Practically since the spring, there have been reports about the release date of the Ryzen 9000X3D and the fact that Zen 5 something can change on the X3D models, in short, they could be more interesting to the user than Ryzen 5000X3D and Ryzen 7000X3D in some way.
As for the release date, there has been talk of alternately January (CES 2025) and October (answer to desktop processors Lake Arrow from Intel). Different sources said there were different possibilities and basically stood by them. Now we learn from leaker zhangzhonghao of the ChipHell discussion forum that the situation has been clarified at Gigabyte’s X870 motherboard release event, where AMD employees appeared as well as Gigabyte employees. There should have been a pretty clear indication that the octa-core Ryzen 7 9800X3D would be released in October in response to Arrow Lake, while the multi-chiplet models (twelve-core and six- Ryzen 9 9900X3D and Ryzen 9 9950X3D will arrive next year. The reason is that the models with more chiplets will be equipped with that technological innovation.
An estimate
The fact that the models with two CCDs (processor chiplets) will be different from the model with one CCD would be in line with our speculation in June, which arose in response to the official announcement that the X3D pro Zen 5 it will be “interesting, different and improved”.
| Placing processor chiplets on a silicon substrate comes to mind as another possibility [interposer]which would introduce SRAM. There would be no need to develop new chiplets there (chiplets equipped with an interface for a silicon pad AMD is preparing for the APU Strix Halo, which is currently being prepared for the beginning of 2025 and is responsible for the integration of final grade). cache), but it would increase the costs with the price of the plug. But the advantages would be obvious: (1) No lowering of the clock frequency, (2) cache available for both processor chiplets, (3) compared to the cache on the central chiplet, a much faster interface between the CPU cores and the cache of the last level). It is a de facto solution that was expected in the desktop until Zen 6, so it is doubtful whether AMD will want to use it like this (let’s say) ahead of time. However, it would definitely be interesting, different and improved. |
While for a solution with one CCD, the silicon substrate could unnecessarily increase the price of the processor, for a pair of CCDs (Ryzen 9) it would be easier to distribute in the total price of the product and would bring benefits special to him too. . In that case, a larger cache would be available with both CCDs, not just one, so that application workloads demanding data transfer or latency would benefit from it. At the same time, such a solution would make a more significant difference – Ryzen 9 X3D would differ from Ryzen 7 X3D not only in the number of cores, but also in support of higher clocks (even gaming performance could be higher).
It remains to be seen whether AMD will embark on such a bold solution Zen 5it is still a question. Technically, it should be solved at the beginning of 2025, because something similar is planned for the APU chiplet Strix Halo. Rather than a technical question, it is therefore a strategic decision – ie whether something like this should be launched on the market earlier than with a generation Zen 6. According to older reports, this should be the first to provide such a solution.
2024-09-25 22:08:09
#Ryzeny #9000X3D #released #AMD #October #January
