The God of Computers

[SeanLMG]

 

In today's video two idiots play with over $60,000 of computer parts.

 

Check out the Comino Grando RM: https://lmg.gg/aA8Sl

[emosun]

that shock watch sticker was tripped on the crate

[carloskay]

Which are the others benchmarks that they are using in the video?? i saw cinebecnh but i dindt see the others

[AnonymousGuy]

Can we get some detail on who made the quick disconnects?  Im interested in buying ones used in proper servers. 

[T8z5h3]

1 hour ago, SeanLMG said:

 

In today's video two idiots play with over $60,000 of computer parts.

 

Check out the Comino Grando RM: https://lmg.gg/aA8Sl

Can I ask that the next time/any time  you guys look play with a setup of this class you run Nvidia-smi screenshot it and post it to the form thead of the release please... it would help give some context to the viewers what your running.

[T8z5h3]

58 minutes ago, carloskay said:

Which are the others benchmarks that they are using in the video?? i saw cinebecnh but i dindt see the others

I am surprised  those benchmarks worked. 
1. Those cards should by default be in TCC (Tesla compute cluster)

and should not even be seen by task manager or any direct X or Open GL/Val can app or game.

 

2. They got to be running the Nvidia Grid driver for those cards to be set to WDDM mode to even run some of those benchmarks. Overall I am impressed as I got my Nvidia Tesla T4 in WDDM mode but it will only run open GL games (might work for Valkin) I need to find the fix for direct X 

[power666]

The system was not running true quad SLI as there was not a 4-way nvLink bridge installed.  Rather it was two dual SLI.  Same number of GPUs in total but they are seen by applications as two logical GPUs which the driver then splits the work load again in half for each individual physical GPU.  I suspect that stable results would have been achieved by either removing the nvLink bridges so the system would see four physical GPUs or disabling a pair with nvLink bridges between them.  

 

On a more positive note, the logical GPU would have had 96 GB of memory to leverage as memory is shared over the nvLink connector.  With a true quad nvLink setup, there would be 192 GB of memory for the GPUs to leverage.

[Psittac]

downloaded installed and purchased signal RGB!!!  Quick demonetized example of it in action

 

 

[Wh0_Am_1]

Hey if you need a crazy intensive Blender file I got you. I have a file that takes 150+ hours on a RTX 2060 to render. Since this has received some attention, here is the blender file (last tested on 2.93) This is the folder, here is a texture guide. Happy burn testing!

for multi GPU support, RTX, or using a different rendering device you first select edit in the top left and go to preferences.

 

then go to system, select your render method of choice (select OptiX for RTX on Nvidia) 

Then for the actual render go to Render in the top corner and select "Render image" (or hit F12)

 

 

 

[CompoundChris]

I recently did a school project for a finance class that maxed out our school compute server. I wrote it using R. It used 48 cores and 500+ gigs of RAM(I needed more to run the full sample size lol)  and ran for multiple days(due to a large number of iterations per task). It would be pretty easy to integrate a performance monitoring package and lower the iterations so it could run in a much lower amount of time but still use as many workers as possible. It would also be a cool way to talk about R which is an open-source programing language developed for data science by data scientists and researchers. It's free and pretty intuitive as far as learning goes. It would also be a good way to benchmark across time as everything uses data that can be saved once and used across tests instead of pulling it every time. 

 

Either way, it is enjoyable to see enterprise hardware used on YouTube and it got me into the field I am in today!