AMD & NVIDIA Next-Gen Flagship GPUs Detailed: RDNA 3 Radeon RX 7900 XT With 15360 Cores, Ada Lovelace GeForce RTX 4090 With 18432 Cores

Rumored specifications of the adjacent-generation AMD RDNA 3 powered Radeon RX 7900 XT & NVIDIA Ada Lovelace powered GeForce RTX 4090 graphics cards have been detailed once once again. The rumors come up from Greymon55 who has been actively posting details regarding upcoming hardware such as CPUs and GPUs over on his Twitter feed for a while at present.

AMD RDNA iii & NVIDIA Ada Lovelace GPU Powered Next-Gen Flagship Radeon RX 7900 XT & GeForce RTX 4090 Graphics Cards Detailed

The AMD RDNA 3 powered Navi 31 and Ada Lovelace-powered AD102 GPUs are expected to bring huge performance improvements and volition too exist the almost power-hungry fries ever made. While NVIDIA is aiming for a monolithic approach with its Ada Lovelace architecture, AMD will apply a total MCM design which it has already taken leadership in with the launch of the CDNA 2 powered MI200 serial 'Aldebaran'. AMD will leverage the aforementioned MCM engineering for consumer-end & gaming GPUs now. So allow's talk nigh the latest rumored specifications that we take got from the leaker:

AD102
5nm TSMC
144SM 18432CUDA
384bit 24G GDDR6X 21Gbps？
2.3~ii.5GHz？
85~92T FP32？

— Greymon55 (@greymon55) November 9, 2021

AMD Radeon RX 7900 XT Graphics Card - RDNA 3 Powered Navi 31 Flagship GPU

The AMD Navi 31 GPU, the flagship RDNA 3 scrap, would power the adjacent-gen Radeon RX 7900 XT graphics carte du jour. We accept heard that AMD will drop CU (Compute Units) in favor of WGP (Work Group Processors) on its adjacent-gen RDNA three GPUs. Since Navi 31 is an MCM GPU, it volition feature ii fundamental IPs, a GCD (Graphics Core Die) based on TSMC's 5nm process and an MCD (Multi-Cache Die) based on the TSMC 6nm process node. Earlier rumors advise that AMD has already tapped out its Navi 31 GPU die.

The Navi 31 GPU configuration shown hither features two GCD'southward (Graphics Core Die) and a single MCD (Multi-Enshroud Dice). Each GCD has 3 Shader Engines (6 in total) and each Shader Engine has 2 Shader Arrays (two per SE / 6 per GCD / 12 in total). Each Shader Array is composed of 5 WGPs (10 per SE / 30 per GCD / 60 in total) and each WGP features eight SIMD32 units with 32 ALUs (40 SIMD32 per SA / eighty per SE / 240 per GCD / 480 in total). These SIMD32 units combine to make upwards 7,680 cores per GCD and 15,360 cores in total.

Performance-wise, the GPU is expected to feature a clock speed of 2.four - 2.5 GHz which puts its theoretical performance at effectually 75 TFLOPs (FP32). This is an insane 226% improvement vs a Radeon RX 6900 XT graphics card.

The Navi 31 (RDNA 3) MCD will be linked to the dual GCD's via a next-generation Infinity Cloth interconnect and feature 256-512 MB of Infinity Cache. Each GPU should also feature 4 memory connect links (32-bit). That's a total of 8 32-bit memory controllers for a 256-bit bus interface. It is stated that the carte du jour will characteristic upwardly to 32 GB of GDDR6 memory running at xviii Gbps pin speeds & that delivers upward to 576 GB/s bandwidth. Another rumor that appeared recently suggests that AMD will be using 3D Infinity Cache technology on its RDNA 3 lineup which volition integrate the new cache in vertical stacks on the GPUs, similar to how the Vermeer-10 chips will stack L3 cache over the CCD.

AMD RDNA GPU (Generational Comparison) Preliminary:

GPU Name	Navi 10	Navi 21	Navi 31
GPU Process	7nm	7nm	5nm (6nm?)
GPU Parcel	Monolithic	Monolithic	MCD (Multi-Chiplet Die)
Shader Engines	2	4	six
GPU WGPs	20	40	30 (Per MCD) 60 (In Full)
SPs Per WGP	128	128	256
Compute Units (Per Dice)	40	fourscore	120 (per MCD) 240 (in total)
Cores (Per Die)	2560	5120	7680
Cores (Total)	2560	5120	15360 (two x MCD)
Memory Bus	256-chip	256-flake	256-chip
Retention Type	GDDR6	GDDR6	GDDR6
Memory Capacity	eight GB	sixteen GB	32 GB
Infinity Enshroud	Northward/A	128 MB	256-512 MB
Flagship SKU	Radeon RX 5700 XT	Radeon RX 6900 XTX	Radeon RX 7900 XT
TBP	225W	330W	350-550W
Launch	Q3 2019	Q4 2020	Q4 2022

