Array
(
[content] =>
[params] => Array
(
[0] => /forum/whats-new/latest-activity?before_id=101916
)
[addOns] => Array
(
[DL6/MLTP] => 13
[Hampel/TimeZoneDebug] => 1000070
[SV/ChangePostDate] => 2010200
[SemiWiki/Newsletter] => 1000010
[SemiWiki/WPMenu] => 1000010
[SemiWiki/XPressExtend] => 1000010
[ThemeHouse/XLink] => 1000970
[ThemeHouse/XPress] => 1010570
[XF] => 2030770
[XFI] => 1060170
)
[wordpress] => /var/www/html
)
You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please,
join our community today !
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser .
Better power efficiency for the functions a TPU can do, which are more limited than a GPU. And it's my guess, based on skimpy available...
Nov 26, 2025
O
The Oregonian reported that Intel has comforted hiring Lo Wen-Jen:
In a message to employees Wednesday morning, Tan defended the...
Nov 26, 2025
K
I agree, TPU chips or modules alone are fairly useless without the entire pod and associated software stack.
Nov 26, 2025
T
Where's the original source for this 125 MTr/mm2 density claim ? Can't find it amongst the links listed.
Nov 26, 2025
K
I think "renting chips" just means Meta gets a special deal on using TPUs in Google Cloud, which Google already supports...
Nov 26, 2025
D
The 125 MTr/mm2 came from a Weibo post on March 24 this year by 定焦数码, later reposted on X the same day by @Jukanlosreve.
If the SMIC...
Nov 26, 2025
S
The 125 MTr/mm2 came from a Weibo post on March 24 this year by 定焦数码, later reposted on X the same day by @Jukanlosreve.
If the SMIC...
Nov 26, 2025
A
It's all about Parallel Computing - Compiler & APIs.
"I don’t know if you know about the egress cost that is moving data out of one...
Nov 26, 2025
All-hands meeting coinciding with GA of Ironwood release reveals extent of AI demand
Google will have to double its "serving capacity"...
Nov 26, 2025
All-hands meeting coinciding with GA of Ironwood release reveals extent of AI demand
Google will have to double its "serving capacity"...
Nov 26, 2025
N+3 is still a guess though. There could still be a rude shock.
Nov 26, 2025
The 125 MTr/mm2 came from a Weibo post on March 24 this year by 定焦数码, later reposted on X the same day by @Jukanlosreve.
If the SMIC...
Nov 26, 2025
Huawei's upcoming Mate 80 Pro Max has shown up on Geekbench with a new Kirin 9030 SoC. It is the first smartphone chip to be...
Nov 26, 2025
OMG! Us old engineers worship these senior executives like gods, but turn out, they behave not much different than our teenage kids...
Nov 26, 2025
that's why it's always good to have a healthy dose of anti-authority :) trick is the balance ...
Nov 26, 2025
Xebec reacted to blueone's post in the thread Meta in talks to spend billions on Google's chips, The Information reports with
Like.
Better power efficiency for the functions a TPU can do, which are more limited than a GPU. And it's my guess, based on skimpy available...
hist78 reacted to Daniel Nenni's post in the thread Google must double AI capacity every six months to meet demand - report with
Daniel Nenni posted the thread Google must double AI capacity every six months to meet demand - report in SemiWiki Main Forum ( Ask the Experts! ).All-hands meeting coinciding with GA of Ironwood release reveals extent of AI demand Google will have to double its "serving capacity"...
Haha.
that's why it's always good to have a healthy dose of anti-authority :) trick is the balance ...