As we also need to store the BWT of the reverse genome to calculate the bound, the memory required for calculating intervals is doubled, or about 2.3 GB for a 3 Gb genome.
This bound on memory usage is critical for the visualization of large data sets.
r3 series represent memory optimized amazon instances and are recommended for memory bound applications such as high performance databases and distributed cache, in-memory analytics, genome assembly, and larger deployments of SAP.
The configuration bounds of the processing threads may not align exactly with the data structure bounds, and there is no memory bound protections on GPUs, so such out-of-bounds memory writes can easily overwrite other data.
As an example, in the case of Bowtie2, which is a memory bound application; the key optimisation consists in improving locality of the memory accesses and utilisation of shared memory bandwidth.
Compute and memory bound Microarchitectures exploit parallelism to accelerate compute-bound algorithms but offer fewer advantages to memory-bound algorithms.
The alignment algorithm is entirely memory bound, and its running time is dominated by the random memory access patterns of the sampled occurrence array and the compressed BWT string during the SA interval computation.
As integral image computation is known as being memory bound, but also a very simple algorithm (3 LOADs, 1 STORE, and 3 ADDs), it is quite impossible to reduce its complexity.
A BufferTable instance is provided with an upper memory bound which it will not exceed.
Moreover, we obtain a stronger bound on the memory decay rate.
Memory bound applications such as solvers for large sparse systems of equations remain a challenge for GPUs.
