Rumor Mill Downgrades AMD’s RDNA3 GPU Specs

In what is certainly a first in recent times, new RDNA3 rumors indicate its specs might be a tad lower than we expected. Typically these leaks point in the opposite direction, or at least that’s been the case for Nvidia’s GPUs. On the AMD front we actually haven’t heard anything in almost six months. This new round of leaks are a revision of previous leaks. There’s also the possibility they are closer to reality than previous rumors. That’s just because we’re closer to launch now, so it’s plausible the new specs are what AMD has chosen as its final numbers. Its RDNA3 family of GPUs is expected to launch at the end of 2022.

The “revised” update for AMD’s upcoming GPUs comes from Twitter user Redfire via Videocardz. Overall the new numbers mark a reduction in important specs from previous leaks by 22 percent. The flagship Navi 31 card, aka the RX 7900 XT, was previously thought to have 60 Work Group Processors (WGP) with 15,360 stream processors (SP). That’s been downgraded to just 48 WGP and 12,228 SPs. This would afford the card up to 73 teraflops of FP32 performance if the GPU was clocked at 3GHz. This is an interesting development as it was recently reported both Nvidia and AMD were in a race to 100 teraflops. It’s expected that Nvidia’s AD102 die could hit the target first, but AMD might also be prepping a special card too, similar to Nvidia’s Titan GPUs of the past. For context the RTX 3090 Ti is marketed as offering 40 teraflops of performance, so these next-gen GPUs are going to be quite powerful.

The card below Navi 31 is reportedly going on a similar diet. The RX 7700 XT (Navi 32) was previously thought to have 40 WGP and 10,240 SPs, but is now rumored to offer 32 WGP and 8,192 SPs. The last card, Navi 33, remains unchanged. This was also confirmed by notorious GPU tipster Greymon55, who noted the card would “maintain original specifications.” That card has 16 WGP and 4,096 SPs. We still don’t have any details on memory allocations or clock speeds for these GPUs, unfortunately. Greymon55 also changed his Twitter profile’s bio to reflect these updated specs; a seemingly tacit approval of their veracity.

What’s interesting is Greymon55 also stated that despite AMD lowering the specs for this family of GPUs, its performance goals remain unchanged. In fact, they wrote, they may even be higher. Sadly, we have no idea what this means in the real world. Also it’s important to keep in mind that we do know AMD’s RDNA3 GPUs will be quite intricate. The top SKUs will feature a multi-chip module (MCM) design care of TSMC. These will reportedly be composed of seven chiplets, utilizing TSMC’s 5nm and 6nm processes. The flagship die is rumored to be approximately 800mm squared. For comparison, the RX 6900 XT die is just 520mm squared.

Continue reading

AMD’s New Radeon RX 6000 Series Is Optimized to Battle Ampere

AMD unveiled its RX 6000 series today. For the first time since it bought ATI in 2006, there will be some specific advantages to running AMD GPUs in AMD platforms.

Nvidia Will Mimic AMD’s Smart Access Memory on Ampere: Report

AMD's Smart Access Memory hasn't even shipped yet, but Nvidia claims it can duplicate the feature.

AMD’s Retail Market Share Surges Based on European Reseller Data

AMD's European channel sales figures continue to be excellent, with a particular surge in November and December 2020.

AMD’s Reliance on TSMC Isn’t Harming the Company’s Growth Prospects

It has been difficult to buy high-end PC components for nearly six months. There are a number of reasons for this, including pandemic-related impacts, the related surge in demand for all computing hardware, and supply shortages. A lot of eyeballs have been trained on foundries like TSMC, to the point that national governments have put…