https://helix.eecs.harvard.edu/index.php?title=ISCA2014&feed=atom&action=historyISCA2014 - Revision history2024-03-29T06:37:34ZRevision history for this page on the wikiMediaWiki 1.16.5https://helix.eecs.harvard.edu/index.php?title=ISCA2014&diff=73&oldid=prevSimone at 20:09, 20 June 20142014-06-20T20:09:08Z<p></p>
<table style="background-color: white; color:black;">
<col class='diff-marker' />
<col class='diff-content' />
<col class='diff-marker' />
<col class='diff-content' />
<tr valign='top'>
<td colspan='2' style="background-color: white; color:black;">← Older revision</td>
<td colspan='2' style="background-color: white; color:black;">Revision as of 20:09, 20 June 2014</td>
</tr><tr><td colspan="2" class="diff-lineno">Line 10:</td>
<td colspan="2" class="diff-lineno">Line 10:</td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td class='diff-marker'>-</td><td style="background: #ffa; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper]] ] [ [[media:ISCA2014_Slides.pdf|Slides]] ]</div></td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper]] ] [ [[media:ISCA2014_Slides.pdf|Slides<ins class="diffchange diffchange-inline">]] ] [ [[media:ISCA2014_SlidesFF.pdf|Fast Forward</ins>]] ]</div></td></tr>
</table>Simonehttps://helix.eecs.harvard.edu/index.php?title=ISCA2014&diff=71&oldid=prevSimone at 20:06, 20 June 20142014-06-20T20:06:21Z<p></p>
<table style="background-color: white; color:black;">
<col class='diff-marker' />
<col class='diff-content' />
<col class='diff-marker' />
<col class='diff-content' />
<tr valign='top'>
<td colspan='2' style="background-color: white; color:black;">← Older revision</td>
<td colspan='2' style="background-color: white; color:black;">Revision as of 20:06, 20 June 2014</td>
</tr><tr><td colspan="2" class="diff-lineno">Line 10:</td>
<td colspan="2" class="diff-lineno">Line 10:</td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td class='diff-marker'>-</td><td style="background: #ffa; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper]] ]</div></td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper<ins class="diffchange diffchange-inline">]] ] [ [[media:ISCA2014_Slides.pdf|Slides</ins>]] ]</div></td></tr>
</table>Simonehttps://helix.eecs.harvard.edu/index.php?title=ISCA2014&diff=49&oldid=prevSimone at 13:22, 21 April 20142014-04-21T13:22:01Z<p></p>
<table style="background-color: white; color:black;">
<col class='diff-marker' />
<col class='diff-content' />
<col class='diff-marker' />
<col class='diff-content' />
<tr valign='top'>
<td colspan='2' style="background-color: white; color:black;">← Older revision</td>
<td colspan='2' style="background-color: white; color:black;">Revision as of 13:22, 21 April 2014</td>
</tr><tr><td colspan="2" class="diff-lineno">Line 7:</td>
<td colspan="2" class="diff-lineno">Line 7:</td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>''Proc. International Symposium on Computer Architecture (ISCA), June, 2014''</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>''Proc. International Symposium on Computer Architecture (ISCA), June, 2014''</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div><ins style="color: red; font-weight: bold; text-decoration: none;"><br></ins></div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper]] ]</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>[ [[media:ISCA2014_Paper.pdf|Paper]] ]</div></td></tr>
</table>Simonehttps://helix.eecs.harvard.edu/index.php?title=ISCA2014&diff=47&oldid=prevSimone at 13:21, 21 April 20142014-04-21T13:21:26Z<p></p>
<table style="background-color: white; color:black;">
<col class='diff-marker' />
<col class='diff-content' />
<col class='diff-marker' />
<col class='diff-content' />
<tr valign='top'>
<td colspan='2' style="background-color: white; color:black;">← Older revision</td>
<td colspan='2' style="background-color: white; color:black;">Revision as of 13:21, 21 April 2014</td>
</tr><tr><td colspan="2" class="diff-lineno">Line 1:</td>
<td colspan="2" class="diff-lineno">Line 1:</td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>__NOTITLE__</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>__NOTITLE__</div></td></tr>
<tr><td class='diff-marker'>-</td><td style="background: #ffa; color:black; font-size: smaller;"><div>= <del class="diffchange diffchange-inline">Decoupling Communication From Computation Makes Non</del>-<del class="diffchange diffchange-inline">Numerical </del>Programs <del class="diffchange diffchange-inline">Parallelizable </del>=</div></td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div>= <ins class="diffchange diffchange-inline">HELIX</ins>-<ins class="diffchange diffchange-inline">RC: An Architecture-Compiler Co-Design for Automatic Parallelization of Irregular </ins>Programs =</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei, David Brooks</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei, David Brooks</div></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div><ins style="color: red; font-weight: bold; text-decoration: none;"></ins></div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div><br></div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div><br></div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>''Proc. International Symposium on Computer Architecture (ISCA), June, 2014''</div></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"><div>''Proc. International Symposium on Computer Architecture (ISCA), June, 2014''</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td><td class='diff-marker'> </td><td style="background: #eee; color:black; font-size: smaller;"></td></tr>
<tr><td class='diff-marker'>-</td><td style="background: #ffa; color:black; font-size: smaller;"><div><del class="diffchange diffchange-inline">TBD</del></div></td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div><ins class="diffchange diffchange-inline">Data dependences in sequential programs limit parallelization because extracted threads cannot run independently. Although thread-level speculation can avoid the need for precise dependence analysis, communication overheads required to synchronize actual dependences counteract the benefits of parallelization. To address these challenges, we propose a lightweight architectural enhancement co-designed with a parallelizing compiler, which together can decouple communication from thread execution. Simulations of these approaches, applied to a processor with 16 Intel Atom-like cores, show an average of 6.85x performance speedup for six SPEC CINT2000 benchmarks.</ins></div></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div> </div></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="background: #cfc; color:black; font-size: smaller;"><div><ins class="diffchange diffchange-inline">[ [[media:ISCA2014_Paper.pdf|Paper]] ]</ins></div></td></tr>
</table>Simonehttps://helix.eecs.harvard.edu/index.php?title=ISCA2014&diff=45&oldid=prevSimone: Created page with "__NOTITLE__ = Decoupling Communication From Computation Makes Non-Numerical Programs Parallelizable = Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei,..."2014-03-11T15:35:48Z<p>Created page with "__NOTITLE__ = Decoupling Communication From Computation Makes Non-Numerical Programs Parallelizable = Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei,..."</p>
<p><b>New page</b></p><div>__NOTITLE__<br />
= Decoupling Communication From Computation Makes Non-Numerical Programs Parallelizable =<br />
<br />
Simone Campanoni, Kevin Brownell, Svilen Kanev, Timothy Jones, Gu-Yeon Wei, David Brooks<br />
<br><br />
''Proc. International Symposium on Computer Architecture (ISCA), June, 2014''<br />
<br />
TBD</div>Simone