IBM does 60 deg C watercooling which can be not a lot of thermal delta in nonarctic environments. It’s a lot of km of infrastructure to vent directly if you want to dissipate a nuclear reactor’s worth of power in a single site.
On raw performance might, the M4 really does live up to Apple’s promises, should deliver. Single core is up about 20% compared to all M3 chips and more than 40% compared to M2. The generational computational leap from the previous M2 iPad Pro is at least a 42% jump on single-core and multi-core.
You seem to not be using open source software packaged for multiple architectures or which can be built for your binary target. Most people will be just using a browser and an office suite.
No, that captures just the neuroanatomy. Not the properties like density of ion channels, type, value of the synapse and all the things we don’t know yet.
“Of course, AMD is trying to get into the the AI training and inferencing game itself with the Instinct MI300 chip. And that, perhaps, is the main if modest cause for hope. If AMD can gain some traction in that huge market, it will not only be making lots of money, it will be in a position to do a similar thing to Nvidia and push some of that technology across into its gaming GPUs.”
which strikes me as incorrect. AMD MI is pretty widespread in HPC. With margins lower in the consumer market it makes sense to focus on HPC.
There are diminishing returns in semiconductor photolitho. Moore scaling is long over, absolute real estate see WSI with Cerebras, DC costs and power envelope are all sending a clear message. Quantization is there, so you can go from digital multipliers to analog and go spiking networks, but transformers and Co have little power there.
Also, the kind of economy that can carry Gen AI as business model is not a given, long term.
Yes orders of magnitude, but not too many of them. The real estate of a 300 mm wafer is limited, the structure shrink is saturating and you can’t get too many layers. You still need a packet switched network on the wafer even if the rest is mostly analog. Perhaps spintronics can limit the power requirements too.
A free running cellular automaton (CA) approach in hardware would work, but each cell would be a much souped up SRAM cell, the interactions would be all local and 2D. Considering Cerebras is 40 G SRAM on the 300 mm WSI and is about at the cooling limit I’m afraid you do not have 5 orders of magnitude. Perhaps reversible spintronics can help with the power draw, but you still have to splat a higher dimensional network so not just local interactions into a 2D array.
Hey, I’m always searching a home server (already post here lemmy.ml/post/15083947), I was thinking about a Lenovo P500, but maybe the PSU is a bit too special… My budget around $140, will buy used parts and one of the most important thing is the power efficiency (don’t care about the peak wattage but want to stay around...
I would look into thin clients and Lenovo etc. tiny PC for office on eBay. I run old low power low noise rackmount Supermicros which are nice but hard to find at low prices.
Global Financial Crisis 2 is definitely a fast component in the overall collapse process. Unaccounted consumer debt is a useful facet of a diagnostic. We don’t know where exactly the spark will originate that will blow up GFC2. It could as well start there, though it’s not very probable.
I think we are already at the catabolic collapse stage where the elites are cannibalizing the periphery before the heartland. On the other hand the US has been also being stripped bare in patches for a while. So it is a simultaneous progressive process.
The surging demand for data is guzzling Virginia’s water (grist.org)
iPad Pro with M4 chip boasts impressive performance jump compared to just-released M3 MacBook Air (9to5mac.com)
On raw performance might, the M4 really does live up to Apple’s promises, should deliver. Single core is up about 20% compared to all M3 chips and more than 40% compared to M2. The generational computational leap from the previous M2 iPad Pro is at least a 42% jump on single-core and multi-core.
Full scan of 1 cubic millimeter of brain tissue took 1.4 petabytes of data, equivalent to 14,000 4K movies — Google's AI experts assist researchers (www.tomshardware.com)
Paleocene/Eocene carbon feedbacks triggered by volcanic activity - Nature Communications (2021) (www.nature.com)
Abstract...
NVIDIA switching to open kernel modules by default in future driver update for Turing+ (www.gamingonlinux.com)
AMD's gaming graphics business looks like it's in terminal decline (www.pcgamer.com)
Flexible quasi-2D perovskite solar cells with high specific power and improved stability for energy-autonomous drones - Nature Energy (www.nature.com)
Abstract...
Toward a Cenozoic history of atmospheric CO2 (2023) (www.science.org)
Structured Abstract...
Has Generative AI Already Peaked? - Computerphile (www.youtube.com)
Record-breaking increase in CO2 levels in world’s atmosphere (www.theguardian.com)
Cubic millimetre of brain mapped in spectacular detail (www.nature.com)
Google scientists have modelled a fragment of the human brain at nanoscale resolution, revealing cells with previously undiscovered features.
Telegram apparently censor queer groups (mastodon.social)
cross-posted from: leminal.space/post/6433881
Need recommandations for a home server
Hey, I’m always searching a home server (already post here lemmy.ml/post/15083947), I was thinking about a Lenovo P500, but maybe the PSU is a bit too special… My budget around $140, will buy used parts and one of the most important thing is the power efficiency (don’t care about the peak wattage but want to stay around...
deleted_by_moderator
Americans Are Racking Up 'Phantom Debt' That Wall Street Can't Track (www.bloomberg.com)
cross-posted from: lemmy.world/post/15144297...
The Oil and Energy Macro | Art Berman (www.artberman.com)
Farmers warn food aisles will soon be empty because of crushing conditions: 'We are not in a good position' (www.yahoo.com)