Data-intensive applications users are looking for new solutions to gain a 10x factor of efficiency at server-level. The legacy processing-centric model cannot help, even with more cores because the limiting factor is the data movement.
UPMEM PIM DRAM solution is a new class of memory-centric solutions that already demonstrated a 20x performance improvement for multiple use cases.
UPMEM PIM DRAM solution is based on UPMEM DPU processor instances integrated into the DRAM memory chips, where the data is located. UPMEM DPU processor is both general purpose and optimized for data computing. The PIM chips (combining DRAM and UPMEM DPUs ) are assembled on DIMM modules, and plugged into the memory slots of the server.
PIM in DRAM is very competitive because its production cost is mostly sitting on DRAM processes one, which is half the price of a logic one.
Executing genomics algorithms in seconds instead of hours
The main genomics operations are data-intensive applications.
Mapping DNA fragments reads from a subject against a reference genome (Blast, mapping…) can be accelerated 25 to 40 times, reducing hours of processing to a couple of minutes.
Identifying variances in this genome against a reference genome (Variance calling) can also be accelerated over 20 times, reducing again hours of processing to a couple of minute.
Combining those genomics operations results in virtually real-time genomics.
Divide by 100 database index search response time, multiply throughout by 17
Searching strings of words in indexed document database is a massive well identified application. Implementing such a workload on UPMEM PIM solution drastically leverages the thousands of DPU cores that can work in parallel for each request.
It results in index search with 100 times better response time, 17 times better throughput at marginal additional cost.