We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.AR

Change to browse by:

cs

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Hardware Architecture

Title: AMOEBA: A Coarse Grained Reconfigurable Architecture for Dynamic GPU Scaling

Abstract: Different GPU applications exhibit varying scalability patterns with network-on-chip (NoC), coalescing, memory and control divergence, and L1 cache behavior. A GPU consists of several StreamingMulti-processors (SMs) that collectively determine how shared resources are partitioned and accessed. Recent years have seen divergent paths in SM scaling towards scale-up (fewer, larger SMs) vs. scale-out (more, smaller SMs). However, neither scaling up nor scaling out can meet the scalability requirement of all applications running on a given GPU system, which inevitably results in performance degradation and resource under-utilization for some applications. In this work, we investigate major design parameters that influence GPU scaling. We then propose AMOEBA, a solution to GPU scaling through reconfigurable SM cores. AMOEBA monitors and predicts application scalability at run-time and adjusts the SM configuration to meet program requirements. AMOEBA also enables dynamic creation of heterogeneous SMs through independent fusing or splitting. AMOEBA is a microarchitecture-based solution and requires no additional programming effort or custom compiler support. Our experimental evaluations with application programs from various benchmark suites indicate that AMOEBA is able to achieve a maximum performance gain of 4.3x, and generates an average performance improvement of 47% when considering all benchmarks tested.
Subjects: Hardware Architecture (cs.AR)
Cite as: arXiv:1911.03364 [cs.AR]
  (or arXiv:1911.03364v1 [cs.AR] for this version)

Submission history

From: Hui Zhao [view email]
[v1] Fri, 8 Nov 2019 16:43:54 GMT (1372kb,D)

Link back to: arXiv, form interface, contact.