5 Questions about Dual GPU for Machine Learning (with Exxact dual H100 server)
Jeff Heaton Jeff Heaton
87.6K subscribers
8,032 views
0

 Published On Aug 23, 2023

In this video I cover how to use a dual GPU system for machine learning and deep learning. I look at five questions you might have about a dual GPU system.

1:08 Question 1: Do two GPUs combine into one big GPU?
2:24 Question 2: Can I buy one now, and one later?
4:21 Question 3: Can I mix multiple types of GPU?
4:45 Question 4: How do you cool multiple GPUs?
6:15 Question 5: What about NVLink?

Using a dual GPU system provided by the Exxact corporation.

https://www.exxactcorp.com/category/N...

https://www.exxactcorp.com/category/D...

The exact specs for this system are:

4U Rackmount w/ 4000W Redundant Titanium Level Power Supply
2x AMD EPYC 9654 96-core 2.40GHz CPUs
2x NVIDIA H100 80GB PCIe GPUs
24x 32GB DDR5 4800 2Rx8 LP ECC RDIMM Memory
2x 3.84TB 2.5” U.2 NVMe PCIe 4.0 x4
1x 1.92TB 2.5” U.2 NVMe PCIe 4.0 x4

#Keras #DualGPUs #DeepLearning #NVIDIAH100 #ExxactCorporation #GPUAcceleration #NeuralNetworks #MachineLearning #AI #DataScience #TrainingModels #GPUComputing

show more

Share/Embed