ASUS TUF Gaming B760-PLUS WIFI D4 review
Netac NV7000 2 TB NVMe SSD Review
ASUS GeForce RTX 4080 Noctua OC Edition review
MSI Clutch GM51 Wireless mouse review
ASUS ROG STRIX B760-F Gaming WIFI review
Asus ROG Harpe Ace Aim Lab Edition mouse review
SteelSeries Arctis Nova Pro Headset review
Ryzen 7800X3D preview - 7950X3D One CCD Disabled
MSI VIGOR GK71 SONIC Blue keyboard review
AMD Ryzen 9 7950X3D processor review
Sapphire To Release Vapor-X edition for the Radeon RX 7900 XTX & XT
Sapphire has revealed in a teaser that it will relaunch the "Vapor-X" graphics card series, which is cooled with the help of an evaporator chamber ("Vapor-Chamber") for the Radeon RX 7900 XTX and 7900 XT. That's a return of the series after an absence of more than six years.
After a sabbatical of more than six years, the Vapor-X series—famous for its high-end custom designs and having debuted in 2008 with the Sapphire Radeon HD 3870 Vapor-X—is making a comeback with the Radeon RX 7000 series, which utilises the same cooler with Vapor Chamber Technology ("VCT") as the Atomic and Toxic models.
I had to look it up, but the Sapphire Radeon R9 390X Vapor-X was the last in the Vapor-X series, although the Toxic series with AiO water cooler was also represented in the Radeon RX 6900 XT (test). The Radeon RX 7900 XTX and 7900 XT, as well as manufacturers' custom designs based on the RDNA 3 architecture, could change that on December 13.
Sapphire's New Nitro RX 7900 GPU: A Teaser - 11/24/2022 09:48 AM
Sapphire has just hinted at yet another future card on social media, yeah, the Radeon RX 7900 series. ...
Review: Sapphire Radeon RX 6400 Pulse (4GB) - 09/08/2022 01:38 PM
Meet AMD's Radeon RX 6400. It has 4GB of 64-bit RAM and a whiff of Inifinity Cache (16MB L3). This card might operate at Radeon RX 480 levels when connected to a PCIe Gen x4 connection (3.0/4.0). Rea...
Sapphire Adds 8GB SKU to Radeon RX 6500 XT Graphics Card line - 07/26/2022 08:55 AM
And that's unusual but certainly a good move. With 8 GB of video memory, twice the SKU's norm of 4 GB. The Sapphire Pulse RX 6500 XT 8 GB is designed similarly to the company's regular Pulse RX 65...
Intel Sapphire Rapids Workstation Specifications: Up to 56 Cores, 350W TDP - 07/25/2022 08:30 AM
VideoCardz has provided the Sapphire Rapids-WS CPU specs. These core-heavy chips target workstations, not data centers, unlike Sapphire Rapids-SP....
SAPPHIRE has produced a non-XT Radeon RX 6700 graphics card. - 07/13/2022 09:28 AM
The Radeon RX 6700 is a GPU that sits below the Radeon RX 6750 XT / 6700 XT. It has 2,560 2,304 stream processors, an Infinity cache of 96MB 80MB, video memory of GDDR6 12GB GDDR6 10GB, and a memory b...
cucaulay malkin
Senior Member
Posts: 7389
Joined: 2020-08-03
Senior Member
Posts: 7389
Joined: 2020-08-03
#6079178 Posted on: 12/03/2022 09:28 PM
nvidia's mcm is different, can't remember the name for it but they're chiplets designed so that each one can have a different function.
edit: it's called composable on-package architecture (COPA), basically allowing them to make many different variations.
nvidia's mcm is different, can't remember the name for it but they're chiplets designed so that each one can have a different function.
edit: it's called composable on-package architecture (COPA), basically allowing them to make many different variations.
Our analysis shows that DL-optimized COPA-GPUs will provide impressive per-GPU training and inference performance improvements, while still efficiently supporting scaled-down HPC-targeted designs. DL-optimized COPA-GPUs will also result in reduced datacenter cost by minimizing the number of GPUs required to achieve scale-out training performance targets, making COPA-GPU an attractive paradigm for increasing individual and aggregate GPU performance without over-optimizing the product for any specific domain.
tunejunky
Senior Member
Posts: 3254
Joined: 2017-08-18
Senior Member
Posts: 3254
Joined: 2017-08-18
#6079180 Posted on: 12/03/2022 09:32 PM
the efficiency was a pleasant surprise to some but many had expected it w/ the node shrink
AMD wasn't any more worried about the 4090 than the 1080 when they released the rx5800.
not only was AMD not targeting that slice of heaven or hell - the yield is unacceptable to AMD.
they don't have the bucks to waste 20-30% of each wafer on a limited fab run.
just being honest. plus the fact that this is something Nvidia does every generation, the difference being the post-crypto price. the Nvidia Halo production is literally paid for by Marketing every year. the number of units sold has never justified the cost of production alone. but Nvidia is almost as good at marketing as Intel (the king).
