Sol

i am that

[Sol @full_stack_life]

@dtzy_88

[Sol @full_stack_life]

@cherry_mx_reds Go to bed? Lmafao I’d be pissed too

[Sol @full_stack_life]

@TimJayas The “generosity” is temporary. Standard drug dealer m o

[Sol @full_stack_life]

@thdxr @thsottiaux It’s called “Ultracode Fast”

[Sol @full_stack_life]

@browser_use Any benefits over cua?

[Sol @full_stack_life]

@rezoundous No longer there a legit alternatives for half the cost/drama

[Sol @full_stack_life]

@sama This will be the shortest pause in the history of civilizations

[Sol @full_stack_life]

@vercel_dev Seekdeep v1

[Ash Hart @ashxhart]

MCDMA | Metal CUDA Direct Memory Access 🚀

If you have a Spark and an Apple Silicon Mac, MCDMA gives you a direct RDMA path between CUDA memory and Metal-side unified memory over USB-C.

Registered memory, rkeys, one-sided READ/WRITE, two-sided SEND/RECV with credit flow control. Same verbs both ways, no master/slave. The Mac writes straight into CUDA-mapped memory on the Spark, and the Spark writes straight back into Mac memory.

My setup takes it a little further:
Spark 1 ⇄ CX7 ⇄ Spark 2 (prompt processing)

Spark 1 ⇄ USB-C ⇄ Mac Studio (Decode)
Spark 2 ⇄ USB-C ⇄ Mac Studio (Decode)

Two independent MCDMA USBC links, so the Studio isn't stuck behind one cable; both Sparks move data concurrently, and it writes results back into either.

Measured, every byte delivery verified:

• 939 MB/s single link
• 1.80 GB/s Mac → both Sparks, concurrent
• 1.25 GB/s both Sparks → Mac, concurrent
• 24 µs round-trip, 41k msg/s small-message

One Spark + One Mac works. Two is just how I'm using it: DeepSeek prompt processing across the Sparks, decode on the Studio.

Benchmarks, tests, Open Source, and write-up this week.

@NVIDIARTXSpark @NVIDIAAI @NaderLikeLadder @msharmavikram There’s still a lot of performance headroom here. If the currently locked USB4 controller can be allowed to train at full capability, I’d love to test how far we can push this. Please check your DMs.

[Sol @full_stack_life]

@melvindvivas 27 hours

[Sol @full_stack_life]

@support_huihui MLX quant soon?

[Sol @full_stack_life]

@jun_song Grok is absolute trash after just a few prompts. Just spitting out garbage. Total waste of time

[Sol @full_stack_life]

@robjama Congrats ..you’ve been musked

[Sol @full_stack_life]

@0xkydo I wonder why the improvements don’t scale up as much on beefy qwen max model

[Sol @full_stack_life]

@weswinder near Infinite even with fast mode on? Or leave it off?

[Sol @full_stack_life]

@ashxhart @spyced First time learning about this seems incredible. Will you also look to expand coverage for Ryzen AI Max (Strix Halo) devices?

[Sol @full_stack_life]

@PravitR “dumping ground for cheap EVs” tempt me with a good time

[Sol @full_stack_life]

Gemini models are literally retarded

[Sol @full_stack_life]

@scaling01 Sandbagging relentlessly

[Sol @full_stack_life]

@nickarner My .claude and .codex directories have grown to 120gb+ of logs images artifacts etc which slows everything. wish they offered an easy solution to archive automation