05.08.2014 Views

here - Stefan-Marr.de

here - Stefan-Marr.de

here - Stefan-Marr.de

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

8. Evaluation: Performance<br />

31.62<br />

AmbientTalkST<br />

31.62<br />

LRSTM<br />

Runtime, normalized to corresponding Ad Hoc<br />

implementation, lower is better<br />

10.00<br />

3.16<br />

1.00<br />

10.00<br />

3.16<br />

1.00<br />

Binary Trees<br />

Fannkuch<br />

Fasta<br />

NBody<br />

Binary Trees<br />

Fannkuch<br />

Fasta<br />

NBody<br />

AST-OMOP on CogVM<br />

RoarVM+OMOP (opt)<br />

AST-OMOP on CogVM<br />

RoarVM+OMOP (opt)<br />

Figure 8.5.: Ad hoc vs. OMOP Kernel Benchmarks, logarithmic scale: For the kernel<br />

benchmarks, RoarVM+OMOP (opt) implementations show an average slowdown<br />

of 28% (min 2%, max 2.6x), which is close to the 17% general overhead<br />

for VM support. AST-OMOP on CogVM shows a significantly higher slowdown<br />

of 2.8x compared to ad hoc implementations. The RoarVM+OMOP (opt) implementation<br />

comes close to the i<strong>de</strong>al performance of the ad hoc implementations<br />

while providing a unifying abstraction. This shows that the OMOP can <strong>de</strong>liver<br />

on par performance when implemented efficiently.<br />

218

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!