News AMD Lists Radeon RX 7950 XTX, Other Unreleased Graphics Cards

Secondly, the pull request includes Radeon RX 7800 XT based on the Navi 31 GPU, but there is no sign of a vanilla Radeon RX 7800.

Make some corrections in your article. RX 7800XT is based on NAVI 32, and the remaining mainstream cards will sport NAVI 33 GPU core.

The AMD RDNA 3 GPU portfolio actually is based on three main chips, the Navi 31, Navi 32, and Navi 33.

The Navi 31 GPU is the "GFX1100" and codenamed Plum Bonito, the Navi 32 GPUs are mentioned under "GFX1101" and codenamed Wheat Nas, while the Navi 33 GPUs are designated as "GFX1102" and codenamed Hotpink Bonefish. What funny names by the way.

There's also a GFX1103 design which is aimed at integrated IPs such as APUs, unless I'm mistaken.
  • Radeon RX 7950 XTX RDNA3 gfx1100 (Navi 31)
  • Radeon RX 7950 XT RDNA3 gfx1100 (Navi 31)
  • Radeon RX 7900 XT| RDNA3 gfx1100 (Navi 31)
  • Radeon RX 7900 XT RDNA3 gfx1100 (Navi 31)
  • Radeon RX 7800 XT RDNA3 gfx1101 (Navi 32)
  • Radeon RX 7700 XT RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7600 XT RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7500 XT RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7600M XT RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7600M RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7700S RDNA3 gfx1102 (Navi 33)
  • Radeon RX 7600S RDNA3 gfx1102 (Navi 33)
It's odd that only one AMD Navi 32 GPU exists but it is also likely that AMD will position its Navi 32 chips as the flagship mobility offerings, which might explain its limited nature on the desktop front.

Non-XT models are also likely to appear sooner or later.
 
Last edited:

Eximo

Titan
Ambassador
Non-XT models are also likely to appear sooner or later.
Been thinking about that. They have easy chances for name confusion if they don't give the XT branding to all cards.

As you can imagine, marketing a computer with a Ryzen 5 7600 and an RX 7600 would be perhaps a little misleading. Not that adding XT makes it much better, but at least it doesn't appear in the CPU names (And that is precisely when AMD will launch a R9-7950XT to make this wrong)
 

ET3D

Distinguished
Mar 12, 2012
99
33
18,660
Make some corrections in your article. RX 7800XT is based on NAVI 32, and the remaining mainstream cards will sport NAVI 33 GPU core.
Thanks. Wanted to post this, but you already did.

However, this lineup doesn't make sense to me, at least with the 7700 XT. In most likelihood, based on rumours and current RDNA 3 cards, Navi 33 will be slower than Navi 22. Therefore having a 7700 XT based on Navi 33 doesn't make sense. It also means that Navi 32 won't have a lower end SKU, which is problematic for yields.

My thought is that it's either a typo or, as someone speculated on the Reddit thread, a placeholder for the real list.
 

hannibal

Distinguished
Thanks. Wanted to post this, but you already did.

However, this lineup doesn't make sense to me, at least with the 7700 XT. In most likelihood, based on rumours and current RDNA 3 cards, Navi 33 will be slower than Navi 22. Therefore having a 7700 XT based on Navi 33 doesn't make sense. It also means that Navi 32 won't have a lower end SKU, which is problematic for yields.

My thought is that it's either a typo or, as someone speculated on the Reddit thread, a placeholder for the real list.

I agree with this. It would be strange if there would be only one GPU using navi32 (7800 XT)
Unless navi32 is found out to be useless in which case it could be possible that they will newer put 7800XT to the market and just move 7700xt to tier below and use the best navi33.
But if navi 32 come to the market it does need cut down option.... And it has to be 7700xt or 7800 non xt.
 
  • Like
Reactions: artk2219

hannibal

Distinguished
I could see it like this.

7500=6gb
7600=8gb
7700=10gb
7800=12gb probably 16 though

And on up the stack. Maybe trying to compete with the 4090?

Seems like they should put some effort into professional cards though.

It depends on how they cut the die... But if 7700 use navi 33, the max is 8Gb...
If it is cut down navi32, then who knows.

