For anyone that isn’t following the minutiae of desktop computing, one of the more prominent rumors in AMD land is the presence of upcoming sockets called TRX80 and WRX80. These parts, assumed to be aimed at ‘casual’ and ‘workstation’ users, have formed the basis of many rumors and a lot of speculation, especially on the back of TRX40 which does exist for the latest generation of AMD Threadripper CPUs. Similarly, recent news has been posted about the potential existence of a theoretical LGA1159 socket for Intel, despite the fact that LGA1200 is often cited as the next generation socket for upcoming Comet Lake processors on the desktop. We reached out to a few contacts while we were here at CES to put some weight into these rumors, to confirm them one way or the other, for good.

It’s at this point I want to talk about sources. For anyone who has studied journalism, or even history at a high school level, there is the concept of sources carrying different amounts of weight. A primary source, for example, would be a direct participant in an event who would be able to recount with specific detail what they did/saw or what is planned. A secondary source might be a journalist collating or analyzing data from primary sources, a piece of software including a list of details about other products, or an artist depicting the events taking place from information they have collected. Tertiary sources go beyond this, and might involve individuals discussing an event without any direct experience of a specific instance. When it comes to the journalism that the tech press practices, depending on the publication, different numbers and different levels of sources form the minimum requirements for each publication. This acts as the minimum barrier required in order to present the information to its readers. 

Here at AnandTech, we often use at least two sources for any particular story, preferably both primary sources. If the second source is trusted, for example a fellow media peer in the industry that has a history of accuracy and one that we trust, they can be treated similarly to a primary source, albeit technically a secondary source (we often try to find a separate primary source different to the one they have used). Sometimes we have a lower standard for entry for the sake of humor, such as our report into Intel’s mythical Core i9-9900KFC, which never saw the light of day but I enjoyed writing it in a very tongue-in-cheek style (and I hope that came across). So to put this in to context, if I write something based on primary sources, I nominally act as a second source for others that cite my work.

Back to this specific matter, I had noticed that any mention of TRX80, WRX80, or LGA1159, was written about by journalists with no primary sources about any of it. So in order to find out when these two AMD eight-channel chipsets were going to come out, as well as the existence of yet another consumer Intel socket, I naturally went to ask AMD, Intel, as well as their primary partners and OEMs. The partners and OEMs that work with Intel and AMD are involved in developing and selling future platforms, and as such they have insight into the next 6-9 months of products coming into the market. They also have ideas about the 9-18 month schedule, but as we’ve seen in the past, that is often subject to change. Obviously none of the companies we spoke to were prepared to speak about future products, such as their own LGA1200 motherboards, however that doesn’t apply here.

All of our primary sources in this regard had very puzzled looks on their faces when I mentioned either the TRX80, WRX80, or LGA1159. One of them looked at me in amazement, and specifically said ‘what are you on about?’. I explained the situation, with ‘the internet’ talking about new 8-channel consumer motherboards, or an updated socket for Intel’s Comet Lake and beyond. The answer I got was very clear cut from everyone I talked to: no-where on their roadmaps has it ever said TRX80, WRX80, or LGA1159. None of the companies I talked had even heard of these names, let alone had any products with these features in the pipeline.

On the LGA1159 socket, one source actually knew what I was talking about, at least to some degree. They stated that the reason why people are confused here is that Intel distributed some early engineering samples of six-core Comet Lake CPUs in LGA1150 packaging, and those photographs got leaked onto the internet. The reason Intel could do this is that there's next to no difference between Coffee Lake Refresh and Comet Lake silicon, and as a result some early testing could be done. However, the same source also showed us the different packaging for the 10-core engineering sample, where the layout is somewhat different. As it stands, this is the issue with assuming what you see is always going to be at retail - if the fact that something is an engineering sample isn't taken into account, perhaps you shouldn't write about every rumor you see or hear.

But does this mean that there will never be a product called TRX80, WRX80, or have the LGA1159 socket? No. But what I am told from my sources is that as it stands these products have never existed. There might be a future LGA1159 socket for Intel CPUs, or future upgraded motherboards for AMD CPUs, but as it stands, all of my contacts confirmed that none of them exist on roadmaps today.

