Array
(
[content] =>
[params] => Array
(
[0] => /forum/whats-new/latest-activity?before_id=109304
)
[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] => 2030970
[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 .
F
32 trays per rack, 128GB total; 8 chips per tray, 4GB total; 500MB per chip
Wednesday at 5:56 AM
F
AI6 and Groq are big chips that require a fab operate in unprecedented ways. It blurs the line between R&D and production.
Wednesday at 5:55 AM
P
As electronics is increasingly more talked within the military action context, I feel I have to write that PRC defence industry is far...
Wednesday at 5:49 AM
From read Ian Cuttess - smart guy - Rubin CPX is for the prefill phase -- and the LPU is for the decode phase. Read his article on it...
Wednesday at 5:14 AM
S
From read Ian Cuttess - smart guy - Rubin CPX is for the prefill phase -- and the LPU is for the decode phase. Read his article on it...
Wednesday at 5:08 AM
Maybe my Gemini agent should write the AI-pie strategy book, every second a 1000 tokens :ROFLMAO: .
Or perhaps our host DAN can...
Wednesday at 5:04 AM
Curiously, NVIDIA has not disclosed what the host CPU is at this time, though they have disclosed that it will have (up to) 128GB of...
Wednesday at 4:36 AM
F
32 trays per rack, 128GB total; 8 chips per tray, 4GB total; 500MB per chip
Wednesday at 4:35 AM
This article points out that the Groq solution appears to have displaced Rubin CPX (announced 2025) from the hardware stack...
Wednesday at 4:28 AM
This article points out that the Groq solution appears to have displaced Rubin CPX (announced 2025) from the hardware stack...
Wednesday at 4:23 AM
B
AI6 and Groq are big chips that require a fab operate in unprecedented ways. It blurs the line between R&D and production.
Wednesday at 4:05 AM
Have you seen this paper?
https://cdn.sanity.io/files/chol0sk5/production/4e3e6966d98da9ef4bfd9834dbfc00921da58252.pdf
I have no doubt...
Wednesday at 4:03 AM
Sadly, considering the ongoing chaos in the Middle East, this $1.5 billion Saudi Arabia deal along with many other AI and datacenter...
Wednesday at 3:59 AM
Groq already lost confidence in themselves last year...
Wednesday at 3:59 AM
Samsung Foundry Nabs Nvidia? :ROFLMAO:
Groq started with GF 14 which really is Samsung 14 so it tracks they would use Samsung 4nm...
Wednesday at 3:57 AM
Like.
32 trays per rack, 128GB total; 8 chips per tray, 4GB total; 500MB per chip
user nl reacted to Sølve Folkestad Dahl's post in the thread Inside NVIDIA Groq 3 LPX: The Low-Latency Inference Accelerator for the NVIDIA Vera Rubin Platform with
Xebec replied to the thread Inside NVIDIA Groq 3 LPX: The Low-Latency Inference Accelerator for the NVIDIA Vera Rubin Platform.This article points out that the Groq solution appears to have displaced Rubin CPX (announced 2025) from the hardware stack...