Do we have any info about navi32. Memory bus?
I am very doubtful about anything considering navi32 at this moment... They could drop it out completely in this generation.
 
  • Like
Reactions: artk2219
However, this lineup doesn't make sense to me, at least with the 7700 XT.

Actually as per some AMD internal documents/technical papers, they listed the 7700 SKU under Navi 33, because it is based on RDNA3 gfx1102 ID. And gfx1102 is basically Navi 33, but yeah, it doesn't make much sense to include the 7700 SKU under the lowest tier GPU core.

I thought about this before, but it seems that AMD will use a heavily cut-down NAVI 32 variant for the 7700 SKU.

BTW, AMD GFX1103 RDNA 3 APU was also once spotted in a Linux patch with DCN 3.1.4 engine, I think last year. Here we get to see another string ID labeled as gfx1103. So maybe they might also introduce gfx1103 SKUs in future, if these entries are valid and anything to go by.

But like I mentioned before, GFX1103 design will be aimed at integrated IPs such as APUs. So no desktop variant thus far.

2022-07-10_6-07-48.png
 
  • Like
Reactions: gg83 and artk2219
Thanks. Wanted to post this, but you already did.

However, this lineup doesn't make sense to me, at least with the 7700 XT. In most likelihood, based on rumours and current RDNA 3 cards, Navi 33 will be slower than Navi 22. Therefore having a 7700 XT based on Navi 33 doesn't make sense. It also means that Navi 32 won't have a lower end SKU, which is problematic for yields.

My thought is that it's either a typo or, as someone speculated on the Reddit thread, a placeholder for the real list.
Confidence is low at this point, but I'd love to see a compelling 7700 XT to replace my 5700 XT.
This is a response to you both.

The problem is that AMD did something really slimy with the RX 7000-series. The RX 7900 XTX should have been named the RX 7900 XT and the RX 7900 XT should have been named the RX 7800 XT. The tech press was so up in arms about nVidia trying to sell two RTX 4080 cards with different performance levels (and rightfully so) that they didn't realise the crap that AMD was trying to pull.

However, AMD really screwed themselves because now they can't properly deliver the correct performance uplift on the rest of their product stack. The 7800 will really be a 7700, the 7700 will really be a 7600 and the 7600 (that they're releasing next) is really a 7500. AMD saw what nVidia was doing and decided to get in on it instead of ignoring nVidia and just doing their own thing like they should have.

Now before anyone thinks otherwise, I'm not blaming nVidia for AMD's actions, like, at all. This was a choice that AMD made and it was a terrible one. By the time anyone realised what AMD had done, it was too late for them to be forced to pull the same 180 that nVidia did when they re-named their RTX 4080 12GB as the RTX 4070 Ti.

As a result, every single card that AMD releases in the RX 7000-series is going to be a grave disappointment because the uplift is going to be barely worth talking about. I'm actually taking some pleasure in this because AMD is now adrift in the video card space as they can't hope to compete with their own 6000-series which are all fantastic values ATM.

The RX 7600 / 7600 XT are rumoured to be in the same performance ballpark as the RX 6700 /6700 XT but with only 8GB of VRAM. With the RX 6700 being currently available for $280 and the RX 6700 XT being currently available for $320, AMD has no choice but to offer the RX 7600 for no more than $200 and the XT for no more than $250.

They have no way out of this situation because in the current environment, people will happily pay a bit extra to get more than 8GB of VRAM. If they try to charge more, the cards will get terrible reviews and nobody will buy them until they get down to those price points anyway. That means all that AMD will be doing is screwing themselves because they won't get more money than $200 and $250 for these cards anyway.

Meanwhile, the tech press will, quite rightly, tell people to stay away from these cards and so sales of the card will be virtually non-existent, just like the RTX 4070. OTOH, if they do price them as I said, they'll get more money because the tech press will hail them as great buys which means that more people will buy them.

They're squirming right now and this can be seen by the fact that the sole AIBs that are releasing 7600-series cards are the AMD-Only brands like ASRock, Powercolor, Sapphire, and XFX. Other AIBs like ASUS, Gigabyte and MSi have opted out for now, probably because they see the clusterfrack coming and they know that whatever profit they receive from these cards won't be worth the effort to make them.

