rokyed

Firstly, thanks Luke and Linus for sending me here. So, I've built a new setup: CPU: AMD EPYC 7401P MOBO: GIGABYTE MZ31-AR0 GPUS: 5 x Nvidia GTX 1080 RAM: 64GB DDR4 ECC (2333Mhz) STORAGE: 4 x 480GB Kingston SSD PSU: EVGA Supernova P2 80+ Platinum 1600W (the computer 100% theoretical load should be 1300-1400w [did not managed to test it, problem explained below]) UPS: CyberPower PR2200LCDRT2U ( I wouldn't run this computer without it, it costed too much to trust wall electricity) Software: Storage is in RAID 0 for speed. I'm using Proxmox (linux + qemu) in order to virtualize the computer. I'm using Ubuntu 18.04 Server (the one with terminal setup, not with graphical setup) + CUDA drivers in order to be able to run Tensorflow on it. Speed: Blender Rendering (one frame): Mac Book Pro (Retina, 15-inch, Mid 2014) 2.2Ghz: 11 minutes and 13 seconds Ubuntu(virtualized, 32cores): 2 minutes and 4 seconds Ubuntu(virtualized, cuda 4X gpu): 1 minute and 46 seconds Current problems: TLDR: I suck at setting the passthrough, since this is my first beast of a server, and had no experience with "board management control", virtualization, qemu and iommu groups, also i've learned how to properly set up a UPS with Proxmox so in case of power loss, the system will properly shut down. I'm able to virtualize 4 out of 5 GPUs because, I still didn't figure out how to properly set the virtual instance's pci bus, because activating the 5th gpu, makes the virtual network show as present but won't work. Past problems: Fitting 5 GPUs on that motherboard is not feasible, so the only way is with extensions. I've modified a Rosewill server case to make it be able to handle as many GPUs as the board can handle (logically) Anyway here's the juicy footage: I have a series made on youtube for it: https://www.youtube.com/watch?v=0S8pEEBmIes&list=PLOFkZ9MQr4DWjY4V3ertyAB7SKukZ6dEf (not trying to promo my channel, also no ads on my thing, just want to share the footage)

For the ones who want to know how this project progressed, here: Thanks for the suggestions.

Thanks but the budget is not a concern here, getting the best of the best would ensure that it would last longer (as technology and relevance) Thanks for the tips. Also now knowing that from 16x to 8x will impact the perfomance less than i expected, it's only natural to go with such a beast so in the future I could add more gpus ( like 3 more at least )

I am planning on using linux only on it. Also very interesting that fact. Thanks.

additional question: is https://www.pc-canada.com legit and reliable?

Hi guys, I hope this is the right way to express myself and find solutions to my problem. I will be building a new computer for machine learning ( i already have the GPUs (5 x GTX1080) that i got from a mining rig). My target was to get: AMD EPYC 7551P (32 core) and the motherboard from GIGABYTE MZ31-AR0 This proves to be problematic, because I can't find any seller in Montreal, Canada. Any kind of solution that can handle 5x 16x PCIe would work. I need all this raw power in order to iterate fast. Thanks.

seems cool, because you can have full expansion slot for the GPU, but you have to stick with the small form factor psu... I am also looking for a 2U but with horizontal full expansion slots, so i can put my gaming pc in it.

user case: sometimes, I want to listen again to what's said in the video, in other words, I want to go back 5-10 seconds. youtube's impementation: j = -10 seconds k = pause/play l = +10 seconds m = mute numbers 0 to 9 = start at (number * 10% from total length) maybe there are other that i don't know if for any reason you guys don't have time, I am willing to implement such myself because having such functionality allows me to control the video player with greater precision and speed than using the mouse. really thanks

do you guys need any help? PM me if yes

Hi Luke and Linus, not sure where to put this but here ye goes <canvas class="connecting-dots"/> not updating it's dimensions upon resizing Bug severity: Extremely low Steps to reproduce: 1. open browser 2. make window small 3. go on https://www.floatplane.com 4. make window big 5. watch and be amazed #MTL Btw Luke, awesome job on the responsiveness of the website, I haven't yet looked at how the server responds, but awesome job!!