NVIDIA GeForce RTX 4090 Graphics Carte - Ada Lovelace Powered AD102 Flagship GPU

Based upon previous rumors, there take been whispers that NVIDIA would utilize TSMC'southward N5 (5nm) process node for its Ada Lovelace GPUs. This includes the AD102 SKU too which will be an entirely monolithic design. In his latest tweet which talks nearly the specific GPU configurations, the AD102 GPU is said to feature a clock speed every bit loftier as 2.5 GHz (2.iii GHz average heave). The specific tweet states that the GPU clock for Ada Lovelace 'AD102' could be 2.3 GHz or greater so allow'due south take that as a baseline and previously leaked specifications to figure out where the performance should land.

The NVIDIA AD102 "ADA GPU" appears to have 18432 CUDA Cores based on the preliminary specs (which can alter), housed within 144 SM units. This is nigh twice the cores present in Ampere which was already a massive footstep up from Turing. A 2.3-2.five GHz clock speed would requite u.s.a. upward to 85 to 92 TFLOPs of compute functioning (FP32). This is more than twice the FP32 functioning of the existing RTX 3090 which packs 36 TFLOPs of FP32 compute ability.

The 150% performance jump looks huge but one should remember that NVIDIA already gave a large leap in FP32 numbers this generation with Ampere. The Ampere GA102 GPU (RTX 3090) offers 36 TFLOPs while the Turing TU102 GPU (RTX 2080 Ti) offered xiii TFLOPs. That's over a 150% increment in FP32 Flops merely the real-earth gaming performance increase for the RTX 3090 averaged at around 50-60% faster over the RTX 2080 Ti. So one thing we shouldn't forget is that Flops don't equal GPU gaming performance these days. Furthermore, we don't know if 2.iii-2.5 GHz is the average heave or the height boost with the sometime meaning that at that place could be even higher compute potential for AD102.

Aside from that, the leaker also states that the NVIDIA GeForce RTX forty flagship would retain a 384-bit bus interface, like to the RTX 3090. What's interesting is though that the leaker mentions G6X which means that NVIDIA won't be moving to a new retentivity standard until after Ada Lovelace and utilize the higher pin-speeds of G6X of 21 Gbps for its next-generation cards before we see a newer standard (e.thou. GDDR7). The menu will characteristic 24 GB of memory so nosotros can either wait single-sided 16Gb DRAM or dual-sided 8Gb DRAM modules.

NVIDIA CUDA GPU (RUMORED) Preliminary:

GPU	TU102	GA102	AD102
Compages	Turing	Ampere	Ada Lovelace
Procedure	TSMC 12nm NFF	Samsung 8nm	5nm
Graphics Processing Clusters (GPC)	vi	seven	12
Texture Processing Clusters (TPC)	36	42	72
Streaming Multiprocessors (SM)	72	84	144
CUDA Cores	4608	10752	18432
Theoretical TFLOPs	16.1	37.6	~xc TFLOPs?
Memory Type	GDDR6	GDDR6X	GDDR6X
Memory Bus	384-flake	384-bit	384-bit
Retentivity Capacity	11 GB (2080 Ti)	24 GB (3090)	24 GB (4090?)
Flagship SKU	RTX 2080 Ti	RTX 3090	RTX 4090?
TGP	250W	350W	450-650W?
Release	Sep. 2018	Sept. 20	2022 (TBC)

The NVIDIA Ada Lovelace GPUs volition ability the next-generation GeForce RTX twoscore graphics cards that will go head-on with AMD'southward RDNA 3 based Radeon RX 7000 series graphics cards. In that location's notwithstanding some speculation regarding the use of MCM by NVIDIA. The Hopper GPU, which is primarily aimed at the Datacenter & AI segment, is allegedly taping out soon and volition characteristic an MCM architecture. NVIDIA won't be using an MCM design on its Ada Lovelace GPUs so they will keep the traditional monolithic design.

Which next-generation GPUs are you looking frontward to the most?

Source: https://wccftech.com/amd-nvidia-next-gen-flagship-gpus-detailed-rdna-3-radeon-rx-7900-xt-with-15360-cores-ada-lovelace-geforce-rtx-4090-with-18432-cores/

Posted by: ricecolip1997.blogspot.com