well, luckily for the US, we have massive hydro power dams in the west that are low because of 20 years of drought, so we can actually do pumped storage.
Ideally, we would pump salt water from oceans to the salt water reservoirs and use surplus energy to desalinate to refill the fresh water ones.
Better, just pump sea water to the desert, and let it spread out and evaporate. The mountains will catch the water vapor and fill your reservoirs.
Releasing in Death Valley, you don't even need to pump. A siphon once started will just run indefinitely. (The high point of the siphon would need to be less than 30 feet above sea level.)
You might need to bulldoze up the salt in the fall and take it somewhere.
My thoughts exactly. Solar -> Desalination, with all other energy consumption in the middle. Granted, I am not up to date on desalination. I know MIT released something recently but idk the details. Is energy the bottleneck now or is it still material?
scrolling through the news on different subreddits is quite fun and informative. Quick and easy way to get caught up on local and world news, sports, politics, etc. So all scrolling isn't bad
probably closer to 3500-4000 USD. Maybe $5k if theyre feeling greedy. But I think they want to undercut intel so theyll keep their margins reasonable so they can crush the Xeon line
I feel your comment is just pointlessly snarky without mentioning for what workloads it's not enough.
Nothing is ever enough but we'll always be limited by the economics of scale and in case you're in the exclusive demographic where this doesn't apply to you than you'd have resources for workarounds to performance limitations without complaining (server grade chips, FPGAs, supercomputers, dedicated ASICs, clusters, etc.).
Then you can buy an EPYC 7702P. It's not like they don't make 64 core CPUs. Or get a dual-socket board and buy two 7702s if 128 cores are more to your liking.
Try running kubernetes dev environment with all the crap your team has hoarded on it, elastic, Prometheus, 5 different database engines and dashboard, etc.
Then watch the system chug.
Funny story. The Vega64 has 8GB of RAM and 4MB of L2 cache (last-level cache) across 4096 SIMD-cores. That's 1kB of L2 cache and 2MB RAM per core.
It gets worse: although there are 4096 cores, the Vega64 isn't fully loaded until you stick 4-threads-per-core (Or more precisely, 16-wavefronts per Compute Unit (256-SIMD-cores)). That means you will actually need to run 16384 SIMD-threads before the Vega64 is fully utilized.
That's less than 512kB main-RAM per core, and less than 256 bytes of L2 cache per core. Better hope your threads are sharing a lot of memory...
--------
AMD addresses this problem with NAVI / RDNA: NAVI is fully utilized at 1-thread per core. So you only need 2816 threads on the 5700 XT to fully utilize the system.
> Federally funded research and development centers (FFRDCs) are public-private partnerships which conduct research for the United States Government. They are administered in accordance with U.S. Code of Federal Regulations, Title 48, Part 35, Section 35.017 by universities and corporations. There are currently 42 recognized FFRDCs that are sponsored by the U.S. government.
FFRDCs have a special legal ability to provide “neutral” input on technical matters. In theory Congress could hire Mitre or other FFRDCs to perform this service for them. That might take some tweaks to the law because FFRDC contracts seem to be aligned to specific executive agencies.
Personally, I’m a big believer in Federal in-sourcing combined with civil service reform. Using contractors for public interest positions, even contractors with special designations like FFRDCs, clouds the issue with contract management.
Of course, it’d be very difficult to design Civil Service reform properly and politically impossible to implement it, so creative contracting is probably the way to go for now.
Ideally, we would pump salt water from oceans to the salt water reservoirs and use surplus energy to desalinate to refill the fresh water ones.