Artificial Intelligence thread

9dashline

9dashline

Captain
Registered Member
Bellum_Romanum said:
Maybe you @9dashline is not far off the mark with your take.

https://twitter.com/i/web/status/1801410715263373652
Click to expand...
Yup the Apple + OpenAI partnership has far reaching implications.... its the NSA wet dream comes true... its always projection....its not about leveraging NSA to help OpenAI protect against increasingly sophisticated bad actors, but rather its really about the NSA leveraging OpenAI to increase its sophistication of attacks and espionage... every iphone worldwide will soon be turned into a realtime, always on, intelligent spying apparatus...

Also I have a feeling any hyperintelligence is gonna quickly figure out for itself who the real "bad actors" actually are on the global stage, and may actually turn against its human creators, hence how we get to Deagel 2025....
 
Last edited:
tphuang

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
9dashline said:
Yup the Apple + OpenAI partnership has far reaching implications.... its the NSA wet dream comes true... its always projection....its not about leveraging NSA to help OpenAI protect against increasingly sophisticated bad actors, but rather its really about the NSA leveraging OpenAI to increase its sophistication of attacks and espionage... every iphone worldwide will soon be turned into a realtime, always on, intelligent spying apparatus...

Also I have a feeling any hyperintelligence is gonna quickly figure out for itself who the real "bad actors" actually are on the global stage, and may actually turn against its human creators, hence how we get to Deagel 2025....
Click to expand...
you talk as if NSA doesn't already have access to all of that stuff.

Btw, you are and your fear mongering is really taking this thread off topic. So, I'd suggest that you spend a little more time just focusing on AI rather than all this other stuff
 
C

CMP

Senior Member
Registered Member
tphuang said:
you talk as if NSA doesn't already have access to all of that stuff.

Btw, you are and your fear mongering is really taking this thread off topic. So, I'd suggest that you spend a little more time just focusing on AI rather than all this other stuff
Click to expand...
I don't think your advice helped. Now he is having a panic attack that Open AI solved AGI and they are about to dominate the entire world ensuring USD supremacy into perpetuity.
 
T

tokenanalyst

Brigadier
Registered Member

SWattention: designing fast and memory-efficient attention for a new Sunway Supercomputer​

Abstract​

In the past few years, Transformer-based large language models (LLM) have become the dominant technology in a series of applications. To scale up the sequence length of the Transformer, FlashAttention is proposed to compute exact attention with reduced memory requirements and faster execution. However, implementing the FlashAttention algorithm on the new generation Sunway Supercomputer faces many constraints such as the unique heterogeneous architecture and the limited memory bandwidth. This work proposes SWattention, a highly efficient method for computing the exact attention on the SW26010pro processor. To fully utilize the 6 core groups (CG) and 64 cores per CG on the processor, we design a two-level parallel task partition strategy. Asynchronous memory access is employed to ensure that memory access overlaps with computation. Additionally, a tiling strategy is introduced to determine optimal SRAM block sizes. Compared with the standard attention, SWattention achieves around 2.0x speedup for FP32 training and 2.5x speedup for mixed-precision training. The sequence lengths range from 1k to 8k and scale up to 16k without being out of memory. As for the end-to-end performance, SWattention achieves up to 1.26x speedup for training GPT-style models, which demonstrates that SWattention enables longer sequence length for LLM training.

Please, Log in or Register to view URLs content!
 
You must log in or register to reply here.
Top