It should be pointed out that our sources did request anonymity. This is par for the course for this sort of response, even when confirming when a particular ‘product’ doesn’t exist. Should these sources be able to provide information about actual future products, and I may wish to call upon them again, and they'd rather not act as a magnet for future questions from all the other tech media.

POST A COMMENT

48 Comments

View All Comments

  • HeyYou,It'sMe - Monday, January 13, 2020 - link

    Apparently they didn't do their research. It's listed on the USB-IF standards committee. Seems like Anandtech is taking a swipe at Gamers Nexus, which was the original source. GN has a super high accuracy rate, has its own sources, and has the backing of USB-IF listings...a standards committee. So. Odd that they would criticize them like this, but not have the cojones to call out GN directly. This is like a subtweet article lol. Reply
  • Korguz - Monday, January 13, 2020 - link

    maybe posting a link to this would help ? Reply
  • DigitalFreak - Monday, January 13, 2020 - link

    That's what you get when you read WCCFTech Reply
  • phoenix_rizzen - Monday, January 13, 2020 - link

    With Ryzen desktop CPUs pushing 16 cores, and desktop APUs coming soon (possibly with 8 cores?), it would be nice to see support for quad-channel memory. Especially in the APUs.

    Was actually surprised the mobile APUs still only have dual-channel memory. Will be interesting to see some benchmarks of Ryzen Mobile 4000 systems in single-channel and dual-channel setups to see how performance changes; and to extrapolate from there to see how (if?) it would improve with the extra two channels.
    Reply
  • Alexvrb - Monday, January 13, 2020 - link

    The APUs are only 8 cores max, and most of the bandwidth is needed for the tiny iGPU anyway. Extra memory channels come at a cost, especially in terms of power, which means less watts for your CPU/GPU cores. Further if your memory isn't all soldered down (which I abhor) then it also means more slots and sticks to actually utilize. That makes it unpalatable, especially for the mobile space.

    With that said, Ryzen Mobile 4000 supports LPDDR4X up to 4266, which offers a TON more bandwidth. If manufacturers take advantage of it AND utilize a proper dual-channel configuration, they'll see a massive boost without the power consumption of an exotic-for-this-segment quad channel configuration.

    On the desktop, dual channel is still fine for the vast majority of cases. If you can afford their 16 core chip, you can afford high-end RAM. You've got a lot of options too for low latencies at 1:1 IF or much higher clocks but higher latencies at 2:1. If that's not good enough for your workload, they already HAVE a quad-channel HEDT platform. Anyway for regular consumers, dual channel DDR4 is good enough until DDR5 platforms arrive.
    Reply
  • nevcairiel - Tuesday, January 14, 2020 - link

    Additional memory channel require a lot more CPU pins and a lot of extra complexity in routing the memory from the socket to the DIMM modules.

    As unfortunate as it is, I doubt mainstream will move to quad-channel. DDR5 will almost double the bandwidth, so that'll probably hold us over for a while again.
    Reply
  • HeyYou,It'sMe - Monday, January 13, 2020 - link

    This all seems odd, as those chipsets were, and still are, listed on the USB-IF standards committee. Perhaps the author missed that bit. Companies don't submit things to standards bodies for...nothing. Reply
  • hammer256 - Tuesday, January 14, 2020 - link

    Is there a link? I can't seem to find anything on USB-IF's product search page. Reply
  • npz - Tuesday, January 14, 2020 - link

    This was the original search results page that started the rumor:
    https://cdn.mos.cms.futurecdn.net/nNUm7Ly8gcs77KYZ...

    Notice how TRX40 turned out to be real after being listed way back in 4/30/2019

    however, now the search results look different. "AMD Premium Chipset" doesn't show up, though make sure you click on the "Show Development Products" switch to the side
    https://www.usb.org/products

    What does appear for AMD in 8/12/2019 and 9/26/2019 are "AMD 2019-C3" and "AMD 2019-C4" and expanding the link (down arrow) to see what it covers only shows:
    TID: 2125
    Model Number: Superset

    TID: 1642
    Model Number: Superset, subset2
    Reply
  • npz - Tuesday, January 14, 2020 - link

    That said, regarding the Intel LGA1159 rumor, I believe Ian's explanation from that source stating how it stemmed from a confusion of a very early engineering sample. Reply

Log in

Don't have an account? Sign up now