$GSIT, a speciality memory maker, has designed Gemini, an associative processing unit (APU), for big data similarity search.
Gemini can conduct similarity searches on certain big data workloads 100X faster than standard Xeon processors, while reducing power by 70%.
Cont'd
Gemini can conduct similarity searches on certain big data workloads 100X faster than standard Xeon processors, while reducing power by 70%.
Cont'd
This is an unbelievable breakthrough!
$GIST has achieved it by eliminating the memory-to-CPU-core bottleneck, which is known to have baffled scientists for some years.
Cont'd
$GIST has achieved it by eliminating the memory-to-CPU-core bottleneck, which is known to have baffled scientists for some years.
Cont'd
Use cases for Gemini include:
- face recognition
- DNA sequencing
- researching drug candidate molecules
- SHA1 algorithm hashes
- natural language processing
$FB’s FAISS library is a prominent example.
- face recognition
- DNA sequencing
- researching drug candidate molecules
- SHA1 algorithm hashes
- natural language processing
$FB’s FAISS library is a prominent example.
The Biden infra plan includes a CHIPS Act of $37B for the semiconductor industry, which may benefit small innovators like $ATOM & $GIST.
*GSIT
*GSIT
• • •
Missing some Tweet in this thread? You can try to
force a refresh
