• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion RDNA4 + CDNA3 Architectures Thread

Page 427 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
DisEnchantment

DisEnchantment

Golden Member
1655034287489.png
1655034259690.png

1655034485504.png

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
github.com

History for llvm/lib/Target/AMDGPU - llvm/llvm-project

The LLVM Project is a collection of modular and reusable compiler and toolchain technologies. - History for llvm/lib/Target/AMDGPU - llvm/llvm-project
github.com github.com
Or Phoronix

More AMD "GFX940" Enablement Work Landing In LLVM - Phoronix

www.phoronix.com www.phoronix.com

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 :grimacing:

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

1655034362046.png

Previous thread on CDNA2 and RDNA3 here

Question - Speculation: RDNA3 + CDNA2 Architectures Thread

Man I have been dying to make this one for a while now. First rumours for RDNA3 are here so new thread time! Just going to start off with this one for now: kopite7kimi on Twitter: "@VideoCardz Ah, I mean a simple mcm design with 10240 cores is not enough. Because the lift from RDNA2 to RDNA3...
forums.anandtech.com forums.anandtech.com
 
Last edited:
Geranium said:
OEMs are definitely shilling for Nvidia. There are more Strix Point laptop with dGPU than Strix Point only laptop; And Strix Point is designed for thin and light-weight laptop.

Let me give you a example:
Specification of Picasso based ASUS laptop : Ryzen 5 3500U + 4GiB Single-Channel RAM + MX230 + 1TB HDD. And you read right no SSD in a laptop launched in 2019, but has a Nvidia GPU. No SSD or dual channel memory, but hey it has a nvidia gpu which is not that faster than the iGPU and consumes more power. Nvidia shilling at best.
Click to expand...

I always hated how laptops seem to "require" a dGPU. Especially the ones with strong iGPU's to begind with. Thankfully it doesn't seem to be that way anymore. At least not the last time I looked.
 
Thunder 57 said:
I always hated how laptops seem to "require" a dGPU. Especially the ones with strong iGPU's to begind with. Thankfully it doesn't seem to be that way anymore. At least not the last time I looked.
Click to expand...
Sadly situation is bad with Strix Point when Strix Point has more GPU core than the Rembrant or Phoenix.
 
Racan said:
Click to expand...

i like the focus of that guy (200W). I'd love a Fury Nano like card, efficient, not ridiculously sized, and reasonable priced. I feel like that should be possible, especially with DLSS/FSR now, where it can still perform quite well and would do good running somewhat older games. But I've been sitting out the GPU market for a decade at this point.

I'll be curious if the 5000 series stuff is software/drivers or hardware. Guess we'll maybe see, if software stabilizes things and improves performance, or only with new cards (Super versions probably next year). Or maybe they can iron it out befoare then. If software can do it, then 5000 series should turn out like RDNA1, where early software issues hampers it but then it stablizes and turns out decent-ish. Especially if prices improve (ok, they won't).

If I were Intel, I'd take a gamble and make some large GPUs on their older mature processes, where sure they'll be large but keep them clocked efficiently and it could work in their favor. The combination of older mature process and bypassing tariffs could help them with sales by giving them cost advantages.
 
darkswordsman17 said:
i like the focus of that guy (200W). I'd love a Fury Nano like card, efficient, not ridiculously sized, and reasonable priced. I feel like that should be possible, especially with DLSS/FSR now, where it can still perform quite well and would do good running somewhat older games. But I've been sitting out the GPU market for a decade at this point.
Click to expand...
Yeah, a little over 200W for a graphics card is also my limit, for the same reasons the guy in the video mentioned. But size-wise, I don’t personally see the need for something like the Fury Nano. I prefer the silence that a larger cooler can offer, and there are some very space-efficient small form factor cases out there that can still fit those big cards.
 
Geranium said:
OEMs are definitely shilling for Nvidia. There are more Strix Point laptop with dGPU than Strix Point only laptop; And Strix Point is designed for thin and light-weight laptop.

Let me give you a example:
Specification of Picasso based ASUS laptop : Ryzen 5 3500U + 4GiB Single-Channel RAM + MX230 + 1TB HDD. And you read right no SSD in a laptop launched in 2019, but has a Nvidia GPU. No SSD or dual channel memory, but hey it has a nvidia gpu which is not that faster than the iGPU and consumes more power. Nvidia shilling at best.
Click to expand...

I assume that's part The Green Sticker and part using up (partially) busted IGPs. That's how you end up with those Raptor Lake+4050s which are like 700 bucks.
 
SolidQ said:
High memory temp research from Igor
www.igorslab.de

High memory temperatures on many RDNA4 cards and what exactly is behind them - Cause research and background information | Page 3 | igor´sLAB

As I was also surprised by the high memory temperatures on many (but not all) of the RDNA4 cards I have tested so far and have received various inquiries from readers, this was the trigger for me to…
www.igorslab.de www.igorslab.de
Click to expand...
I didnt realize memory temp had a target. Must be why my mercury runs high fan speeds all the time even if I set to 240W power limit and core temps are great - memory temps still go into the mid or high 80's
 
gaav87 said:
5600xt->5700xt bios mod xD
Click to expand...
That was actual legit bios mod with signature hack, fully featured editing et al, this is just taking out your trusty CH341a and flashing a random 9070xt bios. You could do the same for 7900xtx, it is just some people burned their flash chips because they didn't use 1.8V adaptor while trying to flash xtx bios which has scared the masses away from crossflashing fun.
 
tsamolotoff said:
That was actual legit bios mod with signature hack, fully featured editing et al, this is just taking out your trusty CH341a and flashing a random 9070xt bios. You could do the same for 7900xtx, it is just some people burned their flash chips because they didn't use 1.8V adaptor while trying to flash xtx bios which has scared the masses away from crossflashing fun.
Click to expand...
Ye i was first in Poland to do 5600xt->5700xt 😛
 
from MLID about 9600XT
baea68a4ce494f8a57e06ea5fdf1852d.jpg
 
FSR4 backported to RDNA3 in Linux by modders. Lots of issues with performance and artifacts, but it technically seems to be possible.


Also another interesting video - Radeon Image Sharpening 2 (RIS 2) tested on RX 9070 -

People will call it oversharpened, and I agree that setting it as 100% is oversharpened a bit, but 50% looks good. I would use it over zero sharpening every time. FSR4 does a decent job of reducing TAA blur, but there is still more to be done and RIS2 is another step forward to reducing blur.
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top