Editing Talk:amd/athlon/3000g

{{talk header}}

Tried to edit this one repeatedly. The name 3000G implies a 12nm Picasso chip, but in reality it is a 14nm Raven Ridge chip. See the AMD spec for this: https://www.amd.com/en/products/apu/amd-athlon-3000g

It says it right there, it's 14nm so it can only be Raven Ridge. <small><span class="autosigned">—&nbsp;Preceding [[WikiChip:Signatures|unsigned]] comment added by [[User:131.174.168.130|131.174.168.130]] ([[User talk:131.174.168.130|talk]] • [[Special:Contributions/131.174.168.130|contribs]]) </span></small><!-- Template:Unsigned -->


:That's wrong, and the AMD website is wrong. The 3000G is 12nm Picasso and uses B1 stepping, not 14nm Raven Ridge (B0 stepping). --[[User:David|David]] ([[User talk:David|talk]]) 12:58, 27 November 2019 (EST)


::Allright, thanks, I didn't know that. But, are we sure? I mean, I believe you because you're probably quite knowledgeable, but the B1 vs B0 stepping, isn't that just a character string that's embedded in the processor that AMD can change into whatever they like at production? I mean, either the stepping is wrong or doesn't tell us what it should, or it's the 14nm mentioned on the AMD website that is wrong. Or both, because there are pictures of a delidded 3000G, and it looks to me like a much smaller die than Raven Ridge or Picasso. See here: https://mobile.twitter.com/momomo_us/status/1199354469227487232
::Many thanks for a great website btw! [[Special:Contributions/131.174.168.130|131.174.168.130]] 08:43, 28 November 2019 (EST)


Wish we had a decent die shot though, Raven2 is still a bit of a mystery to me. Funny processor to toy around with though. [[Special:Contributions/131.174.168.130|131.174.168.130]] 07:59, 23 January 2020 (EST)

The more I play with this chip, the more I think it's beefed up Dali. Specs are almost identical. The imc will do 3733MHz stable, 3866MGHz tops. The GPU will do 1700MHz stable, 1800MHz max (what I saw at least). That's both way over Raven Ridge. And way over Zen 1. [[Special:Contributions/131.174.168.130|131.174.168.130]] 04:27, 11 February 2020 (EST)

It could also be Banded Kestrel, which Computerbase seems to think it is: https://www.computerbase.de/2019-11/amd-athlon-3000g-test/ [[Special:Contributions/131.174.168.130|131.174.168.130]] 09:35, 11 February 2020 (EST)

My 3000G is certainly a 12 nm Picasso part. This is confirmed by the CPUID of 810F81h, and my overclocking attempts hitting 4.1 GHz (This is an extreme stretch for Raven Ridge, but common for Picasso.) --[[Special:Contributions/24.55.221.133|24.55.221.133]] 09:20, 1 June 2020 (EDT)
::I've done some more digging, I'm starting to think the silicon is Dali/Banded Kestrel/Raven2. Delidded 3000G's have a smaller die than Picasso parts, approximately the same size as Dali/Banded Kestrel (Which are 2 Zen cores w Vega 3 graphics). The 3000G would then be the fully unlocked version of this silicon. It would also explain it's lower price point, having much less silicon. Dali parts share the same CPUID as the 3000G and Picasso parts, so it isn't a reliable indicator of Zen+. --[[User:Svmlegacy|Svmlegacy]] ([[User talk:Svmlegacy|talk]]) 14:16, 12 June 2020 (EDT)

Above being said, it's clear that the 3000G doesn't use the standard Picasso silicon, based on delidded pictures. The die shots should be removed from this particular part, until we get die shots of a 3000G (And potentially Dali/Banded Kestrel) to confirm. --[[User:Svmlegacy|Svmlegacy]] ([[User talk:Svmlegacy|talk]]) 14:16, 12 June 2020 (EDT)

Die shots available, it's Dali -> https://www.hardwareluxx.de/index.php/news/hardware/prozessoren/53852-dali-die-shots-amd-athlon-3000g-in-allen-details.html