96 TFlop and 64GB

  • 3 Replies
  • 5768 Views
*

frankinstien

  • Replicant
  • ********
  • 653
    • Knowledgeable Machines
96 TFlop and 64GB
« on: June 15, 2023, 11:39:59 pm »
The fact that you can get a GPU for free is amazing but the horse power is pretty limited and the better GPUs are very expensive to rent let alone buy and hard to reserve. Particularly if you have to train a model for a month! I have an old server with dual Xeons and 128GB of ram and it's just sitting there doing nothing. I recently priced the Instinct MI25s that boast a 24 TFlops F16 with 16GB for $80! If I purchase 4 of those I'd have 96TFlops and 64GB of ram.  Of course the model's architectures would have to be MoE(Mixture of Experts) and the routing done by the CPU end. But in comparison to an A100 40GB on eBay at $7500 to $14,000 to a $320 expense I'm just about to pull the trigger.

The problem however is power, the 4 Instinct MI25 GPUs would draw 300 watts each. However...I do have some 2500 watts in solar panels hanging around.  :) Of course I could buy more GPUs, say double it which amounts to $640 that then leads to 192 TFlops F16 and 128GB of gpu ram, with power draw of 2400 watts!  :D
« Last Edit: June 16, 2023, 12:06:08 am by frankinstien »

*

WriterOfMinds

  • Trusty Member
  • ********
  • Replicant
  • *
  • 617
    • WriterOfMinds Blog
Re: 96 TFlop and 64GB
« Reply #1 on: June 16, 2023, 02:22:09 pm »
The fact that you can get a GPU for free is amazing but the horse power is pretty limited

Where can one get a GPU for free, I'm intrigued. Are you just referring to the limited free trial/low-use plans of cloud services?

*

frankinstien

  • Replicant
  • ********
  • 653
    • Knowledgeable Machines
Re: 96 TFlop and 64GB
« Reply #2 on: June 17, 2023, 02:34:46 am »
Google Collab...

*

MagnusWootton

  • Replicant
  • ********
  • 646
Re: 96 TFlop and 64GB
« Reply #3 on: July 03, 2023, 02:31:23 pm »
my gpu on my comp is overheating bad,  cant do any largish workloads with it,  better not get knockoff video cards they have defects.

 


Requirements for functional equivalence to conscious processing?
by DaltonG (General AI Discussion)
November 19, 2024, 11:56:05 am
Will LLMs ever learn what is ... is?
by HS (Future of AI)
November 10, 2024, 06:28:10 pm
Who's the AI?
by frankinstien (Future of AI)
November 04, 2024, 05:45:05 am
Project Acuitas
by WriterOfMinds (General Project Discussion)
October 27, 2024, 09:17:10 pm
Ai improving AI
by infurl (AI Programming)
October 19, 2024, 03:43:29 am
Atronach's Eye
by WriterOfMinds (Home Made Robots)
October 13, 2024, 09:52:42 pm
Running local AI models
by spydaz (AI Programming)
October 07, 2024, 09:00:53 am
Hi IM BAA---AAACK!!
by MagnusWootton (Home Made Robots)
September 16, 2024, 09:49:10 pm
LLaMA2 Meta's chatbot released
by spydaz (AI News )
August 24, 2024, 02:58:36 pm
ollama and llama3
by spydaz (AI News )
August 24, 2024, 02:55:13 pm
AI controlled F-16, for real!
by frankinstien (AI News )
June 15, 2024, 05:40:28 am
Open AI GPT-4o - audio, vision, text combined reasoning
by MikeB (AI News )
May 14, 2024, 05:46:48 am
OpenAI Speech-to-Speech Reasoning Demo
by MikeB (AI News )
March 31, 2024, 01:00:53 pm
Say good-bye to GPUs...
by MikeB (AI News )
March 23, 2024, 09:23:52 am
Google Bard report
by ivan.moony (AI News )
February 14, 2024, 04:42:23 pm
Elon Musk's xAI Grok Chatbot
by MikeB (AI News )
December 11, 2023, 06:26:33 am

Users Online

362 Guests, 0 Users

Most Online Today: 497. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles