AMD Ryzen 3000 series CPU and X570 motherboard spotted in SiSoftware database

Shawn Knight

Posts: 15,280   +192
Staff member
In brief: The Ryzen 3000 series processor spied in SiSoftware's datbase was mated to an MSI MEG X570 Creation motherboard during testing. Specifications weren’t included although odds are, it’ll likely be MSI’s top-of-the-line board with all the bells and whistles when AMD’s Ryzen 3000 arrives later this year.

Another unreleased processor from AMD has found its way to the Internet courtesy of SiSoftware’s Official Live Ranker.

The desktop CPU in question, identified as 2DS104BBM4GH2_38/34_N, is said to be an engineering sample sporting a base frequency of 3.4GHz and a boost frequency of 3.8GHz. The codename further indicates this is a four-core part with 4 x 512KB of L2 cache, 16MB of L3 cache and a TDP of 65W.

Considering this is a quad-core part (and assuming it is legitimate), we’re likely looking at an entry-level Ryzen 3000 chip.

An Epyc Rome processor, AMD’s 64-core, 128-thread chip, was spotted in the database last week.

AMD previewed its third-generation Ryzen platform at CES 2019 in January. CEO Lisa Su showed in-hand an exposed CPU with a chiplet design featuring two dies. The eight-core chip AMD demonstrated outpaced Intel’s Core i9-9900K in Cinebench R15 and was more energy efficient.

Permalink to story.

 
The next comparable AMD result directly below that is also a 4 core 8 thread processor. Ryzen 2400G @ 4.2GHz.

Who knows what memory speeds the 2400G had compared to this, or other factors that can affect the result. The 3.8GHz AMD sample slightly bested the 2400G @ 4.2GHz on all the CPU tests.

Interesting that the same 2400G 4 cores and 8 threads needed a bit over 10 percent more clock speed to get slightly worse results however. That's the kind of IPC improvement that is expected and hoped from Zen 2. Maybe that is significant and representative. Maybe it isn't.
 
Last edited:
The next comparable AMD result directly below that is also a 4 core 8 thread processor. Ryzen 2400G @ 4.2GHz.

Who knows what memory speeds the 2400G had compared to this, or other factors that can affect the result. The 3.8GHz AMD sample slightly bested the 2400G @ 4.2GHz on all the CPU tests.

Interesting that the same 2400G 4 cores and 8 threads needed a bit over 10 percent more clock speed to get slightly worse results however. That's the kind of IPC improvement that is expected and hoped from Zen 2. Maybe that is significant and representative. Maybe it isn't.


Another thing to note is that this an engineering sample which means that it is not the finalized silicon. Could be an early sample or could be one that has been designed for maximum stability in testing. Typically, finalized silicon tends to perform far better than engineering samples.
 
Last edited:
The next comparable AMD result directly below that is also a 4 core 8 thread processor. Ryzen 2400G @ 4.2GHz.

Who knows what memory speeds the 2400G had compared to this, or other factors that can affect the result. The 3.8GHz AMD sample slightly bested the 2400G @ 4.2GHz on all the CPU tests.

Interesting that the same 2400G 4 cores and 8 threads needed a bit over 10 percent more clock speed to get slightly worse results however. That's the kind of IPC improvement that is expected and hoped from Zen 2. Maybe that is significant and representative. Maybe it isn't.
Speaks pretty highly of the 2400g imo:)
 
The next comparable AMD result directly below that is also a 4 core 8 thread processor. Ryzen 2400G @ 4.2GHz.

Who knows what memory speeds the 2400G had compared to this, or other factors that can affect the result. The 3.8GHz AMD sample slightly bested the 2400G @ 4.2GHz on all the CPU tests.

Interesting that the same 2400G 4 cores and 8 threads needed a bit over 10 percent more clock speed to get slightly worse results however. That's the kind of IPC improvement that is expected and hoped from Zen 2. Maybe that is significant and representative. Maybe it isn't.
Speaks pretty highly of the 2400g imo:)
When a processor with a higher clock speed and twice the core count is beaten by a low end engineering sample?
 
Back