正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Concurrency vs Parallelism

Alex Xu

295,221 subscribers

40,263 次观看 • 1 年前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

parallelism 2015 | 2026 GIRLS LIKE GIRLS

parallelism 2015 | 2026 GIRLS LIKE GIRLS

VICTÓRIA

23,991 次观看 • 3 个月前

Excited to share our new blog: Scaling Video Training with Parallelism We discuss how parallelism helps scale long-video training systems, for both understanding and generation.

Excited to share our new blog: Scaling Video Training with Parallelism We discuss how parallelism helps scale long-video training systems, for both understanding and generation.

Yukang Chen

12,802 次观看 • 4 个月前

•one for the parallelism in #meemsemohabbat [ishq hai- ft. talha and roshi]

•one for the parallelism in #meemsemohabbat [ishq hai- ft. talha and roshi]

sarah

22,148 次观看 • 1 年前

tristan&isolde were a constant parallelism to our dollophead fr😭😭

tristan&isolde were a constant parallelism to our dollophead fr😭😭

Jack Blues 🫶🇵🇸

23,283 次观看 • 1 个月前

Training LFMs at scale means solving parallelism across every layer of the architecture. And not all layers are the same. Our CTO Mathias Lechner (Mathias Lechner) sits down with Liquid's founding engineer Paul Pak (Paul Pak) to talk training infrastructure: Data, tensor, pipeline, expert, and context parallelism, and how they make context parallelism work across hybrid architectures with both attention and convolution operators.

Training LFMs at scale means solving parallelism across every layer of the architecture. And not all layers are the same. Our CTO Mathias Lechner (Mathias Lechner) sits down with Liquid's founding engineer Paul Pak (Paul Pak) to talk training infrastructure: Data, tensor, pipeline, expert, and context parallelism, and how they make context parallelism work across hybrid architectures with both attention and convolution operators.

Liquid AI

18,169 次观看 • 4 个月前

🔥 Outstanding interactive visualizer from Alex which demonstrates Optimistic vs Pessimistic Concurrency Control! Rialo supports both modes giving developers maximum flexibility based on their use case! This interactive demo will be live soon for anyone to play with!

🔥 Outstanding interactive visualizer from Alex which demonstrates Optimistic vs Pessimistic Concurrency Control! Rialo supports both modes giving developers maximum flexibility based on their use case! This interactive demo will be live soon for anyone to play with!

ade | rialo.io

36,735 次观看 • 10 个月前

Planning for #iOSConfSG? Join Donny Wals (Donny Wals 👾) for a hands-on Swift Concurrency Workshop on Jan 15th. Master: ✅ async/await ✅ Structured Concurrency ✅ Actors, Sendable & Data Race Protection ✅ Tips for smooth Swift 6 migration 🎟️ #SwiftDev

Planning for #iOSConfSG? Join Donny Wals (Donny Wals 👾) for a hands-on Swift Concurrency Workshop on Jan 15th. Master: ✅ async/await ✅ Structured Concurrency ✅ Actors, Sendable & Data Race Protection ✅ Tips for smooth Swift 6 migration 🎟️ #SwiftDev

iOS Conf SG

112,504 次观看 • 1 年前

I made a video demystifying database consistency, isolation levels, record contention, and pessimistic vs. optimistic concurrency control tradeoffs. (I also said "dollars" instead of "pounds" 40 times.) Check it out if you want a basic working knowledge of this stuff.

I made a video demystifying database consistency, isolation levels, record contention, and pessimistic vs. optimistic concurrency control tradeoffs. (I also said "dollars" instead of "pounds" 40 times.) Check it out if you want a basic working knowledge of this stuff.

Jamie Turner

122,554 次观看 • 8 个月前

The art of parallelism in Finale Undone. Finally I made an edit from the new card🥹🤍

The art of parallelism in Finale Undone. Finally I made an edit from the new card🥹🤍

mi! 🐦‍⬛🚗🌹

42,306 次观看 • 4 个月前

Day 3 of learning Go after office hours Studied about Concurrency in Go: - how Go handles concurrency - Locks and Mutex - Goroutines - Channels (sender & receiver pattern) It's very intuitive in Go coming from some experience in Java. Will deep dive more in the coming days!

Day 3 of learning Go after office hours Studied about Concurrency in Go: - how Go handles concurrency - Locks and Mutex - Goroutines - Channels (sender & receiver pattern) It's very intuitive in Go coming from some experience in Java. Will deep dive more in the coming days!

avrl ☘

82,127 次观看 • 5 个月前

Think DeepSpeed is only good for ZeRO and its usability, not training throughput? At #PyTorchCon North America 2026, Masahiro Tanaka will cover tensor, sequence, and expert parallelism in DeepSpeed and walk through benchmarks. His session, “DeepSpeed Is Not Just ZeRO: Tensor, Sequence, and Expert Parallelism with ZeRO-Level Usability,” will cover AutoTP, AutoSP, and AutoEP, how these capabilities can be combined with ZeRO, and where each parallelism strategy fits. Register for PyTorch Conference North America 2026:

Think DeepSpeed is only good for ZeRO and its usability, not training throughput? At #PyTorchCon North America 2026, Masahiro Tanaka will cover tensor, sequence, and expert parallelism in DeepSpeed and walk through benchmarks. His session, “DeepSpeed Is Not Just ZeRO: Tensor, Sequence, and Expert Parallelism with ZeRO-Level Usability,” will cover AutoTP, AutoSP, and AutoEP, how these capabilities can be combined with ZeRO, and where each parallelism strategy fits. Register for PyTorch Conference North America 2026:

