$5000/stick or something.
1x 64 GB stick for only 10k USD.
I'm still on DDR4. Fuck me, first make the RAM affordable, and then I might consider fully upgrading my PC (as it would need a CPU and motherboard upgrade too at that point).
So far I haven't felt the limit of my 5600x and 32 GB of DDR4.
So many non-gaming workloads these days are capacity driven more than bandwidth driven. Any AI, LLM, image generation, video editing, etc. would be much better off with 32GB of DDR3 than 8GB of DDR4/5.
32GB of DDR4 is a good sweet spot to be in today.
What is all this about Dance Dance Revolution?
I'm tired boss.
DDR6?
I'm on a desktop from 2012 because I can't afford to upgrade.
4 times the bandwidth, 16 times the price?
Optimist over here
and yet, you can only sell so many kidneys
Slight correction, you can only sell so many of your own kidneys. But there's no limit otherwise!
oh, I wasn't talking about mine !
In this economy ?
How nice for data centers. If my DDR4 PC breaks my backups are a chromebook and retro handhelds.
My back up is my steam deck :(
Look at this guy with his fancy expensive new steam deck, me over here with my 3DS
Hey that 3ds is a fantastic little device. Respect it!
I... Uh... Have both. ^^;
Like it’s going to even be released the public. Lmao. Most we will get is in laptops and it’ll be $6000.
Who wants a laptop witb 1KB of DDR6 RAM?
wtf. I laughed so much at this. Omfg 😂
Even Bill wouldn't be willing to settle for less than 640Kb.
Still less space than a nomad. Lame.
(jk the nomad was 6gb)
Who cares? It's not like you can afford to buy it.
Honestly CAMM makes way more sense. It’s just a better interface, electrically. Why do desktops need sticks if they’re going to be higher voltage and slower…
And we will be able to afford this. Some day.
One thing I've been wondering about, after seeing an article a while back about Nvidia talking to TMSC about manufacturing memory is whether they might do an ASIC aimed at being a neural net or part of one.
The limiting factor today is memory bandwidth. But that's because you have to ship all the data over a long, narrow memory bus with big fat traces.
If you can stick a bunch of adders and multipliers on a single chip and put memory on the same chip, then you don't have to slow things down and generate the heat that shipping everything off-chip requires.
If so, that might be one form that AI-optimized memory might take.
Nvidia bought GRoQ for exactly this. 13,000 tokens a second from an expansion card that uses 200 watts. AMD bought Taalas for the same reason.
Nah, Groq is toast. Those numbers are for toy models distributed across tons of cards, and splitting MoE LLMs across so many cards in SRAM an inefficient approach IMO.
The best model they serve is the kinda obsolete GPT-OSS 120B, at 350 token/s: https://openrouter.ai/openai/gpt-oss-120b
Which you can get with Deepseek, Muse or other open weights models these days, if you really need it.
The only ASIC maker I kind like is Cerebras, especially their scheme for hybrid inference in conjunction with AMD. Architecturally, that makes a lot of sense.
EDIT:
Upon further investigation, they host more models than I thought:
https://console.groq.com/docs/models
They have MiniMax 2.7 at 200 tokens/s, which isn't too bad... but still largely obsolete now.
Yep. Samsung is already researching this. They have been for some time.
Also, I feel Bitnet would be useful here, as it would only require small adders, no huge multipliers.
They could even take a hybrid approach, processing higher-precision attention and small parts of the LLM on the GPU 'traditionally' while specialized adders in memory handle the huge ternary MLPs (or whatever the sparse equivalent is in other ML architectures). This would also give them architectural flexibility (as attention schemes are evolving wildly, the MLPs kinda aren't).
No, you can't have it! Who do you think you are, filthy peasant?!?
Man the datacenters will love those 😭
Most people can only afford DDR3 RAM now. Maybe DDR6 will be affordable in a couple decades assuming it doesn't all get burned up in data center satellites.
They're going to send it all to Mars. It's bait for their mousetrap, they will need workers to build their Martian cities, and they know we'll eventually get desperate enough to try and build a new PC, and then we'll be trapped there.
would windows 11 get more usable with this?
It will be able to run more AI and store more telemetry to send back to Microsoft, so it will be more usable... for Microsoft. And really, that's all that matters. So the answer to your question is "yes." You won't notice any difference, but the amount of money Microsoft is losing from their AI bullshit will be incrementally less, and isn't that why you use Windows to begin with, and doesn't that make you feel good somehow? /s
Technology
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.