DPU maker claims 100x speedup vs. Xeon for big data similarity search

[erek]

"For comparison, a Nvidia A100 GPU server can complete 104 x 4,096 bits per 1.4GHz clock cycle, providing a 7TB/sec memory bandwidth. The Gemini chip’s memory bandwidth leaves the A100 trailing in the dust, and the Xeon CPU is even further behind.

*Hamming Distance When a computer runs a search it deals with a search term, represented as a binary string, and it looks for equivalent or similar search terms. The similarity can be expressed as the difference between the search term string and a target string, described as the number of bit positions in which the two bits are different.

It works like this; envisage two strings of equal length; 1101 1001 and 1001 1101. Add them together, 11011001 + 10011101, to get = 01000100. This contains two 1s and the Hamming distance is 2. Other things being equal, strings with smaller Hamming distances are more likely to represent things that are similar than strings with greater Hamming distances. The ‘things’ can be facial recognition images, genomes, drug candidate molecules, SHA1 algorithm hashes and so forth."

https://blocksandfiles.com/2020/11/30/similarity-search-xeons-and-gpus/

 

[cdabc123]

"For comparison, a Nvidia A100 GPU server can complete 104 x 4,096 bits per 1.4GHz clock cycle, providing a 7TB/sec memory bandwidth. The Gemini chip’s memory bandwidth leaves the A100 trailing in the dust, and the Xeon CPU is even further behind.

*Hamming Distance When a computer runs a search it deals with a search term, represented as a binary string, and it looks for equivalent or similar search terms. The similarity can be expressed as the difference between the search term string and a target string, described as the number of bit positions in which the two bits are different.

It works like this; envisage two strings of equal length; 1101 1001 and 1001 1101. Add them together, 11011001 + 10011101, to get = 01000100. This contains two 1s and the Hamming distance is 2. Other things being equal, strings with smaller Hamming distances are more likely to represent things that are similar than strings with greater Hamming distances. The ‘things’ can be facial recognition images, genomes, drug candidate molecules, SHA1 algorithm hashes and so forth."

https://blocksandfiles.com/2020/11/30/similarity-search-xeons-and-gpus/

Artical lacked the good details but it looks like a asic with a decent amount of cache. The fact that they are not marketing such as a mining chip makes me think they really cant compete in any market. I loled when they used sha256 hashs as a comparison to a CPU.

 

[Lakados]

Artical lacked the good details but it looks like a asic with a decent amount of cache. The fact that they are not marketing such as a mining chip makes me think they really cant compete in any market. I loled when they used sha256 hashs as a comparison to a CPU.

Yes and no it seems to be a very single purpose chip designed for finding duplicate entries in massive datasets.
When training an AI the quality of the initial dataset is key, and having duplicate entities or ones so similar as to trigger the same response it doesn’t add anything other than processing time. So a system for cleaning data ahead of time could be a pretty big deal.
This could also be used by big data for filtering through billions of accounts to look for duplicate or fake accounts to be flagged for merger or removal. There’s a fair number of practical applications for it.