4 ms·
I work on NLP projects and have similar issue with memory. The reason I don't use cloud computing is that you constantly have to worry about moving data around
by lqdc13 10y ago
I work on NLP projects and have similar issue with memory.
The reason I don't use cloud computing is that you constantly have to worry about moving data around, turning stuff on and off, API changes, data movement costs, product lineup changes (should I store this stuff on S3 or local HD), locally compiling all your stuff again or containerization and associated time overhead.
Local large box makes way more sense if you're doing solo projects. And if traveling, it would way more sense to rent a VPS from Hetzner. Flat cost structure and tons of bang for the buck.
BTW, I you can get 128 gb LGA2011-v3 motherboards.
https://www.newegg.com/Product/Product.aspx?Item=N82E16813157543 https://www.newegg.com/Product/Product.aspx?Item=N82E1681315...
and it seems like it works:
http://www.pcworld.com/article/2938855/hardcore-hardware-we-stuffed-this-pc-with-128gb-of-cutting-edge-ddr4-ram.html http://www.pcworld.com/article/2938855/hardcore-hardware-we-...
even though the spec says max 64gb:
https://ark.intel.com/products/82932/Intel-Core-i7-5820K-Processor-15M-Cache-up-to-3_60-GHz https://ark.intel.com/products/82932/Intel-Core-i7-5820K-Pro...
- dsacco 10y agoI can confirm, I run machine learning workloads on a home machine with 128GB of RAM on an X99-E WS/USB 3.1 motherboard. My manual explicitly states that it has 128GB RAM support, as well.