'm guessing AMD is going back to the R600 era with that naming scheme.

Not sure why though. I just wish they'd settle for one naming scheme and follow through with it.

I have no idea what else to expect from here onwards. They make no sense.

 

[snarfbot]

Unlikely.

How about 5790?

 

[Bondrewd]

Unlikely.

Very, very likely.
Wait ~2Q.

 

[Rootax]

But, will it run Crys... I mean Quake II RTX ?

 

[Bondrewd]

But, will it run Crys... I mean Quake II RTX ?

Wait like a few days.

 

[Deleted member 13524]

But, will it run Crys... I mean Quake II RTX ?

I doubt, but it still might.

https://www.guru3d.com/news-story/a...bout-their-raytracing-strategy-during-e3.html

I'm not going to steal Lisa's thunder, but we'll share that with you. PCIe 4, GDDR6, other architectural features that we want to bring to that architecture that wants to be sustainable across multiple generations and we'll be good. And then in time obviously, we need to start talking about what our ray tracing strategy is moving forward and some other elements. So we look forward to E3 and we hope everybody wants to tune in on Monday.