NVIDIA’s next-gen GeForce RTX 40 Founders Edition cooler for high-end GPUs reportedly leaks

Subsequent-Gen NVIDIA GeForce RTX 40 Sequence Founders Version Design Reportedly Leaked By QbitLeaks. The brand new design exhibits us a preview of the up to date Founders Version GPU cooler that will likely be featured on high-end Ada Lovelace GPUs such because the RTX 4090 and 4080.

NVIDIA’s Large GeForce RTX 40 Founders Version cooler is rumored to be leaking and coming quickly to an RTX 4090 and RTX 4080 graphics card!

In response to the leaker, this new cooler design that was photographed will likely be a part of the GTC ’22 keynote to be held two weeks later. No particulars are offered as to which card it’s, however the up to date shroud and fan design matches the leaked RTX 4080 cooler just a few days in the past. It seems like NVIDIA will likely be reusing its current cooler design and giving it a slight visible overhaul.

We have seen in earlier leaks that the heatsink below the shroud has been up to date with a bigger thermal contact floor that covers the GPU, VRAM, and VRMs. The followers have been up to date to a 7-blade fan design and are additionally barely bigger, however the general look of the shroud and design stays basically the identical as the present Founders Version graphics playing cards. This up to date design ought to assist present significantly better cooling for Ada Lovelace GPUs which is able to devour a variety of energy. The playing cards may even use a PCIe Gen 5.0 connector interface and a Gen 5.0 energy interface via the brand new 16-pin connectors.

NVIDIA GeForce RTX 4090 Founders Version Graphics Card Cooler / Shroud Leak:

NVIDIA GeForce RTX 4080 Founders Version Graphics Card Cooler / Shroud Leak:

The leaker says this picture is a teaser for GTC, so we may undoubtedly see CEO Jensen speaking about his line of next-gen video games through the occasion. GTC was once an information heart and HPC particular occasion, however with NVIDIA’s latest drop in income, primarily because of the falling gaming market, the corporate refocused its most important occasion and made it devoted to players.

NVIDIA GeForce RTX 4080 “anticipated” specs

The NVIDIA GeForce RTX 4080 is predicted to make use of a scaled-down AD103-300 GPU configuration with 9,728 cores or 76 SMs enabled out of a complete of 84 items, whereas the earlier configuration supplied 80 SMs or 10,240 cores. Whereas the complete GPU comes with 64MB L2 cache and as much as 224 ROPs, the RTX 4080 may find yourself with 48MB L2 cache and decrease ROPs additionally because of its scaled-down design. The board needs to be based mostly on the PG136/139-SKU360 PCB.

As for reminiscence specs, the GeForce RTX 4080 is predicted to rock 16GB GDDR6X capacities which might be scaled to 23Gbps speeds on a 256-bit bus interface. This can present as much as 736 GB/s of bandwidth. That is nonetheless a bit slower than the 760 GB/s bandwidth supplied by the RTX 3080 because it comes with a 320-bit interface however a low 10 GB capability. To compensate for the decrease bandwidth, NVIDIA may combine a next-generation reminiscence compression suite to compensate for the 256-bit interface.

For energy, the TBP is now set to be rated at 340W, a 20W improve from the earlier 320W spec we bought. This brings the TBP to the identical stage as the prevailing RTX 3080 graphics card (as much as 350W). Now, it is unknown if the opposite RTX 40 sequence graphics playing cards may even obtain the sooner GDDR6X reminiscence processing, however we do know that Micron has began full mass manufacturing of GDDR6X reminiscence modules as much as 24 Gbps, they should go someplace.

• NVIDIA GeForce RTX 4080 TBP “anticipated” – 340W
• NVIDIA GeForce RTX 3080 TBP “Official” – 350W

NVIDIA GeForce RTX 4080 Sequence Preliminary Specs:

Graphics card title	NVIDIA GeForce RTX 4080Ti	NVIDIA GeForce RTX 4080	NVIDIA GeForce RTX 3090Ti	NVIDIA GeForce RTX 3080
GPU title	Ada Lovelace AD102-250?	Ada Lovelace AD103-300?	Ampere GA102-225	Ampere GA102-200
Course of node	TSMC4N	TSMC4N	Samsung 8nm	Samsung 8nm
Die dimension	~450mm2	~450mm2	628.4mm2	628.4mm2
Transistors	To be decided	To be decided	28 billion	28 billion
CUDA Cores	14848	9728?	10240	8704
TMU / ROP	TBD/232?	TBD / 214?	320 / 112	272 / 96
Tensor / Cores RT	TBD / TBD	TBD / TBD	320 / 80	272 / 68
Primary Clock	To be decided	To be decided	1365MHz	1440MHz
Enhance clock	~2600MHz	~2500MHz	1665MHz	1710MHz
Calculation FP32	~55TFLOPs	~50 TFLOPs	34 TFLOPs	30 TFLOPs
RT TFLOP	To be decided	To be decided	67 TFLOPs	58 TFLOPs
Tensor TOP	To be decided	To be decided	273 TOPs	238 TOPs
Reminiscence capability	20 GB GDDR6X	16 GB GDDR6X? 12GB GDDR6X?	12 GB GDDR6X	10 GB GDDR6X
Reminiscence bus	320 bit	256 bit? 192 bit?	384 bit	320 bit
Reminiscence velocity	21.0 Gbps?	23.0 Gbps?	19 Gbps	19 Gbps
Bandwidth	840 GB/s	736 GB/s? 552 GB/s?	912 Gbps	760 Gbps
PAD	450W	340W	350W	320W
Value (MSRP / FE)	US$1199?	$699?	$1199	US$699
Launch (Availability)	2023?	July 2022?	June 3, 2021	September 17, 2020

