3rd Gen AMD EPYC processors with AMD 3D V-Cache technology Released at 4200 to 8800 USD

Published by

Click here to post a comment for 3rd Gen AMD EPYC processors with AMD 3D V-Cache technology Released at 4200 to 8800 USD on our message forum
data/avatar/default/avatar22.webp
AMD STock - BUY BUY BUY!
https://forums.guru3d.com/data/avatars/m/266/266726.jpg
[youtube=ttvJzyNYEik]
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
Weird how the 16 core isn't the cheapest.
https://forums.guru3d.com/data/avatars/m/248/248291.jpg
This strategy of adding huge amounts of cache to everything, is paying off quite nicely for AMD. No wonder Nvidia and Intel are copying the strategy.
https://forums.guru3d.com/data/avatars/m/271/271560.jpg
well this pricing makes one thing clear - they're counting on volume to make up for the lower margin (higher cost of production) and they're counting on the 5800X3D to deliver that margin back, at least some of it. the price of the 5800X3D is exactly what i expected, given the production costs (brand new TSMC factory running ONLY AMD for the rest of this year), but Milan-X blows my mind by coming in 10 - 15% lower than my (and many others) expectations. go ahead print that money AMD & TSMC
https://forums.guru3d.com/data/avatars/m/271/271560.jpg
i have a very sneaky suspicion that Ryzen 4 will have at least three SKUs with 3D cache it will slay and they will also rule HEDT (even more) with TR with the 3D cache.
https://forums.guru3d.com/data/avatars/m/201/201426.jpg
waltc3:

Try this article on for size...;) https://www.phoronix.com/scan.php?page=article&item=amd-epyc-7773x-linux&num=1 AMD is making a habit of trouncing Intel these days...;)
How long did we wait for this? I know I have waited a long time to see AMD bring the heat!!!! Wish I could afford the threadripper version of these. Cause my 1920x is way slower then my 5600x in gaming. And Id loved to have a 3rd gen IPC Threadripper as a server, and game rig because I miss having HDET for my gaming rig. Last one I had was my X79 3930k setup. As I switched to 1500x/x370 in 2017. Been Ryzen since. My TR was never my main rig,
https://forums.guru3d.com/data/avatars/m/277/277212.jpg
This is terrific news to read. My project is ramping up production and we are using a box with two 16-core 7373s that just rocks! We have a couple of machines with 32-core Threadrippers and the EPYCs are considerably faster on my application. Their improved memory system and tons of cache really help. These machines with even more cache will really speed it up. I can't wait to try one of these. Unfortunately, lead times for EPYC machines are obnoxious right now so there's no telling when we might see one. I would be surprised if we could get one this year. It's great that AMD can sell practically everything they make but it's not so good when you are waiting.
https://forums.guru3d.com/data/avatars/m/268/268248.jpg
@Gomez Addams mind if I ask what workload you do ? Fluid dynamics ? Weather patterns including atmosphere composition etc etc ? I know such workloads will benefit from as much cache and ram you can feed em! That said I do not believe you owe me an answer I am just curious!
https://forums.guru3d.com/data/avatars/m/277/277212.jpg
It's actually a brute-force production optimization application that figures out the best way to process a product. The algorithms involve tons and tons of math on a fairly large amount of data but it's not too extreme. There are literally thousands of possibilities (sometimes in the hundreds of thousands) and it tries them all and then selects the one the provides the most value. The program scales to as many processor cores are as available. It works on the GPU too and runs pretty well on a 3090 with 10K cores. However, the memory use is too disparate for it to really shine on the GPU so it's actually faster on 64 EPYC cores than on 10K CUDA cores.
https://forums.guru3d.com/data/avatars/m/268/268248.jpg
Gomez Addams:

It's actually a brute-force production optimization application that figures out the best way to process a product. The algorithms involve tons and tons of math on a fairly large amount of data but it's not too extreme. There are literally thousands of possibilities (sometimes in the hundreds of thousands) and it tries them all and then selects the one the provides the most value. The program scales to as many processor cores are as available. It works on the GPU too and runs pretty well on a 3090 with 10K cores. However, the memory use is too disparate for it to really shine on the GPU so it's actually faster on 64 EPYC cores than on 10K CUDA cores.
I see so the more processing power and memory you can cram the more instances it can run each time ! Interesting and thanks for the answer !