GPUs Latest news

NVIDIA RTX 5000 Ada as a workstation graphics card with 15360 cores and 32 GB VRAM? The portfolio makes interesting reading

The RTX 4000 Ada is a high-performance graphics card model with NVIDIA’s AD104 graphics processor. This processor offers high computing power and is optimized for use in professional applications such as design, animation, rendering and simulations. By using the AD104 GPU, the RTX 4000 Ada is able to handle complex tasks quickly and efficiently. However, a special feature of the RTX 4000 Ada is its efficiency. The card has a TDP of 70 watts and does not require any additional power connections.

This is due to the improved energy efficiency of the AD104 processor. In addition, the RTX 4000 Ada features a small form factor (SFF) design, which means it is compact and space-saving and ideal for use in small desktop workstations. The RTX 6000 Ada is the high-end model in the Ada desktop workstation family, delivering incredible performance for demanding applications like AI training and big data processing. The RTX 4000 Ada, on the other hand, offers slightly lower performance, but is still very powerful and efficient enough for many professional applications. It is ideal for professional users who need a powerful, yet space-saving graphics card for their workstation. And yet, until now, there was still a gap in the portfolio

Leaker Kopite7kimi has shared information about a possible new graphics card from NVIDIA called RTX 5000 Ada. The graphics card is to use the AD102 GPU die, which was chosen for its high core count. The AD102 GPU die is already installed in the top-of-the-line RTX 6000 Ada GPU, but there is still room for more iterations down the line because the chip contains a total of 144 SMs for 18,432 cores. The high number of presumably 15360 cores allows the RTX 5000 Ada to be very powerful and support a wide range of applications, including gaming, AI, VR, and more. The multiple use of the AD102 GPU die also shows NVIDIA’s commitment to expand the portfolio sensibly.

The NVIDIA RTX 5000 Ada will feature a greatly reduced die size and will be powered by 32 GB of standard GDDR6 memory running over a 256-bit bus interface. According to this specification, we can expect a maximum bandwidth of up to 640 GB/s, since NVIDIA uses a similar DRAM specification as in the RTX 6000 Ada. The reduced die size of the RTX 5000 Ada indicates that it is a slightly lower-end variant compared to the RTX 6000 Ada.

However, this does not mean that the RTX 5000 Ada is not capable of performing demanding tasks in the field of graphics applications and machine learning. The 32GB of standard GDDR6 memory used on the RTX 5000 Ada provides fast data transfer speeds and high capacity to support demanding applications. In addition, the 256-bit bus interface and similar DRAM specification to the RTX 6000 Ada can maximize bandwidth and optimize graphics card performance.

The TDP stands for Thermal Design Power and indicates how much power a graphics card requires and how much heat it produces in the end. NVIDIA cards currently have a TDP of around 200-250W, but there is always room for optimization and improvement. The statement implies that NVIDIA has the ability to further optimize the cards and increase the TDP to allow for higher specifications and more performance. The future RTX *500 series could possibly have higher VRAM capacities and more cores to increase performance.

However, it is noted that this is just a speculation and depends on NVIDIA what specifications they will implement in their future graphics cards. It is also assumed that the new cards will have a higher price level, possibly between 2800 and 4200 euros. However, this price has not yet been officially confirmed and depends on various factors, such as the components used and the manufacturing costs. However, it remains to be seen what specifications and prices NVIDIA will set for its future graphics cards.

NVIDIA RTX Ada Workstation Portfolio:

GRAPHICS CARD RTX 6000 RTX 5000 * (Leak)
RTX 4000
GPU Ada Lovelace GPU Ada Lovelace GPU Ada Lovelace GPU
GPU Process TSMC 4N TSMC 4N TSMC 4N
The Size 608mm2 (AD102) 608mm2 (AD102) 295mm2 (AD104)
GPU Cores 18176 Cores 15360 Cores 6144 Cores
Tensor Cores 568 cores TBD 192 cores
Boost Clock 2.50 GHz TBD ~1.70 GHz
Single Precision 91.1 TFLOPs TBD 19.2 TFLOPs
Tensor Performance 1458 TFLOPs TBD 306.8 TFLOPs
VRAM 48 GB GDDR6X 32 GB GDDR6 20 GB GDDR6
NVLINK VRAM N/A N/A N/A
Memory bus 384-bit 256-bit 160-bit
Memory Bandwidth 960 GB/s 640 GB/s 320 GB/s
TDP 300W 175-250W 70W
Launch Price $6800 US $3000-$4500 US $1250 US
Launch Date Q1 2023 TBD Q1 2023

Bisher keine Kommentare

Kommentar

Lade neue Kommentare

Redaktion

Artikel-Butler

1,786 Kommentare 8,642 Likes

Die RTX 4000 Ada ist ein leistungsstarkes Grafikkartenmodell mit dem AD104-Grafikprozessor von NVIDIA. Dieser Prozessor bietet eine hohe Rechenleistung und ist für den Einsatz in professionellen Anwendungen wie Design, Animation, Rendering und Simulationen optimiert. Durch die Verwendung des AD104-Grafikprozessors ist die RTX 4000 Ada in der Lage, komplexe Aufgaben schnell und effizient zu bewältigen. Eine (den ganzen Artikel lesen...)

Antwort 1 Like

Danke für die Spende



Du fandest, der Beitrag war interessant und möchtest uns unterstützen? Klasse!

Hier erfährst Du, wie: Hier spenden.

Hier kannst Du per PayPal spenden.

About the author

Samir Bashir

Werbung

Werbung