NVIDIA GeForce RTX 4090 “anticipated” specs

The NVIDIA GeForce RTX 4090 will use 128 SM out of the 144 SM for a complete of 16,384 CUDA cores. The GPU will include 96MB of L2 cache and a complete of 384 ROPs, which is simply insane. Clock speeds aren’t but confirmed, however provided that the TSMC 4N course of is used, we anticipate clocks between 2.0 and three.0 GHz.

So far as reminiscence specs go, the GeForce RTX 4090 is predicted to rock 24GB GDDR6X capacities which will likely be clocked at 21Gbps speeds on a 384-bit bus interface. This can present as much as 1 TB/s of bandwidth. This is similar bandwidth as the prevailing RTX 3090 Ti graphics card and on the subject of energy consumption, the TBP is alleged to be rated at 450W, which implies the TGP could possibly be decrease than that. The board will likely be powered by a single 16-pin connector that delivers as much as 600W of energy. We’re prone to get 500W+ customized designs like we noticed with the RTX 3090 Ti.

As for its characteristic set, NVIDIA GeForce RTX 4090 and RTX 4080 graphics playing cards will rock all trendy NV characteristic units comparable to newest 4th Gen Tensor cores, third Gen RT cores, newest NVENC encoder and the NVCDEC decoder, and help for the newest APIs. They may even include all trendy RTX options comparable to DLSS, Reflex, Broadcast, Resizable-BAR, Freestyle, Ansel, Highlights, Shadowplay and G-SYNC.

• NVIDIA GeForce RTX 4090 TBP “anticipated” – 450W
• NVIDIA GeForce RTX 3090 TBP “Official” – 350W

NVIDIA GeForce RTX 4090 Ti and RTX 4090 “preliminary” specs:

Graphics card title	NVIDIA GeForce RTX 4090Ti	NVIDIA GeForce RTX 4090	NVIDIA GeForce RTX 3090Ti	NVIDIA GeForce RTX 3090
GPU title	Ada Lovelace AD102-350?	Ada Lovelace AD102-300?	Ampere GA102-350	Ampere GA102-300
Course of node	TSMC4N	TSMC4N	Samsung 8nm	Samsung 8nm
Die dimension	~600mm2	~600mm2	628.4mm2	628.4mm2
Transistors	To be decided	To be decided	28 billion	28 billion
CUDA Cores	18432	16128	10752	10496
TMU / ROP	TBD / 384	TBD / 384	336 / 112	328 / 112
Tensor / Cores RT	TBD / TBD	TBD / TBD	336 / 84	328 / 82
Primary Clock	To be decided	To be decided	1560MHz	1400MHz
Enhance clock	~2800MHz	~2600MHz	1860MHz	1700MHz
Calculation FP32	~103 TFLOPs	~90 TFLOPs	40 TFLOPs	36 TFLOPs
RT TFLOP	To be decided	To be decided	74 TFLOPs	69 TFLOPs
Tensor TOP	To be decided	To be decided	320 TOP	285 TOPs
Reminiscence capability	24 GB GDDR6X	24 GB GDDR6X	24 GB GDDR6X	24 GB GDDR6X
Reminiscence bus	384 bit	384 bit	384 bit	384 bit
Reminiscence velocity	24.0 Gbps	21.0 Gbps	21.0 Gbps	19.5 Gbps
Bandwidth	1152 GB/s	1008 GB/s	1008 GB/s	936 Gbps
TGP	600W	450W	450W	350W
Value (MSRP / FE)	US$1999?	US$1499?	US$1999	US$1499
Launch (Availability)	2023?	October 2022?	March 29, 2022	September 24, 2020

NVIDIA GeForce RTX 40 sequence graphics playing cards, together with the RTX 4080 and RTX 4070 graphics playing cards, will likely be among the many first graphics playing cards moreover the RTX 4090 to be launched to players. The RTX 4090 is to this point anticipated to launch on October 22, however an unveiling is predicted throughout NVIDIA’s GTC keynote later this month.