June 23rd, 2024

Finnish startup says it can speed up any CPU by 100x

A Finnish startup, Flow Computing, introduces the Parallel Processing Unit (PPU) chip promising 100x CPU performance boost for AI and autonomous vehicles. Despite skepticism, CEO Timo Valtonen is optimistic about partnerships and industry adoption.

Read original articleLink Icon
Finnish startup says it can speed up any CPU by 100x

A Finnish startup called Flow Computing claims to have developed a chip, the Parallel Processing Unit (PPU), that can potentially increase CPU performance by up to 100 times without the need for recoding. This innovation could have significant implications for AI technologies and autonomous vehicles. The PPU works as a companion chip that optimizes processing tasks in real-time, converting serial processing into parallel operations without additional power consumption or excessive heat. By managing tasks at nanosecond intervals, the technology allows for multiple processes to occur simultaneously, boosting throughput without altering the CPU's clock speed or architecture. While the bold claims have sparked skepticism, Flow Computing's CEO, Timo Valtonen, remains confident in the technology's potential. The startup has secured initial funding and is seeking partnerships within the industry to further develop and scale its solution. The success of this technology hinges on its adoption by chip manufacturers, which may require adjustments to current production methods. Despite challenges, the promise of significant performance enhancements with minimal modifications makes Flow Computing's innovation an intriguing prospect for the industry.

Related

Unisoc and Xiaomi's 4nm Chips Said to Challenge Qualcomm and MediaTek

Unisoc and Xiaomi's 4nm Chips Said to Challenge Qualcomm and MediaTek

UNISOC and Xiaomi collaborate on 4nm chips challenging Qualcomm and MediaTek. UNISOC's chip features X1 big core + A78 middle core + A55 small core with Mali G715 MC7 GPU, offering competitive performance and lower power consumption. Xiaomi's Xuanjie chip includes X3 big core + A715 middle core + A510 small core with IMG CXT 48-1536 GPU, potentially integrating a MediaTek baseband. Xiaomi plans a separate mid-range phone line with Xuanjie chips, aiming to strengthen its market presence. The successful development of these 4nm chips by UNISOC and Xiaomi marks progress in domestically produced mobile chips, enhancing competitiveness.

Now is a good time to start a service business

Now is a good time to start a service business

Initiating a service business is advantageous now, according to Zach Ocean. Service businesses offer immediate revenue, avoiding R&D phases. AI advancements enable semi-autonomous workers, enhancing service business growth potential.

20x Faster Background Removal in the Browser Using ONNX Runtime with WebGPU

20x Faster Background Removal in the Browser Using ONNX Runtime with WebGPU

Using ONNX Runtime with WebGPU and WebAssembly in browsers achieves 20x speedup for background removal, reducing server load, enhancing scalability, and improving data security. ONNX models run efficiently with WebGPU support, offering near real-time performance. Leveraging modern technology, IMG.LY aims to enhance design tools' accessibility and efficiency.

AI is exhausting the power grid

AI is exhausting the power grid

Tech firms, including Microsoft, face a power crisis due to AI's energy demands straining the grid and increasing emissions. Fusion power exploration aims to combat fossil fuel reliance, but current operations heavily impact the environment.

Merlin Labs aims to test an AI-powered KC-135 within a year

Merlin Labs aims to test an AI-powered KC-135 within a year

Two startups, Merlin Labs and EpiSci, collaborate to develop self-flying tankers and advanced AI for dogfighting. They aim to create AI pilots to reduce human pilot dependency in Air Force missions.

Link Icon 4 comments
By @Nokinside - 4 months
I have been looking at their patents. Seems well defended and methodological. Someone like Intel, Nvidia or AMD might buy them soon.

Speedup is realistic. Multicore SMP or NUMA are not good for memory access patterns they optimize.

Their thick control flow model that should work for exclusive matrix-addition and log-prefix style memory access patterns. In comparison to the baseline the speedup is 150% in log-prefix algorithm, over 190% in fft-style butterfly algorithm, 50-100% in matrix addition and threshold filtering. silicon area and power consumption are estimated to be low.

Light reading material:

Optimizing Memory Access in TCF Processors with Compute-Update Operations Optimizing Memory Access in TCF Processors with Compute-Update Operations https://ieeexplore.ieee.org/document/9150423

The REPLICA on-chip network https://ieeexplore.ieee.org/document/7792877/

Preliminary Performance and Memory Access Scalability Study of Thick Control Flow Processors https://ieeexplore.ieee.org/document/10305463/

Realizing multioperations and multiprefixes in Thick Control Flow processors https://linkinghub.elsevier.com/retrieve/pii/S01419331230005...

By @wmf - 4 months
By @elromulous - 4 months
I could not be more skeptical of this. Do folks know more about this claim?
By @bboreham - 4 months
Is it middle-out?