Tiny Corp Details More Of Their Planned Tinybox System Specs

[phoronix]

Phoronix: Tiny Corp Details More Of Their Planned Tinybox System Specs

After putting their AMD GPU powered Tinybox "on hold" only to decide a few days later to offer both AMD and NVIDIA graphics options for Tinybox compute systems, George Hotz' Tiny Corp has now shared more specifications for these planned "green" and "red" Tinybox designs...

https://www.phoronix.com/news/Tinybox-Green-Red-Specs

[skeevy420]

I'm loling at the "mediocre" description.

Michael, they're not technically typos, but all the tiny corp product names should be lowercase. All lowercase, tiny, product names is their shtick with branding.

[zeroepoch]

128GB RAM is very underwhelming for a system like this. These days that's like the minimum for a data science workstation with just 1 maybe 2 GPUs. The fact that they list more GPU RAM basically makes this point. You have to load models into CPU memory first before you can transfer it to GPU memory.

[CommunityMember]

I'm loling at the "mediocre" description.

Michael, they're not technically typos, but all the tiny corp product names should be lowercase. All lowercase, tiny, product names is their shtick with branding.

e e cummings approves this message

[Quackdoc]

In a response on Twitter/X they shared that there isn't an Intel Arc Graphics option currently due to lacking an Intel discrete graphics card with sufficient video memory bandwidth.

That's a bit disappointing but understandable, Hopefully the next generation cards will work out.

[User42]

Michael, they're not technically typos, but all the tiny corp product names should be lowercase. All lowercase, tiny, product names is their shtick with branding.

Proper names in English are capitalized. Tiny is free to use lowercase all they want for branding of course but general literature should probably follow general rules instead of the stylistic rule of the moment of a specific company.

[sophisticles]

So many things wrong:

1) They sell a computer they call "tinybox", which is 12U in height, or 21 inches, hardly tiny.

2) Their website looks like it was designed by a 10 year old back in the early 90's. I tend to favor minimalism, but this is ridiculous.

3) They list 144gb vram, how do they get to that? 6 cards that have 24gb vram each? If so, then why 128gb system ram?

I can't imagine someone spending $25 to buy something from them, much less $25,000.

The make System76 look good in comparison.

[Quackdoc]

So many things wrong:

1) They sell a computer they call "tinybox", which is 12U in height, or 21 inches, hardly tiny.

Who cares? It's a box from tinycorp, and the relative size for what it packs is pretty good if you are one of the people who are actually in the market for something like this.

2) Their website looks like it was designed by a 10 year old back in the early 90's. I tend to favor minimalism, but this is ridiculous.

Who cares? The site does what it needs to do. I personally much prefer this over the craptastic sites with janky CSS and crappy JS. Anything more is just fluff anyways.

3) They list 144gb vram, how do they get to that? 6 cards that have 24gb vram each? If so, then why 128gb system ram?

Who cares? It's clearly designed for ML/AI based on the needs of tinygrad.

I can't imagine someone spending $25 to buy something from them, much less $25,000.

That's because your head is so far up your ass you can't imagine that people have different needs and wants then you.

The make System76 look good in comparison.

They both look good

[Weasel]

128GB RAM is very underwhelming for a system like this. These days that's like the minimum for a data science workstation with just 1 maybe 2 GPUs. The fact that they list more GPU RAM basically makes this point. You have to load models into CPU memory first before you can transfer it to GPU memory.

That doesn't mean anything, you can slice it and load it in slices on the GPU.

But yes, 128 GB of RAM is a bit low for this price.

[varikonniemi]

They are trying to enter a market space Bitcoin miners perfected 5 years ago. Give them time, they surely will get "AI" folks served at some point when the miners already have their datacenters up and running for the past year.