Discussion RDNA4 + CDNA3 Architectures Thread

Page 350 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DisEnchantment

Golden Member
Mar 3, 2017
1,774
6,757
136





With the GFX940 patches in full swing since first week of March, it is looking like MI300 is not far in the distant future!
Usually AMD takes around 3Qs to get the support in LLVM and amdgpu. Lately, since RDNA2 the window they push to add support for new devices is much reduced to prevent leaks.
But looking at the flurry of code in LLVM, it is a lot of commits. Maybe because US Govt is starting to prepare the SW environment for El Capitan (Maybe to avoid slow bring up situation like Frontier for example)

See here for the GFX940 specific commits
Or Phoronix

There is a lot more if you know whom to follow in LLVM review chains (before getting merged to github), but I am not going to link AMD employees.

I am starting to think MI300 will launch around the same time like Hopper probably only a couple of months later!
Although I believe Hopper had problems not having a host CPU capable of doing PCIe 5 in the very near future therefore it might have gotten pushed back a bit until SPR and Genoa arrives later in 2022.
If PVC slips again I believe MI300 could launch before it

This is nuts, MI100/200/300 cadence is impressive.



Previous thread on CDNA2 and RDNA3 here

 
Last edited:

Keller_TT

Member
Jun 2, 2024
143
162
76
Going to say it again, there's no point to chasing halo without a few generations of "consistently decent" before that. You have to give people confidence you're worth a damn before you can convince them you're the best.
Bingo.

I'm primarily concerned about FSR4 image quality in one title:
Assetto Corsa Evo by Kunos Simulazioni. It's in Early Access and has problems with FSR 3. It's outright not usable in anything less than Native AA. DLSS 4 elevates the game's new rendering engine to just ridiculously gorgeous details. DLSS4 Q is >> DLSS3.x DLAA.

AMD was a mess in VR for sim racing. So I had to sell my Radeon and get myself a 4080. It's gotten better but far from ideal, and NV has some super cool tech for VR that benefits overall from DLSS upscaling.

This is a long standing weakness of AMD. It's time to just iron out ALL pain points. My 7950 X3D is a real treat for proper, heavy racing sims. It's terrible if Radeon can't even be a consideration for such a defining use of tech and one that is a real joy to experience.

I'm rn in the mood. This year, I'm spending big on a dedicated sim racing room & set-up to last me minimum 5 years. For disclosure, I hold a ~ $3 Mil position in $AMD, and I'm really wishing AMD doesn't screw it up this time. Will decide a few things after tomorrow.
 

APU_Fusion

Golden Member
Dec 16, 2013
1,691
2,483
136
You launch 5070, you wait for Computex, announce SUPER, ship August. (frankly they can do it faster, depends on how much optics they want to sacrifice).
That's it.
For the rest of the generation, AMD is wholly irrelevant.
Then why are you this thread? Weird flex. Also, good job disregarding a reasonable request not to use rather ugly language when asked. How is that 40% ipc doing?

I was looking to get a 5090 (coming from old 1070 system) but the prices and gaslighting by nvidia has me looking at team read even with its doomed to fail so why release them cards.
 

Keller_TT

Member
Jun 2, 2024
143
162
76
I was looking to get a 5090 (coming from old 1070 system)
Please, no. It's way too much over maximum wattage that's reasonable for home entertainment, and I was happy in a way that it was unobtainable. If a 6080 can deliver between 4090 and 5090 performance for 300W, that's getting somewhere. One more reason I abhor vanity "halo" SKUs for the sake of it as it's just so egregious.

I really wish there is a heavy premium in both GPU pricing and energy prices for domestic cards using >300W+10% leeway.
 

APU_Fusion

Golden Member
Dec 16, 2013
1,691
2,483
136
Please, no. It's way too much over maximum wattage that's reasonable for home entertainment, and I was happy in a way that it was unobtainable. If a 6080 can deliver between 4090 and 5090 performance for 300W, that's getting somewhere. One more reason I abhor vanity "halo" SKUs for the sake of it as it's just so egregious.

I really wish there is a heavy premium in both GPU pricing and energy prices for domestic cards using >300W+10% leeway.
Yes, I couldn’t quite believe the 625 watts for so little gain over the 4090 and the marketing scams with fake frame generation plus lying about missing rops. Rather bad form.

Don’t want to highjack thread. Definitely will consider 9070xt unless the price it way too high
 
Reactions: Tlh97 and Keller_TT

gdansk

Diamond Member
Feb 8, 2011
4,117
6,826
136
Polaris was a fab filler. A result of the 2016 wafer supply agreement. Marginal profit was better than paying GF to make nothing.

AMD has no need for fab filler now. A deliberate low margin, high volume strategy isn't likely to return.
 

GodisanAtheist

Diamond Member
Nov 16, 2006
7,991
9,183
136
NV will launch what they have for whatever price they want. AMD is like a rounding error to them, they're not going to steer the ship around the iceberg when they have a breaker that can just go straight through.
 
Reactions: marees

gdansk

Diamond Member
Feb 8, 2011
4,117
6,826
136
NV will launch what they have for whatever price they want. AMD is like a rounding error to them, they're not going to steer the ship around the iceberg when they have a breaker that can just go straight through.
They can take 1% of their allocation and make MSRPs real again for everything under the 5080. That's all it takes to kill N48.
 

Gideon

Platinum Member
Nov 27, 2007
2,009
4,976
136
They can take 1% of their allocation and make MSRPs real again for everything under the 5080. That's all it takes to kill N48.
Really? essentially the entire desktop GPUs market below the 5080 are only 1% of all Nvidias wafers?

Even when true, I see no reason why Nvidia would want to sacrifice its profits and start a price-war to go after a party having ~10% of the Discrete GPU market. GB203 is pricier than NV48. GDDR7 is a lot pricier (and i bet volume limited still). And I'm sure if they dump their prices too low they can be sued for monopolistic tactics (e.g. using their monopoly in Server GPUs to force out competition in discrete graphics on unsustainable margins).

I don't believe they would actually lower prices proactively. Perhaps when they actually start to lose a measurabel chunk sof marketshare but not proactively.

I really don't get the logic people here seem to apply: "under no circumstance should AMD ever price anything below Nvidia, ever!" or dear god, Nvidia might lower prices a bit ... in a while ... or with a super refresh a year from now ..

Make sure it's only a an arbitrary fixed $50 from Nvidia (no matter whether that $50 is 7% or 20% of the price mind you) or AMD is immediately doomed. Doomed an bankrupt!
 

Josh128

Senior member
Oct 14, 2022
844
1,314
106
They can take 1% of their allocation and make MSRPs real again for everything under the 5080. That's all it takes to kill N48.
How does that kill N48 if its as fast as 5070Ti and $50-$100 cheaper?? The only thing that would kill RDNA 4 is if NV price their equivalent SKUs cheaper than N48. GB 203 is, to public knowledge thus far, slightly larger than N48 and uses an equal amount of more expensive memory. So the only way it can equal pricing on N48 is if they are happy to take on a profit margin even lower than AMD will for N48. Have you seen NV's stock price movement today? Huang is not going to do anything that will require reducing margins and thus further erode shareholder confidence, especially at this low end, guaranteed.

Contrary to what you posted above, it certainly would sacrifice their profits. If AIBs margins tank, they simply wont produce those SKUs and buy those chips from NV anymore-- NV loses sales and profit.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |