RTX 5090 and others

[arielbit]

NVIDIA board partners expect GeForce RTX 5090 and RTX 5080 to launch in fourth quarter

Board manufacturers anticipate the launch of NVIDIA’s latest RTX 50 series graphics cards in the fourth quarter, initially focusing on high-end products like the 5090 and 5080. Analysts are optimistic that NVIDIA’s biennial major GPU upgrades will boost shipment volumes for board manufacturers this year and drive up the average selling price (ASP) of graphics card products.
— Money UDN (translated)
https://videocardz.com/newz/nvidia-board-partners-expect-geforce-rtx-5090-and-rtx-5080-to-launch-in-fourth-quarter

remember folks, the market might top before, during or after (plus minus a certain time frame), because nvidia will sell those cards as miners panic buy at a premium due to profit hehehehe

[FP91G]

NVIDIA board partners expect GeForce RTX 5090 and RTX 5080 to launch in fourth quarter

Board manufacturers anticipate the launch of NVIDIA’s latest RTX 50 series graphics cards in the fourth quarter, initially focusing on high-end products like the 5090 and 5080. Analysts are optimistic that NVIDIA’s biennial major GPU upgrades will boost shipment volumes for board manufacturers this year and drive up the average selling price (ASP) of graphics card products.
— Money UDN (translated)
https://videocardz.com/newz/nvidia-board-partners-expect-geforce-rtx-5090-and-rtx-5080-to-launch-in-fourth-quarter

remember folks, the market might top before, during or after (plus minus a certain time frame), because nvidia will sell those cards as miners panic buy at a premium due to profit hehehehe

If there is a profit in mining, then miners will again buy any video cards. Prices for top-end Nvidia video cards are not falling, and in many countries there is a very small selection of good 4090 models for mining. And the 5090 video card will also be in good demand, even at inflated prices.

[safar1980]

Nvidia RTX 5090 and 5080 will be announced at same time, says leaker
The new Nvidia RTX 5090 and 5080 are reportedly going to be announced at the same time, according to the latest word on the tech grapevine. The release date of the new GPUs isn’t yet known, but if this rumor is true, we should know the basic details about the top two graphics cards in the new Nvidia Blackwell lineup on the same day.

We’re expecting Nvidia to release at least one of its new GPUs before the end of 2024, and if it’s the RTX 5090 there’s a good chance that it’s going to jump straight onto our best graphics card guide as the fastest gaming GPU. So far, rumors are pointing to the RTX 5080 release date coming before the RTX 5090, but this is the first time we’ve heard that the two GPUs could be announced (not released) on the same day.

[FP91G]

NVIDIA Testing GeForce RTX 50 Series "Blackwell" GPU Designs Ranging from 250 W to 600 W

According to Benchlife.info insiders, NVIDIA is supposedly in the phase of testing designs with various Total Graphics Power (TGP), running from 250 Watts to 600 Watts, for its upcoming GeForce RTX 50 series Blackwell graphics cards. The company is testing designs ranging from 250 W aimed at mainstream users and a more powerful 600 W configuration tailored for enthusiast-level performance. The 250 W cooling system is expected to prioritize compactness and power efficiency, making it an appealing choice for gamers seeking a balance between capability and energy conservation. This design could prove particularly attractive for those building small form-factor rigs or AIBs looking to offer smaller cooler sizes. On the other end of the spectrum, the 600 W cooling solution is the highest TGP of the stack, which is possibly only made for testing purposes. Other SKUs with different power configurations come in between.

We witnessed NVIDIA testing a 900-watt version of the Ada Lovelace AD102 GPU SKU, which never saw the light of day, so we should take this testing phase with a grain of salt. Often, the engineering silicon is the first batch made for the enablement of software and firmware, while the final silicon is much more efficient and more optimized to use less power and align with regular TGP structures. The current highest-end SKU, the GeForce RTX 4090, uses 450-watt TGP. So, take this phase with some reservations as we wait for more information to come out.
https://www.techpowerup.com/322284/nvidia-testing-geforce-rtx-50-series-blackwell-gpu-designs-ranging-from-250-w-to-600-w?cp=2

[OgNasty]

NVIDIA Testing GeForce RTX 50 Series "Blackwell" GPU Designs Ranging from 250 W to 600 W

According to Benchlife.info insiders, NVIDIA is supposedly in the phase of testing designs with various Total Graphics Power (TGP), running from 250 Watts to 600 Watts, for its upcoming GeForce RTX 50 series Blackwell graphics cards. The company is testing designs ranging from 250 W aimed at mainstream users and a more powerful 600 W configuration tailored for enthusiast-level performance. The 250 W cooling system is expected to prioritize compactness and power efficiency, making it an appealing choice for gamers seeking a balance between capability and energy conservation. This design could prove particularly attractive for those building small form-factor rigs or AIBs looking to offer smaller cooler sizes. On the other end of the spectrum, the 600 W cooling solution is the highest TGP of the stack, which is possibly only made for testing purposes. Other SKUs with different power configurations come in between.

We witnessed NVIDIA testing a 900-watt version of the Ada Lovelace AD102 GPU SKU, which never saw the light of day, so we should take this testing phase with a grain of salt. Often, the engineering silicon is the first batch made for the enablement of software and firmware, while the final silicon is much more efficient and more optimized to use less power and align with regular TGP structures. The current highest-end SKU, the GeForce RTX 4090, uses 450-watt TGP. So, take this phase with some reservations as we wait for more information to come out.
https://www.techpowerup.com/322284/nvidia-testing-geforce-rtx-50-series-blackwell-gpu-designs-ranging-from-250-w-to-600-w?cp=2

I'm looking forward to upgrading my GPUs a bit as I'm still mining away with 1080tis and VIIs.  Basically breaking even or maybe making a few cents using my solar power, but no real profit to speak of.  Hopefully we'll see these GPUs release when Bitcoin is peaking so we can maybe pick up some GPUs and take advantage of Bitcoin's high price to secure hashing power for the next cycle.  I did pick up a 3070ti for my gaming rig, but that isn't any better than my Radeon VIIs at mining profitably.  I continue to wait and watch to see what sort of magic comes with the 5xxx series.

[FP91G]

NVIDIA Testing GeForce RTX 50 Series "Blackwell" GPU Designs Ranging from 250 W to 600 W

According to Benchlife.info insiders, NVIDIA is supposedly in the phase of testing designs with various Total Graphics Power (TGP), running from 250 Watts to 600 Watts, for its upcoming GeForce RTX 50 series Blackwell graphics cards. The company is testing designs ranging from 250 W aimed at mainstream users and a more powerful 600 W configuration tailored for enthusiast-level performance. The 250 W cooling system is expected to prioritize compactness and power efficiency, making it an appealing choice for gamers seeking a balance between capability and energy conservation. This design could prove particularly attractive for those building small form-factor rigs or AIBs looking to offer smaller cooler sizes. On the other end of the spectrum, the 600 W cooling solution is the highest TGP of the stack, which is possibly only made for testing purposes. Other SKUs with different power configurations come in between.

We witnessed NVIDIA testing a 900-watt version of the Ada Lovelace AD102 GPU SKU, which never saw the light of day, so we should take this testing phase with a grain of salt. Often, the engineering silicon is the first batch made for the enablement of software and firmware, while the final silicon is much more efficient and more optimized to use less power and align with regular TGP structures. The current highest-end SKU, the GeForce RTX 4090, uses 450-watt TGP. So, take this phase with some reservations as we wait for more information to come out.
https://www.techpowerup.com/322284/nvidia-testing-geforce-rtx-50-series-blackwell-gpu-designs-ranging-from-250-w-to-600-w?cp=2

I'm looking forward to upgrading my GPUs a bit as I'm still mining away with 1080tis and VIIs.  Basically breaking even or maybe making a few cents using my solar power, but no real profit to speak of.  Hopefully we'll see these GPUs release when Bitcoin is peaking so we can maybe pick up some GPUs and take advantage of Bitcoin's high price to secure hashing power for the next cycle.  I did pick up a 3070ti for my gaming rig, but that isn't any better than my Radeon VIIs at mining profitably.  I continue to wait and watch to see what sort of magic comes with the 5xxx series.

I'm sure the 5080 and 5090 graphics cards will have the classic situation of being expensive and overheating very badly.
And miners will have to buy new thermal pads again and install additional cooling and eventually lose their warranty. It's better to buy inexpensive 3070 now if the price of electric power is cheap.

[safar1980]

Leakers claim flagship NVIDIA Blackwell GPU might get sophisticated PCB design
Earlier today, a rumor from Chiphell suggested that NVIDIA’s new RTX 50 Founders Edition might feature a unique memory chip layout resembling the Android mascot, as pointed out by the well-known leaker Kopite7kimi. The original rumor also hinted at a board design split into three parts but lacked specifics.

Kopite7kimi has confirmed that the PCB is divided into three sections: the main board, an I/O rigid board, and a separate PCIe slot component. The latter might function as a riser without any circuits, while the other boards are required to keep up Blackwell’s increased number of components, particularly the new GDDR7 memory.

[FP91G]

Last week, it was revealed that the NVIDIA RTX 5090 PCB will be more complex than the RTX 40 series. Instead of featuring a single board, the RTX 50 series will use three interconnected boards. This design is specific to the Founders Edition model from NVIDIA and may not be used in the reference models from other manufacturers.

NVIDIA likely had strong reasons for opting for such a complex cooler design for the RTX 5090. Prototypes for the GB202 Blackwell GPU, seemingly designed for the upcoming flagship RTX 5090, suggest it will occupy only two slots. This is a significant reduction compared to the bulky RTX 4090 Founders Edition or custom 4090 models, which typically take up 3 to 4 slots
https://videocardz.com/newz/nvidia-geforce-rtx-5090-founders-edition-rumored-to-feature-dual-slot-and-dual-fan-cooler

[elmog]

NVIDIA RTX 5090 to Leverage TSMC’s 3nm (N3) Node, Launching in Late 2024
The upcoming RTX 50 series GPUs, codenamed Blackwell, are planned for a late 2024 release. Like Ada, it will also be a monolithic design. According to Kopite7kimi, these GPUs will be fabbed on TSMC’s 3nm (N3) node. You can expect transistor counts of over a billion and densities of nearly 150 billion/mm². Blackwell should offer core clocks of over 3GHz and bus widths of up to 512 bits, resulting in memory bandwidths rivaling HBM memory.
https://www.hardwaretimes.com/nvidia-rtx-5090-to-leverage-tsmcs-3nm-n3-node-launching-in-late-2024/

AVFA123

[FP91G]

GeForce RTX 5090 rumored memory specs
14x 8Gbit (2GB) GDDR7 modules
28GB total capacity
28 Gbps speed
1568 GB/s maximum memory bandwidth
The 28GB specs on the new flagship may not look as impressive as 32GB. However, this means that NVIDIA will still have room for potential “Ti/SUPER” upgrades in the future. In fact, the 512-bit memory may not be used by gaming GPUs but instead offered for ProViz RTX (formerly Quadro) series, which can utilize a lot more memory capacity.

The last time NVIDIA used a 448-bit memory bus was during the GTX 200 series era. Models like GTX 260 or GTX 275 featured 896 MB of memory. The 512-bit bus was used by GTX 280. The following launches were equipped with up to 384-bit bus (GTX 400 and newer).
https://videocardz.com/newz/nvidia-rtx-5090-new-rumored-specs-28gb-gddr7-and-448-bit-bus

[safar1980]

GeForce RTX 50 Blackwell GPU series leaked
 Kopite7kimi hinted that the configuration of the GB203 GPU is different from what was previously thought. Instead of 96 Streaming Multiprocessors, the actual design has 84 SMs. Despite our best efforts, the leaker did not confirm the exact configurations of the remaining GPUs at the time.

However, today the leaker provided the list of all five upcoming Blackwell GB20X processors along with their configurations. It is now confirmed that not all of them will have the same configuration as the RTX 40 series, but there are at least one will have an identical core count.

The GB202 is still expected to feature 12 GPCs and 8 TPC per GPC, resulting in a total SM count of 192 (assuming 2 per TPC). This, however, does not mean NVIDIA will use such a configuration for the actual product, like the RTX 5090. The retail SKU will likely feature a cut-down configuration to maintain good yields. We also know that NVIDIA Is planning to use a 448-bit memory bus, although the GPU supposedly offers a 512-bit.

The GB203 features a 7×6 configuration, which means a total of 84 Streaming Multiprocessors. This GPU is confirmed to have a 256-bit memory bus, making it close to the AD103 GPU. The new feature is GDDR7 memory support.

[FP91G]

NVIDIA is reportedly cutting back production of its RTX 40 series graphics cards by as much as 50% as it prepares for the RTX 50 series. Such a move is not wholly unheard of before the release of new products but it does mean that PC owners who’ve been holding out for the lowest prices for the current ADA-based GPUs may want to make their move soon since prices could go up significantly in the coming months. Except for NVIDIA’s flagship GeForce RTX 4090, many of its current RTX 40 series lineups can be had at MSRP or less, but that may change within weeks or months.
https://www.thefpsreview.com/2024/08/02/nvidia-is-reportedly-cutting-back-on-rtx-40-series-inventory-price-increases-could-be-imminent/

[FP91G]

According to Wu Haijun, Chairman of Shenzhen Hasee Computer Co., a Chinese company that produces desktops, Mini-PCs, and laptops, the next-generation GeForce RTX 50 graphics cards in the popular X60 laptop segment will require less power. This was stated during a meeting with the media and reported by Golden Pig Upgrade.

It was mentioned that the GeForce RTX XX60 model from NVIDIA’s next-gen laptop series will require 115W of power, compared to 140W in the RTX 40 series. Furthermore, the Chairman stated that these series will use GDDR7 memory.
https://videocardz.com/newz/geforce-rtx-5060-laptop-gpu-confirmed-to-feature-gddr7-memory-and-25w-lower-tgp-than-rtx-4060

[philipma1957]

According to Wu Haijun, Chairman of Shenzhen Hasee Computer Co., a Chinese company that produces desktops, Mini-PCs, and laptops, the next-generation GeForce RTX 50 graphics cards in the popular X60 laptop segment will require less power. This was stated during a meeting with the media and reported by Golden Pig Upgrade.

It was mentioned that the GeForce RTX XX60 model from NVIDIA’s next-gen laptop series will require 115W of power, compared to 140W in the RTX 40 series. Furthermore, the Chairman stated that these series will use GDDR7 memory.
https://videocardz.com/newz/geforce-rtx-5060-laptop-gpu-confirmed-to-feature-gddr7-memory-and-25w-lower-tgp-than-rtx-4060

GDDR7 should be interesting to work with.

[FP91G]

According to Wu Haijun, Chairman of Shenzhen Hasee Computer Co., a Chinese company that produces desktops, Mini-PCs, and laptops, the next-generation GeForce RTX 50 graphics cards in the popular X60 laptop segment will require less power. This was stated during a meeting with the media and reported by Golden Pig Upgrade.

It was mentioned that the GeForce RTX XX60 model from NVIDIA’s next-gen laptop series will require 115W of power, compared to 140W in the RTX 40 series. Furthermore, the Chairman stated that these series will use GDDR7 memory.
https://videocardz.com/newz/geforce-rtx-5060-laptop-gpu-confirmed-to-feature-gddr7-memory-and-25w-lower-tgp-than-rtx-4060

GDDR7 should be interesting to work with.

This would probably be interesting in mining Ethereum and its forks. New algorithms are not very memory-dependent, but mainly depend on the capabilities of the video card core, which is why top Nvidia video cards are in the very first positions of mining calculators.

[safar1980]

Nvidia RTX 5090 and 5080 Designs Will Be Finalized Soon, TDPs Lower Than Reported

Two years ago, gamers had a date on their calendars for Nvidia's Ada Lovelace RTX 40-series reveal—Sept. 20—and yet here we are in early September with an empty calendar. Not cool, Nvidia—not cool. However, according to a new report Nvidia is finalizing the design of its top two Blackwell GeForce cards this month, indicating a launch can't be that far away. Both cards' TDP will also be lower than people expect but still higher than Ada Lovelace.

The newest report about Nvidia's plans comes from the Chinese site benchlife.info, which is known to report on the GPU industry via supply chain insiders. The site's newest post says Nvidia is finalizing the design for the RTX 5090 and RTX 5080 this month, hinting at a holiday launch for the GPUs, though the post says they'll launch at CES in January 2025, which we hope is not the case.

 

[FP91G]

GeForce RTX 5090 to feature 21760 CUDA cores, 32GB GDDR7 memory and 600W, RTX 5080 gets 16GB VRAM

One of the most reliable NVIDIA leakers has now confirmed the specs for two of NVIDIA’s upcoming Blackwell graphics cards, specifically the RTX 5090 and RTX 5080. According to Kopite7kimi, the RTX 5090 is now said to feature a 512-bit memory bus, not 448-bit as mentioned earlier. The RTX 5090 will also feature 32GB of GDDR7 memory.

RTX 5090 with 170 SMs, 32GB VRAM, return of 512-bit bus
The flagship Blackwell gaming model is expected to use the GB202-300 GPU with 21,760 FP32 (CUDA) cores, which is fewer than the 24,576 cores of the full chip (a reduction of 13%). Additionally, the card is rumored to have a 600W spec, though the leaker has not confirmed whether that’s the TBP or TGP, as both can refer to different metrics.

Interestingly, although the power has increased from 450W to 600W, the card is still said to feature a 2-slot cooler (Founders Edition). This likely means that NVIDIA is using some kind of non-standard cooling design, perhaps a liquid cooler.

https://videocardz.com/newz/nvidia-geforce-rtx-5090-and-rtx-5080-specs-leaked

[safar1980]

NVIDIA RTX 5090/5080 in January, 5070/Ti in February, 5060 series in March
The company reportedly fast-forwards its plans to launch the update to RTX series within the first quarter of next year.

If this rumor is true, gamers won’t have to wait much longer for the replacement of the entire RTX 40 series. According to a new post on Board Channels, which is usually reliable for launch rumors (though not always perfect), NVIDIA has been discussing high-end and mid-range updates with its partners, specifically the RTX 5070 and 5060 series.

It’s claimed that in January, gamers can expect the RTX 5090/D and RTX 5080 to be released within the same month. In February, NVIDIA is expected to launch the RTX 5070 and 5070 Ti. Previously, there was no mention of the Ti variant, so this rumor is relatively new.

Perhaps the biggest revelation from the Board Channels post is the news about the RTX 5060 and 5060 Ti. These SKUs haven’t been mentioned in any previous rumors, even from the most respected and reliable leakers, so this part remains unconfirmed.

NVIDIA RTX 50 Launch Dates:
RTX 5090/5090D/5080: January 2025
RTX 5070/5070Ti: February 2025
RTX 5060/5060Ti: March 2025