Monday, March 16th 2026
AMD "RDNA 5" to Heavily Boost Shader Performance in Games with New Dual-Issue Pipeline
AMD is refining its RDNA 5 architecture, which is likely in its final stages of design. Thanks to a submission to the LLVM compiler, we are learning that AMD's upcoming RDNA 5/UDNA architecture features architectural changes that will enhance compute utilization, resulting in significantly higher game shader performance. The codename for RDNA 5 is GFX1310, and it now implements a full Dual-Issue VALU pipeline for Wave32. This allows vector operations (VOPs) to be issued simultaneously to the GPU's X and Y arithmetic logic unit (ALU) lanes. The new design expands the range of fused multiply-add (FMA) and other VOP instructions eligible for dual-issue and relaxes some register constraints, enabling compilers and shader code to perform more intensive floating-point work in Wave32 mode. This means the FP32 compute utilization of RDNA 5 can be much higher than before, greatly benefiting applications that are FP32-heavy, which means gamers are about to see a significant performance boost.
AMD originally implemented Dual Issue VALU in the RDNA 3 architecture with the Radeon RX 7000 series graphics cards. However, the pipeline was not fully functional, as the dual-issue implementation supported only a limited subset of VOP instructions, excluded several important FMA variants, required Wave32 mode and strict register-bank separation, and was often bypassed by compilers and not fully exposed to drivers. As a result, many shaders could not exploit X/Y pairing, and measured FP32 throughput often fell well below the hardware's theoretical peak. This was detrimental to performance and reduced compute capability, especially in applications that rely heavily on FP32 compute. This is particularly true for modern games that process vertex and pixel shaders primarily using FP32 compute. With the new RDNA 5 design, gaming performance should align more closely with theoretical compute performance.
While the RDNA 5 / UDNA specifications are still unknown, we are gradually piecing together the ISA design through compiler patches. AMD plans to release RDNA 5 possibly this year, featuring next-gen neural rendering, ML-based upscaling, multi-frame generation, and advanced ray tracing technologies taken to the extreme with the "FSR Diamond" project. We are eager to see what will come from this and, with the new instruction processing and design that AMD is working on, we are also excited to see the performance targets AMD will present.
Source:
Coelacanth’s Dream Blog
AMD originally implemented Dual Issue VALU in the RDNA 3 architecture with the Radeon RX 7000 series graphics cards. However, the pipeline was not fully functional, as the dual-issue implementation supported only a limited subset of VOP instructions, excluded several important FMA variants, required Wave32 mode and strict register-bank separation, and was often bypassed by compilers and not fully exposed to drivers. As a result, many shaders could not exploit X/Y pairing, and measured FP32 throughput often fell well below the hardware's theoretical peak. This was detrimental to performance and reduced compute capability, especially in applications that rely heavily on FP32 compute. This is particularly true for modern games that process vertex and pixel shaders primarily using FP32 compute. With the new RDNA 5 design, gaming performance should align more closely with theoretical compute performance.
40 Comments on AMD "RDNA 5" to Heavily Boost Shader Performance in Games with New Dual-Issue Pipeline
RDNA4 is really good hardware wise, they just really need to reinforce the software side of it. They need to make FSR4 work on INT8 so whiners can shut up already and this would give game devs additional reason to support FSR better in games. Also where is my ML enhanced AFMF 3 ?!?!?! Come on AMD. Also there is no way they are releasing RDNA5 in 2026. RDNA4 was released in 2025, exactly 1 year ago. Makes no sense from perspective of RDNA4 sales, all the RAM shortages and I just don't think they have anything ready in such short notice.
And yeah release INT8 FSR4 to do right by your customers and help entice prospective customers. I'll be the first to applaud you for doing it, and it'll make my incoming 6900XT all the more usable in modern titles.
Great (I mean it), but aren't these things considered heresy, fake frames, BS, etc, when Nvidia does them?
But I’m told AMD doesn’t gatekeep GPU features.
This idea that consumers are entitled if they don't just accept whatever hardware vendors shove down their throat needs to die.
If not, we'll probably end up with a repeat of RDNA 2 vs Ampere. AMD matches the 6090 on raster, but RT performance is 2 tiers worse and AI features continue to lag behind.
Sometimes it's better to just not support shit.
If Nvidia improves their shaders and RT cores, or significantly bumps up the RT core count (which I really hope they do) with Rubin, AMD will probably be significantly behind again.
And RDNA 3 was released in Dec 2002. And RDNA 5 is expected in 2H 2027 close to 5 years is a good time for it.