xentr_theme_editor

  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

Who wants CHIPS? Everyone!

Some people like salsa on their chips, me I like one trillion transistors on a chip!

Have you come across any information on how the AI GPU demand has affected gaming GPU production? All I know about the AI GPUs is they are in high demand and expensive (and nobody buys just one of them).

I'm asking since I'm considering a slow accumulation of parts for a PC build as I try to leap the gap into the newer PC tech from my old computers. I don't have anything planned yet, other than to figure out what level PC to target - probably mid level with an eye on longevity other than the GPU. What I have now works well enough graphically for a the few games I play and basic computing (recently upgraded GPU & better monitor). Both Win 10 machines are old tech.
 
Last edited:
xentr_thread_starter
I don't know enough about the differences between an AI GPU and a gaming GPU. But I do know that companies gravitate very quickly to the hint of large profit and large demand. So I'm guessing, unless there's a vast difference between the 2 types of GPUs in availability & cost of parts, that the gaming market will come out second best to the AI market.
 
I beleive in a lot of cases the entry level AI GPUs are more akin to the workstation cards. Circuitboard/electronics mostly the same but the software/drivers/licensing are drastically different.
 
I beleive in a lot of cases the entry level AI GPUs are more akin to the workstation cards. Circuitboard/electronics mostly the same but the software/drivers/licensing are drastically different.
What's a workstation card cost? I suspect those aren't even in the same ballpark as an AI GPU. The NVIDIA H100 AI GPUs are around $30,000.00 US or more, EACH. Info from a few searches.

"We will be building a cluster of around 22,000 H100s," said Mustafa Suleyman, the founder of DeepMind and a co-founder of Inflection AI, reports Reuters. "This is approximately three times more compute than what was used to train all of GPT-4. Speed and scale are what's going to really enable us to build a differentiated product."

Microsoft spent hundreds of millions of dollars on tens of thousands of Nvidia A100 chips to help build ChatGPT.
 
Last edited:

Latest posts

Back
Top