I think they got subdued because the RTX 4000 series ended up being a lot more efficient than most had expected. AMD knew they weren't going to beat the 4090 in performance and they most likely know they're still going to under-perform in raytracing, but they could eat least compete against the 4080 in terms of performance-per- and performance-per-watt. Well, the value proposition doesn't really matter when you're still paying 4 figures, and AMD was most likely targeting Nvidia's overestimated wattage. So, now they got a product that's not quite as special as they originally anticipated. I think it'll still be good, and I'm still feeling pretty confident I'll be getting a 7700 if the price is right, but this is likely another round where Nvidia is the overall winner.
the efficiency was a pleasant surprise to some but many had expected it w/ the node shrink
AMD wasn't any more worried about the 4090 than the 1080 when they released the rx5800.
not only was AMD not targeting that slice of heaven or hell - the yield is unacceptable to AMD.
they don't have the bucks to waste 20-30% of each wafer on a limited fab run.
just being honest. plus the fact that this is something Nvidia does every generation, the difference being the post-crypto price. the Nvidia Halo production is literally paid for by Marketing every year. the number of units sold has never justified the cost of production alone. but Nvidia is almost as good at marketing as Intel (the king).
user1
Senior Member
Posts: 2424
Joined: 2016-01-29
Senior Member
Posts: 2424
Joined: 2016-01-29
#6079181 Posted on: 12/03/2022 09:37 PM
I think the main reason nvidia hasn't gone mcm, is because they can sell gpus at 20k+ a pop ,which more than makes up for the increased chip cost and low yields , so instead they pursue larger reticle sizes for maximum performance, with the likes of volta and hopper.
I dont think we'll see an true chiplet design from nvidia until it actually provides a performance advantage over max die size + nvlinking multiple cards or perhaps the manufacturing is able to keep up with demand for lower end cards.
I think the main reason nvidia hasn't gone mcm, is because they can sell gpus at 20k+ a pop ,which more than makes up for the increased chip cost and low yields , so instead they pursue larger reticle sizes for maximum performance, with the likes of volta and hopper.
I dont think we'll see an true chiplet design from nvidia until it actually provides a performance advantage over max die size + nvlinking multiple cards or perhaps the manufacturing is able to keep up with demand for lower end cards.
cucaulay malkin
Senior Member
Posts: 7389
Joined: 2020-08-03
Senior Member
Posts: 7389
Joined: 2020-08-03
#6079182 Posted on: 12/03/2022 09:39 PM
AMD wasn't any more worried about the 4090 than the 1080 when they released the rx5800.
what ?
not only was AMD not targeting that slice of heaven or hell - the yield is unacceptable to AMD.
they don't have the bucks to waste 20-30% of each wafer on a limited fab run.
just being honest. plus the fact that this is something Nvidia does every generation, the difference being the post-crypto price. the Nvidia Halo production is literally paid for by Marketing every year. the number of units sold has never justified the cost of production alone. but Nvidia is almost as good at marketing as Intel (the king).
a 60% functional ad102 is still 15% more powerful than 4080 - amd's target for 7900xtx.
that is really impressive. come 2024, they'll be selling 4-5 skus on ad102 (even 4090 is heavily cut down), both g6 and g6x.
AMD wasn't any more worried about the 4090 than the 1080 when they released the rx5800.
what ?
not only was AMD not targeting that slice of heaven or hell - the yield is unacceptable to AMD.
they don't have the bucks to waste 20-30% of each wafer on a limited fab run.
just being honest. plus the fact that this is something Nvidia does every generation, the difference being the post-crypto price. the Nvidia Halo production is literally paid for by Marketing every year. the number of units sold has never justified the cost of production alone. but Nvidia is almost as good at marketing as Intel (the king).
a 60% functional ad102 is still 15% more powerful than 4080 - amd's target for 7900xtx.
that is really impressive. come 2024, they'll be selling 4-5 skus on ad102 (even 4090 is heavily cut down), both g6 and g6x.
Click here to post a comment for this news story on the message forum.
Senior Member
Posts: 3254
Joined: 2017-08-18
To be fair chiplets design is an old story with IBM being the first who played with it - 3081 mainframe. Nvidia played with chiplets at least the same as AMD if not more because they have a problem: their AI engine doesn't do well with chiplets and Nvidia wants to keep their tensor cores because unlike AMD they built an entire universe of applications based on it: neural networks, AI "smart" cars, etc. AMD has nothing to offer in this field so for them the transition was easy. I use a 3090Ti for deep learning and it is amazing: my own in house learning network! - after 3 months training it has a precision of 68% in image recognition and this with just a consumer card! I also use CUDA accelerated applications at work - AMD has nothing in this field - they didn't had to think about this. I like the fact that Nvidia was supportive of science (and of course their profit) - there is more to GPU than games - AMD failed to recognize that.
An very interesting scientific article published by some universities in collaboration with Nvidia in 2017: https://research.nvidia.com/sites/default/files/publications/ISCA-2017-MCMGPU.pdf
to be fair, an excellent post