The ARM chip race is getting wild… Apple M4 unveiled
1,003,336
Publicado 2024-05-08
#computerscience #ai #thecodereport
💬 Chat with Me on Discord
discord.gg/fireship
🔗 Resources
Apple M4 www.apple.com/newsroom/2024/05/apple-introduces-m4…
CPU vs GPU • CPU vs GPU vs TPU vs DPU vs QPU
CUDA in 100 Seconds • Nvidia CUDA in 100 Seconds
🔥 Get More Content - Upgrade to PRO
Upgrade at fireship.io/pro
Use code YT25 for 25% off PRO access
🎨 My Editor Settings
- Atom One Dark
- vscode-icons
- Fira Code Font
🔖 Topics Covered
- Fasted PC chips
- ARM vs x86
- RISC vs CISC
- What is the best chip architecture for AI
- Running AI models on device
- What is the M4 chip?
- Is x86
Todos los comentarios (21)
-
3:02 why tf this guy is holding his phone upside down
-
Where does this dude get his stock footage? 😂
-
people during cold war: The arms race will be wild the ARMs race now
-
M4 on iPadOS is like installing thruster engines on a horse with no legs
-
Apple car charging was hilarious.
-
singularity blaster turbo pro max, i knew it all along
-
Idk what to say but I had tacos for dinner and it was a banger. Destroyed my toilet in the process but it was worth it in at the end of the day.
-
It can run a LLM on the laptop but you still can't replace the battery.
-
I legit thought they made apple version of the M4 rifle
-
Ipad Hardware: 😎 Ipad OS: 🤡
-
The car got me 😂
-
Can't wait to buy myself the new Macbook Singularity Blaster Turbo Pro Max (with extra storage of course)
-
The Grindr reference and Apple car were wild. 10/10
-
Why is he starting to sound more like an person and not an AI
-
The M4 seems to be the fastest CPU chip on the market for running AI. You'd still want a dedicated GPU if you can get one. While the M4 gets near 40 TOPS, the 4090 gets around 83 TFLOPS (trillions of floating point operations per second) at around 32 bit precision. I think the M4 is rumored to be at 8-bit precision for their current TOPS quote. No confirmation on the 8-bit rumor though, so it's good to be skeptical of that.
-
1:13 that's actually a myth. While it was true that several years ago x86 was CISC and ARM was RISC and that was one of the reason for the latter being more energy efficient, nowadays the latest and most advanced iterations of x86 and ARM have blurred the line between the CISC and RISC. The x86 processor now is using a decoder to interpret its complex CISC intructions to a RISC form internally and ARM has introduced a lot of CISC-like extensions over the years. The actual reason x86 is more inefficient[citation needed] is due to the fact it needs to a lot of 30+ year old legacy code, though Intel have recently proposed a specification called x86-S which aims to remove a lot of legacy cruft from their processors while still maintaining reasonable backwards compatibility.
-
This is just moving from one proprietary architecture to another. WHERE MY RISC-V SQUAD AT?
-
Well I guess I'll have to throw away my M2 MacBook Pro I bough a year ago /s
-
Why is the dude at 3:03 using his phone upside down?
-
Impressive breakdown of the technology war between major players in the ARM chip market. I find it particularly fascinating that despite the CPU advancements, x86 still holds strong in infrastructure due to its age and versatility.