AMD is squirming because of their own jackassery and I am absolutely loving it. They had the opportunity to avoid this situation but they decided against it. The greed and cynicism involved in their decisions regarding card nomenclature is unlike anything that I'd ever seen from AMD before. They deserve all the suffering that they reap and I for one am going to laugh my butt off.

Hopefully, they learn a lesson from this because after the fantastic launches of Zen3 and RDNA2, this kind of conduct is inexcusable.
 
Last edited:
  • Like
Reactions: sherhi and Jagar123

bit_user

Polypheme
Ambassador
The Navi 31 GPU is the "GFX1100" and codenamed Plum Bonito, the Navi 32 GPUs are mentioned under "GFX1101" and codenamed Wheat Nas, while the Navi 33 GPUs are designated as "GFX1102" and codenamed Hotpink Bonefish. What funny names by the way.
I think AMD's naming scheme is to pair a color with a type of fish. The only one I'm not sure about is Nas, which seems like it could refer to Nonindigenous Aquatic Species.

Regarding refreshes, my fantasy is that AMD has found some bugs in Navi 31 that's keeping it from delivering the expected performance. That and more L3 cache are the only things I can imagine they'd do to make a 7950 XT(X). If you look at the specs of Navi 31, it seems like it should perform way better than it does. So, I'm hoping that's due to bugs/workarounds that they'll fix.
 
Last edited:
  • Like
Reactions: Metal Messiah.

Gillerer

Distinguished
Sep 23, 2013
361
81
18,940
There's no sign of a vanilla Radeon RX 7800 as yet

It seems to me that now that Ryzen and Radeon share the same thousands numbering, AMD is differentiating the product names by having different suffixes: vanilla/X/X3D on CPUs and XT/XTX on GPUs.
 
I think AMD's naming scheme is to pair a color with a type of fish. The only one I'm not sure about is Nas, which seems like it could refer to Nonindigenous Aquatic Species.

Regarding refreshes, my fantasy is that AMD has found some bugs in Navi 31 that's keeping it from delivering the expected performance. That and more L3 cache are the only things I can imagine they'd do to make a 7950 XT(X). If you look at the specs of Navi 31, it seems like it should perform way better than it does. So, I'm hoping that's due to bugs/workarounds that they'll fix.

Yeah, I know those names are taken aquatic species, some of which are mostly rare, but can be found mostly everywhere. Both AMD and Nvidia have been using odd codenames for their CPU/GPU architecture, names taken from Bridges, Artists/Painters, Scientists, Solar system/galaxy/stars just a name a few.

What's next ? I think they will use fruits and vegetables as their codename, LOL, or maybe iconic places/temples.

Anyway, on a serious note, it's highly possible that the high-end Radeon RX 7950 series refresh will reuse the Navi 31 core architecture, while the Radeon RX 7800 XT will be the only Navi 32 chip on the block.

Though, since Navi 31 already has it's peak core count on the die, so the only other changes we can expect from a possible refresh will be higher clocks, faster memory and more power/TDP, as we saw with the RX 6950 XT. And like you said, also increased L3 cache.

Btw, do you think the 7950 XTX/XT might dethrone RTX 4090 as well when it comes to raw gaming performance ?
 
Last edited:
  • Like
Reactions: bit_user
Btw, oddly enough AMD once claimed they are not interested in making flagship GPUs which can compete at least with the likes of RTX 4090.

During an interview with ITMedia, AMD's EVP, Rick Bergman, and AMD SVP, David Wang explained why AMD didn't release an RDNA 3 GPU under its Radeon RX 7000 lineup that competes in the ultra high-end enthusiast segment such as NVIDIA's RTX 4090.

To quote AMD EVP, Rick Bergman (Machine Translated via ITMedia)

Technically, it is possible to develop a GPU with specs that compete with theirs (NVIDIA) . However, the GPU developed in this way was introduced to the market as a graphics card with a TDP (thermal design power) of 600W and a reference price of $1,600 (about 219,000 yen)'', and was accepted by general PC gaming fans . After thinking about it, we chose not to adopt such a strategy.

The RDNA 3-based GPU "Radeon RX 7900XTX" released this time is targeted at $ 999 (about 136,000 yen), which is considered to be the "upper price" assumed by high-end users among general PC gaming fans. Developed . The "Radeon RX 7900XT" below it is said to be $ 699 (about 95,000 yen).

