Small $ 60 camera for #CV tasks

JeVois — video sensor + quad-core CPU + USB video + serial port, all in a tiny, self-contained package size of a small coin.


YouTube: https://www.youtube.com/watch?v=7cTfOckkGlE
Project site: http://jevois.org
Link for shop: https://www.jevoisinc.com


#deeplearning #hardware #tensorflow

The current state of "DIY" ML hardware

(i.e. that you can actually assemble and maintain and use in a small team)

Wanted to write a large post, but decided to just a TLDR.
In case you need a super-computer / cluster / devbox with 4 - 16 GPUs.

The bad
- Nvidia DGX and similar - 3-5x overpriced (sic!)
- Cloud providers (Amazon) - 2-3x overpriced

The ugly
- Supermicro GPU server solutions. This server hardware is a bit overpriced, but its biggest problem is old processor sockets
- Custom shop buit machines (with water) - very nice, but (except for water) you just pay US$5 - 10 - 15k for work you can do yourself in one day
- 2 CPU professional level motherboards - very cool, but powerful Intel Xeons are also very overpriced

The good
- Powerful AMD processor with 12-32 cores + top tier motherboard. This will support 4 GPUs on x8 speed and have a 10 Gb/s ethernet port
- Just add more servers with 10 Gb/s connection and probably later connect them into a ring ... cheap / powerful / easy to maintain

More democratization soon?

Probably the following technologies will untie our hands

- Single slot GPUs - Zotac clearly thought about it, maybe it will become mainstream in the professional market
- PCIE 4.0 => enough speed for ML even on cheaper motherboards
- New motherboards for AMD processors => maybe more PCIE slots will become normal
- Intel optane persistent memory => slow and expensive now, maybe RAM / SSD will merge (imagine having 2 TB of cheap RAM on your box)

Good chat in ODS on same topic.

#hardware

​​Lectures on computer architecture

Videos and slides about computer architecture by Professor Onur Mutlu

Channel: https://www.youtube.com/channel/UCIwQ8uOeRFgOEvBLYc3kc3g/featured
Professor: https://people.inf.ethz.ch/omutlu/

#hardware #lectures

GPU cooling tool

This script lets you set a custom GPU fan curve on a headless Linux server.

If you want to install multiple GPUs in a single machine, you have to use blower-style GPUs else the hot exhaust builds up in your case. Blower-style GPUs can get very loud, so to avoid annoying customers nvidia artificially limits their fans to ~50% duty. At 50% duty and a heavy workload, blower-style GPUs will hot up to 85C or so and throttle themselves.

Now if you're on Windows nvidia happily lets you override that limit by setting a custom fan curve. If you're on Linux though you need to use nvidia-settings, which - as of Sept 2019 - requires a display attached to each GPU you want to set the fan for. This is a pain to set up, as is checking the GPU temp every few seconds and adjusting the fan speed.

This script does all that for you.


Code: https://github.com/andyljones/coolgpus

#hardware #gpu

Trying Out New Ampere GPUs and MIG

Our hands on experience with new Ampere GPUs - 3090 and A100 (with multi instance GPU)

https://habr.com/ru/post/531436/

Please like / share / repost!

#hardware
#deep_learning

Practical ML Conf - The biggest offline ML conference of the year in Moscow.

- https://pmlconf.yandex.ru
- September 7, Moscow
- For speakers: offline
- For participants: offline and online (youtube)
- The conference language is Russian.

Call for propose is open https://pmlconf.yandex.ru/call_for_papers

#conference #nlp #cv #genAI #recsys #mlops #ecomm #hardware #research #offline #online