The countdown to AMD's upcoming next generation Zen 2 & Navi announcement on the 27th of May is quickly ticking down. We're now less than four days away and the leaks just keep on coming.
Today's leak pertains to the company's upcoming mid-range Navi GPU, which some sources pin as Navi 10 whilst others claim is actually called Navi 12.
Regardless, this new chip is expected to power the company's upcoming mid-range Radeon graphics cards which are set to compete with NVIDIA's GeForce RTX 2060 and GeForce RTX 2070. Even though the company's next generation nomenclature is still unconfirmed, for the purposes of brevity and simplicity we're going to label this chip Navi 10 and the highest specced graphics card it powers the RX 3080.
AMD Radeon Navi RX 3080 Specs Leaked - 8 Shader Engines, 2560 SPs, 8GB vRAM & 256-bit Bus
Before we get into the juicy bits as it were, we should first point out that this leak is courtesy of venerable leaker KOMACHI, although his tweet has since been taken down we've managed to grab a hold of all the tasty bits. Before we proceed it goes without saying that you should definitely take this upcoming information with a grain of salt, as always nothing is confirmed until we actually have the hardware in our hands.
So without any further ado, let's dig in.
Navi Allegedly Brings New Shader Engine Design For Better Compute Unit Utilization Efficiency & Higher Pixel Throughput
According to this leak, as well as our own sources, the Navi chip features 40 compute units, each housing 64 stream processors. The new previously unknown detail that this leak brings to the table is that Navi allegedly features 8 shader engines. Each of these engines will house 5 compute units and a raster back-end. Normally, in existing GCN GPUs, a raster back-end would include 16 Render Output Units, or ROPs for short. So it's unclear if Navi 10 will feature 128 ROPs or if the company is cutting down the number of ROPs per render back-end with Navi.
Now, the number of shader engines in Navi is actually a significant detail, because one of the major architectural limitations of AMD's previous GCN implementations is the inability to scale the microarchitecture beyond 4 shader engines per die. This has led to underutilization of stream processors in the company's bigger GPUs such as Fiji and Vega and has long been a known limitation of GCN that the company was working to develop solutions to.
AMD Kicking Things Off With Mid-Size Navi, Bigger Chip to Come
If this leak is accurate, AMD's engineers have successfully overcome this limitation and as a result we should see better performance scaling with the company's upcoming big chips compared to its previous designs. This indicates that we could very well see new Radeon GPUs with an excess of 4096 stream processors, which has been the maximum we've seen since the introduction of Fiji four years ago. In theory, the expansion of shader engines would also allow the company to add more than 64 ROPs to its future designs which would result in significantly improved pixel throughput, another limitation of its previous big chips.
Now, moving on to the memory subsystem. Thanks to previously leaked Navi PCB shots we know that the GPU features a 256-bit memory interface, which would normally be paired with 8GB of vRAM and up to 16GB if the company wishes to, perhaps for the professional versions.
More recently, we've learned that the company is looking to initially introduce at least two new Navi based graphics cards to directly compete with NVIDIA's GeForce RTX 2070 and RTX 2060, at $499 and $399 respectively. You can read all about the details of those cards here. In the meantime, stay tuned for more news and leaks as we get closer to May 27.
AMD Radeon Navi RX 5700 Leaked Specs & Rumored Pricing
|RX 5700 XT||Navi XT||2560||256-bit||8 GB GDDR6||RTX 2070||$499 US|
|RX 5700 Pro||Navi Pro||TBA||TBA||TBA||RTX 2060||$399 US|