The price strategy is the same as the previous RDNA 2 (Radeon RX 6000 series), with the top-end "Radeon RX 6900XT" and "Radeon RX 6800XT" targeting $999 and $699, respectively. However, the target price changes for each GPU generation .

We take this strategy to fit into the mainstream infrastructure (hardware environment) utilized by today's PC gaming enthusiasts . At the same time as demanding high performance, it should be possible to operate with an existing common sense'' power supply unit, ``be able to cool the inside of the case with common sense'', and can be installed without requiring an extremely large case.'' ――The Radeon RX high-end product group was designed with these in mind.
 
  • Like
Reactions: bit_user
120-G2-1000-XR
What's next ? I think they will used fruits and vegetables as their codename, LOL, or maybe iconic places/temples.
Wouldn't it be great if the bottom entry-level GPU was called "Potato"?
:ROFLMAO::ROFLMAO::ROFLMAO:
Anyway, on a serious note, it's highly possible that the high-end Radeon RX 7950 series refresh will reuse the Navi 31 core architecture, while the Radeon RX 7800 XT will be the only Navi 32 chip on the block.
And that's where AMD had completely messed up with their BS naming scheme having TWO cards called "RX 7900". The way it has always worked in the past is that the level-8 card ALWAYS had a version of the top-tier GPU, if not the full version. Now, the RX 7800 XT will really be just a RX 7700 XT with the wrong name and thus the price will be stupid and the performance will be disappointing for a level-8 card because it'll really be a level-7.
Though, since Navi 31 already has it's peak core count on the die, so the only other changes we can expect from a possible refresh will be higher clocks, faster memory and more power/TDP, as we saw with the RX 6950 XT. And like you said, also increased L3 cache.
Unless they have a die in the wings that nobody knows about (which is so unlikely as to be not worth thinking about).
Btw, do you think the 7950 XTX/XT might dethrone RTX 4090 as well when it comes to raw gaming performance ?
I seriously doubt it but even if it does, nVidia always keeps an ace (or two) up its sleeve. That ace would either be called the RTX 4090 Ti or something with the word "Titan" in it.
 
Last edited:

bit_user

Polypheme
Ambassador
Btw, do you think the 7950 XTX/XT might dethrone RTX 4090 as well when it comes to raw gaming performance ?
Well, the RX 7900 XTX has paper specs of 46.7 (61.4) fp32 TFLOPS and 960 GB/s of memory bandwidth. If we compare that to the RTX 4090, it brings to the party 73.1 (82.6) fp32 TFLOPS and 1008 GB/s of bandwidth. So, there's a very real gap, on paper.

The reason RDNA2 performed so well against RTX 3000 is thanks to Infinity cache. However RTX 4000 has now added a comparable amount of L2 cache. So, that levels the playing field, where RTX 4090's superior specs are probably the deciding factor.

So, no. I don't think RX 7950 XTXXX can find enough additional performance to catch the RTX 4090. I just want to see a bigger improvement over the RX 6950 XT (19.4 (23.7) fp32 TFLOPS; 576 GB/s) that's more inline with the delta between their specs. On paper, the current RDNA3 GPUs look like a really massive leap beyond RDNA2. Like, nearly 2x territory.
 
  • Like
Reactions: Metal Messiah.

bit_user

Polypheme
Ambassador
Btw, oddly enough AMD once claimed they are not interested in making flagship GPUs which can compete at least with the likes of RTX 4090.

During an interview with ITMedia, AMD's EVP, Rick Bergman, and AMD SVP, David Wang explained why AMD didn't release an RDNA 3 GPU under its Radeon RX 7000 lineup that competes in the ultra high-end enthusiast segment such as NVIDIA's RTX 4090.

To quote AMD EVP, Rick Bergman (Machine Translated via ITMedia)
Yeah, I saw that. Part of it feels like sour grapes, but I think they probably have been competing with Nvidia for long enough to know that if you want to top them, you really need to bring multiple times more firepower than you expect it will take. Nvidia is one of the most formidable competitors out there, and I'm sure everyone at AMD knows it, by now.
 
  • Like
Reactions: Metal Messiah.