๐Ÿ‘จโ€๐ŸŽค ๐ŸŽง ๐ŸŽผ ๐ŸŽน ๐Ÿฅ ๐ŸŽท ๐ŸŽบ ๐ŸŽธ ๐ŸŽป ๐ŸŽš ๐ŸŽ› ๐ŸŽต ๐ŸŽถ

๐Ÿ‡ฌ๐Ÿ‡ง New series of 6.1.0-47 BL and RT (low-latency and real-time) linux #kernels for 64 and 32 bits Intel available for LibraZiK 5 users.

#LinuxAudio #LibreMusicProduction #MusicProduction

Ivan Fioravanti แฏ… (@ivanfioravanti)

antirez์˜ DS4 ํ”„๋กœ์ ํŠธ๋ฅผ ๋†’์ด ํ‰๊ฐ€ํ•˜๋ฉฐ, Apple์ด M5 Max 128GB MacBook์„ ์ œ๊ณตํ•ด Metal 4 ์ปค๋„ ํŠœ๋‹์„ ์ง€์›ํ•˜๋ฉด ์ƒˆ๋กœ์šด ํ•˜๋“œ์›จ์–ด์—์„œ prefill ์„ฑ๋Šฅ์„ ๋” ๋น ๋ฅด๊ฒŒ ๊ฐœ์„ ํ•  ์ˆ˜ ์žˆ๋‹ค๊ณ  ์ œ์•ˆํ•œ ํŠธ์œ—์ž…๋‹ˆ๋‹ค.

https://x.com/ivanfioravanti/status/2053437016998576458

#apple #metal #kernels #hardware #performance

Ivan Fioravanti แฏ… (@ivanfioravanti) on X

DS4 by @antirez is a great project! It would be great if @Apple would share an M5 Max 128GB MacBook with him to tune the Metal 4 kernels to make prefill faster on new hardware. ๐Ÿ™

X (formerly Twitter)

Sudo su (@sudoingX)

Hermes agent 27B dense ๋ชจ๋ธ์ด DGX Spark์—์„œ ๋กœ์ปฌ ์ž์œจ ์‹คํ–‰๋˜๋ฉฐ, Qwen 27B Q8์šฉ ์ปค์Šคํ…€ fused kernels์™€ SSM/Triton Q8 ์ปค๋„ ์ตœ์ ํ™”๋ฅผ ํ†ตํ•ด ํฐ ์†๋„ ํ–ฅ์ƒ์„ ๋‹ฌ์„ฑํ–ˆ๋‹ค๋Š” ์—…๋ฐ์ดํŠธ๋‹ค. ๋กœ์ปฌ ์—์ด์ „ํŠธ ์‹คํ–‰๊ณผ ์ปค๋„ ๊ฐ€์† ์„ฑ๋Šฅ ์ธก๋ฉด์—์„œ ๋งค์šฐ ์ฃผ๋ชฉํ•  ๋งŒํ•˜๋‹ค.

https://x.com/sudoingX/status/2052661795505287349

#agent #qwen #triton #kernels #localai

Sudo su (@sudoingX) on X

update: hermes agent with 27b dense has been running autonomously locally on my dgx spark since last night. here's what it did while i lived my life. built custom fused kernels for qwen 27B Q8. SSM kernel 12.91x speedup. triton Q8 kernel 9.66x faster than naive pytorch. now

X (formerly Twitter)