PyTorch

15,911 次观看 • 27 天前

PWR Chain brings internet-like scalability to blockchain. 🌐 Here’s how 👇 • Others: Parallelism at smart contract level within one VM • PWR: Parallelism at VM and software application level - Supports any VM or applications - Similar to parallel servers on the internet - Efficient inter-application communication

PWR Chain brings internet-like scalability to blockchain. 🌐 Here’s how 👇 • Others: Parallelism at smart contract level within one VM • PWR: Parallelism at VM and software application level - Supports any VM or applications - Similar to parallel servers on the internet - Efficient inter-application communication

PWR Chain

94,655 次观看 • 2 年前

How Fast is Gemma 4 on a MacBook Pro M4? Benchmarking Google's new MoE (26B-A4B) > Model size: 26.1 GiB > Load time: ~4.2s Comparing single request VS > concurrent requests performance > 32k total context, 4 parallel slots single request behavior > TTFT: 5.68s > prompt: 3,701 tokens @ 652 tok/s > decode: 40.08 tok/s sequential (1 request at a time): > avg duration: 20.5s > p99: 22.1s > throughput: 40.11 tok/s > clean finishes: 100% concurrent (4 parallel requests): > aggregate throughput: 47.25 tok/s > total system throughput: 262.27 tok/s > avg duration: 65.1s > p95 latency: 68.8s > req/sec: 0.058 Head-to-Head: Sequential vs Concurrent throughput: > 40.11 tok/s → 47.25 tok/s (+17.8%) > small gain despite 4x parallelism latency per request: > 20.5s → 65.1s (~3.2x slower) > you pay heavily for concurrency system throughput (true utilization): > ~40 tok/s → 262 tok/s (~6.5x total output) > this is where concurrency wins tokens per second (decode ceiling): > ~40 tok/s steady in both modes > hardware-bound, not scheduler-bound TTFT impact: > ~5.7s baseline → buried under queueing in concurrent > “headers waittime” becomes the bottleneck What this actually means? - You don’t get linear scaling from parallel slots - You trade latency for total output - Mac Unified Memory setup is clearly saturating - Bandwidth + Scheduling overhead show up immediately This is exactly why GPUs dominate here Concurrency without killing latency

How Fast is Gemma 4 on a MacBook Pro M4? Benchmarking Google's new MoE (26B-A4B) > Model size: 26.1 GiB > Load time: ~4.2s Comparing single request VS > concurrent requests performance > 32k total context, 4 parallel slots single request behavior > TTFT: 5.68s > prompt: 3,701 tokens @ 652 tok/s > decode: 40.08 tok/s sequential (1 request at a time): > avg duration: 20.5s > p99: 22.1s > throughput: 40.11 tok/s > clean finishes: 100% concurrent (4 parallel requests): > aggregate throughput: 47.25 tok/s > total system throughput: 262.27 tok/s > avg duration: 65.1s > p95 latency: 68.8s > req/sec: 0.058 Head-to-Head: Sequential vs Concurrent throughput: > 40.11 tok/s → 47.25 tok/s (+17.8%) > small gain despite 4x parallelism latency per request: > 20.5s → 65.1s (~3.2x slower) > you pay heavily for concurrency system throughput (true utilization): > ~40 tok/s → 262 tok/s (~6.5x total output) > this is where concurrency wins tokens per second (decode ceiling): > ~40 tok/s steady in both modes > hardware-bound, not scheduler-bound TTFT impact: > ~5.7s baseline → buried under queueing in concurrent > “headers waittime” becomes the bottleneck What this actually means? - You don’t get linear scaling from parallel slots - You trade latency for total output - Mac Unified Memory setup is clearly saturating - Bandwidth + Scheduling overhead show up immediately This is exactly why GPUs dominate here Concurrency without killing latency

Ahmad

88,866 次观看 • 7 个月前

GLM5 running on a cluster of 4 Mac Studios using tensor parallelism thanks to MLX distributed. Yes, those are the Alex drawers.

GLM5 running on a cluster of 4 Mac Studios using tensor parallelism thanks to MLX distributed. Yes, those are the Alex drawers.

Alex Ziskind

20,594 次观看 • 8 个月前

the parallelism of huiju running to wan just like he did for her 😭 theyre literally the definition of “run for your love” #PerfectCrownEp11 #PerfectCrown

the parallelism of huiju running to wan just like he did for her 😭 theyre literally the definition of “run for your love” #PerfectCrownEp11 #PerfectCrown

ris 🍀

69,512 次观看 • 5 个月前

Speed up sequential async/await code with promise concurrency But be aware of a couple critical gotchas! Read more in my latest article:

Speed up sequential async/await code with promise concurrency But be aware of a couple critical gotchas! Read more in my latest article:

Steve (Builder.io)

459,161 次观看 • 3 年前

The concept of zuko perfectly fitting into water benders trait while Katara fitting into fire benders is so interesting THEIR PARALLELISM IS SO FUCKING COOL

The concept of zuko perfectly fitting into water benders trait while Katara fitting into fire benders is so interesting THEIR PARALLELISM IS SO FUCKING COOL

yuki ☾𖤓

49,781 次观看 • 5 个月前

the parallelism that there is between these two in live, #taekook there is so much subtlety and there are still pple who doubt how close they are?

the parallelism that there is between these two in live, #taekook there is so much subtlety and there are still pple who doubt how close they are?

⭒

40,815 次观看 • 3 年前

this takedown ft mdzs edit is ICONIC. the cinematic parallelism of it all— jgy's deceptive sweetness, nhs dropping the hammer and wwx sweeping in for the ultimate takedown. goosebumps

this takedown ft mdzs edit is ICONIC. the cinematic parallelism of it all— jgy's deceptive sweetness, nhs dropping the hammer and wwx sweeping in for the ultimate takedown. goosebumps

piu•

29,902 次观看 • 1 年前