Reports from sources in Asia indicate AMD has started sharing details with partners about the upcoming RDNA2 based gaming GPUs.
Unfortunately for us PC enthusiasts it seems the news aren’t great regarding the so-called “NVidia Killer”. The performance level of the high-end GPUs based on the codenamed Sienna Cichlid ASICs is around the same as NVidia’s last-gen top SKU, the 2080ti, according to these reports. A best case scenario supposedly shared by AMD is that at most gamers should expect to get about 15% over the 2080ti in a select AMD “optimized” titles. According to the same sources AMD is aiming to launch the “Big Navi” as a direct competitor to the soon to be launched RTX 3080, and not as a competitor to NVidia’s highest performing part as has been widely reported (let alone beat it). Some have suggested that “Big Navi” would be up to 50% faster in traditional rasterization than the RTX 2080ti but according to AMD’s partners that will not be the case.
The “Big Navi” reveal should happen sometime in early September with the actual launch happening on the 7th of October. Pricing for the “Big Navi” GPU will be determined by how much NVidia decides to sell the upcoming RTX 3080 for, with AMD looking to undercut said product.
According to the same sources there will be 2 SKUs based on Sienna Cichlid for the gaming market, one using a full die and another one cut-down, similar to what we saw with the 5700XT and 5700 GPUs from last year. Both will be using GDDR6 memory, with two other HBM based models coming to the Pro market based on the same micro-architecture (more on this below) in the future.
Navy Flounder is a mid-range GPU based on the same ASIC as Sienna and should come to market in Q1 of next year. It’s expected to be the “x7xx” class GPU, and target the mid-range, although details shared with partners on that part are scarce as of right now.
At the time of writing there was no information shared regarding real-time ray tracing capabilities, but a recently published patent suggests AMD might be using hardware acceleration for ray intersection, as seen in fig. 1
In addition to this information I should add that I speculated in a recent video that changes to the SDMA in RDNA2 pointed to the possibility of 3D stacked memory coming with Sienna Cichlid but with more time to look at the code in the public repositories and other sources it seems I was incorrect in that assessment. Although there have also been multiple patents published recently that describe various methods and inventions which suggest that there will be multiple 3D stacked devices coming to market from AMD in the coming months (as one would expect), it appears in the case of Sienna Cichlid the increased bandwidth will be used for chip-to-chip communication, either in a server environment or with a product similar to the Radeon Pro Vega II Duo which launched for Apple workstations last year. This would indicate that that product will not be an Apple exclusive. Based on the recent code commits, there seems to be the possibility of linking up to 2 Sienna based ASICs in this Pro context, for a total of 4 dies. The video memory will be unified/shared between all of them as well as with the host CPU.
Seeing that AMD is currently sharing information on these upcoming products with partners it’s likely we will be hearing more in the coming days/weeks. Follow me on twitter for updates @coreteks
Disclaimers: This article may contain amazon affiliate links. Information in the article is based on reports from industry sources and reported to the best of the author’s ability and in good faith but may contain inaccuracies or mis-translations and should thus be read as “rumor”. Note that there can be changes made to products that are reported on right up to launch day.