The Radeon RX 6850 XT may never come out

For some time now, several well-known leakers, such as Greymon55 and Moore's Law Is Dead, had hinted that a Radeon RX 6850 XT would arrive over the next few months, particularly referring to the second quarter of this year, or at the latest in July. However, it appears that this will not be the case. In fact, MLID has stated that no RX 6850 XT is planned and even coreteks has never heard of this specific model.

Photo Credit: AMD Anyway, the arrival of "refresh" versions of the cards current videos now seem established, even if now we talk more specifically about RX 6950 XT, RX 6750 XT and RX 6650 XT. A little over a week ago, we reported that the alleged RX 6950 XT should offer a boost clock of over 2.5GHz, a whopping 250MHz more than the current RX 6900 XT, which will push TBP (Typical Board Power) up to 350W. Recall that these refresh models should be equipped with faster memory at 18Gb / s.

I can now confirm that #AMD does NOT plan to release an "RX 6850 XT". Also:

1) 6950 XT is on track!

2) 6750 XT & 6650 XT are CURRENTLY planned for Q2

I briefly talk about it on Broken Silicon tomorrow, and I actually suggested most of this last week: https://t.co/GPPkxcov8t

- Moore's Law Is Dead (@mooreslawisdead) February 15, 2022



if (jQuery ("# ​​crm_srl-th_hardware_d_mh2_1"). is (": visible")) {console.log ("Edinet ADV adding zone: tag crm_srl-th_hardware_d_mh2_1 slot id: th_hardware_d_mh2"); } The shortage of semiconductors that hit the market hard has certainly not spared AMD which, recently, has re-supplied "old" Polaris 12 GPUs to some Japanese manufacturers, so much so that new Radeon RX 550s are now on the market, sold at a price of about 137 euros. For more information on this, we recommend that you read our previous article.

On the processor front, AMD recently announced that its third generation EPYC processors will be used to expand the Google Cloud offering with new instances Google Cloud C2D, which will be added to those Google Cloud T2D. This will allow Google to deliver high performance for memory-bound HPC workloads, such as electronic design automation (EDA) and computational fluid dynamics (CFD).