Latest news

Teclab unveils the world’s fastest RTX 4090 with 26 Gbps memory

In the world of graphics cards, there are always projects that go beyond the usual limits and set new standards. One such project was recently realized by Teclab: they developed a customized version of the NVIDIA GeForce RTX 4090 based on the AD102 “Ada” chip and equipped with a faster G6X memory. This graphics card, which they themselves call the GeForce RTX 4090 SUPER, is an impressive demonstration of what is possible when you combine expertise, experience and the right tools. Although NVIDIA has not yet released an official “SUPER” version of the RTX 4090, Teclab shows that there is certainly potential for such a card.

Source: TecLab

The creation of the GeForce RTX 4090 SUPER

The idea for the development of the GeForce RTX 4090 SUPER was born at a time when the RTX 3090 Ti was introduced by NVIDIA. This card appeared just a few months before the introduction of the RTX 40 series and used the GA102 GPU chip. Interestingly, the PCB (Printed Circuit Board) of the RTX 3090 Ti showed more similarities to the later RTX 4090 than to the regular RTX 3090, particularly in terms of memory connections and the card’s power supply.

When the RTX 4090 finally came onto the market, these assumptions were confirmed. The differences between the PCB of the RTX 3090 Ti and the RTX 4090 were minimal, which gave Teclab the idea to make a modification. The aim was to transform the already powerful RTX 3090 Ti into an even more powerful card by replacing the GA102 chip with the newer AD102 chip and adapting the memory connections and power supply accordingly.

The choice of components

Teclab opted for the RTX 3090 Ti HOF OC LAB variant as the basis for its modifications. This model was chosen for several reasons. Firstly, the HOF OC LAB version has a water cooling block, which enables improved cooling without having to resort to exotic cooling methods such as liquid nitrogen (LN2). Secondly, this card already provided a solid basis for the conversion, as it was designed for overclocking and high performance requirements.

The modifications Teclab made to the RTX 3090 Ti to turn it into the GeForce RTX 4090 SUPER included replacing the GA102 chip with the AD102, integrating 24 Gbps GDDR6X memory and making adjustments to the memory connections and voltage regulator. These changes were necessary to bring the card up to the performance level Teclab was aiming for.

A complex conversion process

Although it may seem simple at first glance to replace a chip and memory, the conversion process was anything but trivial. Teclab not only had to have the necessary expertise, but also special tools and circuit diagrams to carry out the conversion successfully. Thanks to its connections with GALAX, a manufacturer of graphics cards, Teclab was able to draw on the necessary resources.

A particularly critical aspect of the conversion was the adjustment of the power routing, memory connections and voltage limits. These had to be carefully adjusted to ensure that all components worked together smoothly. An error in this process could have rendered the entire card unusable.

For cooling, Teclab relied on the existing water cooling block of the GALAX GeForce RTX 3090 Ti HOF OC LAB. However, this was slightly modified to ensure optimum contact between the cooling block and the components. The cooling itself is provided by two 360mm radiators installed in a push/pull configuration. The system is supported by a powerful water pump, which ensures stable cooling even at high overclocks.

Test and performance of the GeForce RTX 4090 SUPER

After completing the modifications, Teclab subjected the GeForce RTX 4090 SUPER to a series of performance tests to determine the actual performance of the card. These included the Unigine Superposition 8K benchmark, a demanding test that measures the capabilities of the graphics card under extreme conditions.

The results were impressive. The GPU was overclocked to 3105 MHz and the G6X memory reached a speed of 25.8 Gbps. These values are particularly remarkable as they are close to the physical limits of the components. Despite this high load, the card remained stable and showed no signs of overheating or other problems.

Compared to conventional RTX 4090 models, which are often cooled with liquid nitrogen, the GeForce RTX 4090 SUPER from Teclab performed excellently. Even without the use of LN2, it outperformed several benchmarks achieved with LN2-cooled RTX 4090s. This shows that the water cooling Teclab used was sufficient to keep the card cool and stable even during extreme overclocks

Unigine Superposition 8K (Higher is better)  
RTX 4090 SUPER WC (Max GPU Max Mem Tweak LOD)
17454
RTX 4090 SUPER (GPU Mem OC) 14537
RTX 4090 SUPER (Mem OC) 14114
RTX 4090 SUPER (Stock) 12574
RTX 4090 Stock 12479

Teclab’s project to develop the GeForce RTX 4090 SUPER is an impressive demonstration of the potential that lies in the modification and optimization of graphics cards. Although it may be unrealistic for most users to carry out such complex modifications themselves, this project illustrates what is possible with the right expertise, the right equipment and a high level of commitment.

The GeForce RTX 4090 SUPER from Teclab is an example of how far the limits of graphics card performance can be pushed. It remains to be seen whether NVIDIA itself will launch an official “SUPER” version of the RTX 4090, but Teclab has already shown that it is possible to develop such a card – and with impressive results.

Overall, Teclab’s work is an inspiring example of the innovation and technical know-how that exists in the world of computer modification. It is a project that shows that even the most powerful components can be further improved if you are prepared to take the necessary steps and not shy away from a challenge

Source: TechLab

Kommentar

Lade neue Kommentare

H
Helldog38

Mitglied

47 Kommentare 19 Likes

Also ich finde die Steigerung von 4090 Stock auf 4090 Super echt enttäuschend.

Erst durch übertakten kommt was dabei raus.

Antwort 1 Like

G
GPUModder

Mitglied

84 Kommentare 32 Likes

Das für mich spannendste ist die enorme Steigerung der 4090 alleine aufgrund Speicher OC. Das bringt neben dem Max Setting, was wahrscheinlich nichts mehr mit Alltag zu tun hat, mit Abstand am Meisten. Ein wenig in die Zukunft geblickt, könnte das der 5090 mit GDDR7 doch deutlich zur Mehrleistung verhelfen. Bin ich gespannt, was sich da noch so alles ergibt in die Richtung. Vor allem wäre es aus meiner Sicht interessant, bei solchen Themen nicht nur immer die 4090 zu beleuchten, sondern auch mal Karten die man sich noch halbwegs leisten kann wie die 70er/ti und vielleicht grad noch 80er Karten. Zudem wäre auch interessant, wenn man mal abweicht von dem synthetischen Benchmark, hin zu Games, was da dann von der Mehrleistung in der Praxis wirklich noch ankommt? Viele Fragen, trotzdem für mich ein sehr interessanter Artikel! :) Danke.

Antwort 1 Like

S
S.nase

Urgestein

1,476 Kommentare 499 Likes

Könnte auch ein Beispiel sein, wieviel Leistungspotenzial aus marktpolitischen Gründen liegen gelassen wird, um den Leistungsabstand zur nächsten GrafikkartenAusbaustufe oder GrafikkartenGeneration nicht zu klein werden zu lassen.

Am Ende will ein Hersteller ja wahrscheinlich auch den swettspot treffen, bei dem das HerstellungsKosten/Leistungsverhältnis am am besten ist. Klar könnte er auch die schnellste 4090 konstruieren. Nur hätte die dann einen Verkaufspreis den keiner mehr zahlen will.

Beide Situationen könnten den wirtschaftlichen Erfolg des Herstellers gefährden.

Antwort 2 Likes

O
Omega

Neuling

8 Kommentare 2 Likes

Abgesehen von dem wirtschaftlichen Aspekt könnte ich mir auch ganz gut vorstellen, dass ein Hersteller, der so aus der Reihe tanzt auch schnell wieder von NVIDIA auf Spur gebracht wird, wenn er so was in Serie bringen würde.

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