AMD has unveiled the model new Radeon RX 7900 XT and Radeon RX 7900 XTX, its newest high-end GPUs based mostly on the all-new RDNA 3 structure and manufactured utilizing a number of scalable ‘chiplets’ which have been fabricated on a 5nm course of. Each can be obtainable all over the world ranging from December 13. Whereas the flagship Radeon RX 7900 XTX is priced at $999 (roughly Rs. 81,875 earlier than taxes), the Radeon RX 7900 XT which is positioned one step beneath it should value $899 (roughly Rs. 73,680). Advised retail costs for India haven’t but been introduced. AMD claims as much as 70 % higher efficiency or 54 % increased efficiency per Watt than the earlier technology based mostly on the RDNA 2 structure.
Competing with Nvidia’s GeForce RTX 4000 collection, the brand new Radeon RTX 7900 XTX options 96 compute models, a clock velocity of as much as 2.3GHz, and 24GB of GDDR6 RAM. Complete board energy consumption is rated at 355W. The marginally extra inexpensive Radeon RX 7900 XT has 84 CUs, a 2GHz clock velocity, and 20GB of GDDR6 RAM. Energy consumption for this mannequin is 300W. AMD’s promotional photographs present two commonplace 8-pin PCIe energy connectors for each fashions, in addition to DisplayPort, HDMI, and USB Sort-C video outputs. Companion manufacturers will be capable to customise their choices.
Each of the Radeon RX 7000 collection fashions introduced to this point help the DisplayPort 2.1 commonplace for as much as 8K output at 165Hz of 4K at 480Hz. There’s additionally hardware-level help for as much as 8K 60fps AV1 encoding, plus simultaneous encode/decode with the AVC and HEVC codecs.
The basic monolithic GPU has been damaged up right into a Graphics Compute Die (GCD), consisting of the compute models that make up the graphics rendering pipeline, in addition to a number of Reminiscence Cache Dies (MCDs). The Radeon RX 7900 XTX can have six MCDs, every on a 64-bit large bus, amounting to 384 bits, whereas the Radeon RX 7900 XT has 5 energetic MCDs which quantities to a 320-bit bus. Whereas the GCD is manufactured on a 5nm node, the MCDs use a 6nm course of, permitting AMD to combine and match parts to optimise manufacturing sources and value.
AMD claims 50 % sooner ray tracing than with the earlier technology because of new AI directions and improved throughput. Compute models can now dual-issue directions, new AI accelerators have been added to every CU, and completely different phases of the graphics pipeline can now run at completely different clock speeds. Extra specifics about efficiency, together with aggressive positioning in opposition to Nvidia’s GeForce RTX 4000-series choices and the new Intel Arc GPUs, will doubtless be introduced at the moment.
The corporate additionally touts a 2.7X enchancment in bandwidth for its next-gen Infinity Cache implementation because of improved algorithms, in comparison with the earlier technology. The corporate additionally says it has caught with GDDR6 RAM within the curiosity of energy effectivity.