LukeTbk's latest activity

A rubin GB300 will take way less time to run things, for a lot of use case say like an LLM that actual data send to it and from it is so small relative to the compute that this part is irrelevant. Of course Rubin-Groq sram system will be way...

An article with a title ram price are dropping dated from this spring, not sure if that reliable in hindsight. that on a site called peq42.com with an anonymous writter, that using as a source a yahoo page that I am not sure exist ...

this is such an hyperbolic and ridiculous way to express something, everyone of course agree because it is trivial: “You want maximum utilization in the cloud if you’re trying to deal with this very constrained environment,” he said. “And...

we will have for people to look at them with scan or what not to make sure they achieved full real 3D or 2.5D (cache side-by-side instead of below), but in both scenario it is not over the compute and it let them run those crazy power level.

good news, bad news, that mean that Intel way to put cache the 3d cache below the cpu instead of above with their really good Foveros packaging, so that the extra cache do not hurt cooling capacity do seem to work has intented in that regard. In...

Lot of peoples that never launched-operated-owned company specialized stuff in high altitude/space with the modern cost of it and never made datacenter or made the chips that goes in them, will tell an entity that did a lot of thoses things...

That not really a counter point, that making the same point, that maximising cloud providing and usage if it work well. (every user of sogni, is using the cloud, every provider of sogni making one), that simply distributed cloud marketplace, not...

Yes vendor/traditional enterprise, anything turnkey, anything people can buy, hyperscaler level of internal usage is quite different and usually way more custom than that, Amazon-Google-Meta run their own stuff on their own silicon in a way we...

A bit of the reverse, the ai popularity is the reason Google search is not growing fast, which was always the expected and why google has been cheap the last 3 years. People talkign to an AI instead of spending time in a browser-google search...

what he said is quite trivial, it would be nice if people building local AI turn around and share it to others people during their downtime.... (i.e. the grid cloud) for a better use rate of limited dram supply, it would also help to have a lot...

Get A New Beginning: Final Cut free from GOG till 11-02-26. https://www.gog.com/en/game/a_new_beginning Bioengineer Bent Svensson had to resign from researching alternative ways of generating energy. Now, the former workaholic lives a secluded...

Could have impacted around 6.3 billions and a feat from North korean hacker (who are really among the best in the world historically) https://www.cbc.ca/news/business/binance-north-korea-9.7402446

Only in a sense for the giant customer. An openAI and others make their own language-compilers, generate .ptx themselve and let nvidia cuda-stack/compiler optmise from there, for them using intel-amd the second they would be cheaper-better would...

It seem to me we have no confirmation about anything here, Nova lake was always meant to have some tile on some sku being on Intel 18A-p node... and the only thing said in the transcript of the investor call was: Nova Lake, our first Intel 18A-P...

This has been a talking point since the 2021 crypto boom (those less known-glamourous part of the supply chain were a major reasons for some of the covid era shortage, in cars and elsewhere)...

Raw result would be meaningless, it all depend on the game setting anyway, 70 fps mean nothing by itself, 3.2 frame per watt even less. The importance of having actual power-and frame in the graph is rooted in the fact that frame per watt is not...

ALmost always either upsell or forced by the competition, they have a long history of protecting parts of their line up, from no FM radio for the itunes segment, flash storage for the memory upsell, from waiting years for the iPad Mini, years for...

who bought the 162 millions tablets sold in 2025 in the world if not people ? and what would that mean about people wanting PS5 that sold less units since launch.... It is an hard segment, because of what people do with them usually mean they do...

yes I think, by wanting a frame per watt graph, they do not mean a graph that only show frames per watt, it would not be more telling for the users at all and would have 100% of the same problems, but those graph we sometime see where you have...

while first run, Apple has quite the very long experience in oled touchscreen here.... MacOS could have some legacy issue here of course, but the updates version will be available to this device if they ever figure it out.