Wikipedia is contradictory on AMD history

Wikipedia should not be used as an ultimate repository of factual knowledge. It is a repository of public knowledge. Anyone can get a wikipedia account and with that add/modify its articles. It can and does contain inaccurate, misleading and poorly worded information. That is also the #1 reason you should never use it as a source of 'proof' in a forum discussion ;)

In this case you linked to separate articles written on different dates that no doubt were written by different people who did not necessarily look into other articles to maintain its integrity.
 
They are two articles and I am well aware of its factuality -- I intend to fix one of the articles just not sure which one to fix. The question could be worded: was Bulldozer (and Zen, I guess) built from total scratch or are there significant blocks from Nx586 remaining?
 
I have no doubt the parts of the old NexGen IP is still used regardless of the 'from scratch' claims. To AMD, leveraging bits or ideas from that older tech would still be (to them) new ideas for using it. In short both can still be accurate.
 
As with everything, context. The term from scratch means it did not follow the previous design. It does not mean that everything that went into it was invented at the time.

The above is like saying I have to invent flour, to bake a cake from scratch.
 
The wording is not the problem it is that people tend to be stuck on understanding what is written.
 
The wording is not the problem it is that people tend to be stuck on understanding what is written.
Ensuring the intended interpretation of an article's meaning is not always a given. That's why there are such things as advanced university degrees in English literature.
 
If you want to give yourself a complex (pick any) you should steer away from correcting articles on any web based platform. There are just to many things that are not that worthwhile and other people can just troll you on purpose because they can :) not because you are right.
 
The NexGen CPU became the AMD K6. That was a long time ago! And the very next AMD CPU, the K7, was not based on K6 to a meaningful degree. Maybe some bits and pieces, but not the overall design.
 

Both articles are wrong.

Nx686 was used as baseline for K6 and derivations: K6-II, K6-III, K7,...

Bulldozer wasn't based in Nx686, but it wasn't a design from scratch either. Bulldozer (and its CMT concept) was derived from unfinished K11 prototype by Andy Glew, when he worked at AMD. His prototype was based in his MCMT approach, which he invented back at Wiscosin University

1_andy_mcmt.png


which itself borrowed ideas from early conjoined-core uarch introduced by Kumar et al at MICRO 37 symposium

1-s2.0-S0065245806690013-gr015.gif
 
Last edited:
Namecalling / Insults
Both articles are wrong.

Nx686 was used as baseline for K6 and derivations: K6-II, K6-III, K7,...

Bulldozer wasn't based in Nx686, but it wasn't a design from scratch either. Bulldozer was derived from unfinished K11 prototype by Andy Glewn, when he worked at AMD. This prototype was based in his MCMT approach, which he invented back at Wiscosin University

View attachment 93698

which itself borrowed ideas from early conjoined-core uarch introduced by Kumar et al at MICRO 37 symposium

View attachment 93699

And if someone would make posts about how many wrong things you have posted we be here for a few centuries at least ...
 
They are two articles and I am well aware of its factuality -- I intend to fix one of the articles just not sure which one to fix. The question could be worded: was Bulldozer (and Zen, I guess) built from total scratch or are there significant blocks from Nx586 remaining?
If you have never edited a Wikipedia article before, boy are you in for a treat. It's one of the biggest cesspools out there.
 
Back
Top