[www] r354465 - [EuroLLVM'19] Program update.

Wed Feb 20 06:23:43 PST 2019

Author: aadg
Date: Wed Feb 20 06:23:43 2019
New Revision: 354465

URL: http://llvm.org/viewvc/llvm-project?rev=354465&view=rev
Log:
[EuroLLVM'19] Program update.

Modified:
    www/trunk/devmtg/2019-04/index.html
    www/trunk/devmtg/2019-04/talks.html

Modified: www/trunk/devmtg/2019-04/index.html
URL: http://llvm.org/viewvc/llvm-project/www/trunk/devmtg/2019-04/index.html?rev=354465&r1=354464&r2=354465&view=diff
==============================================================================

--- www/trunk/devmtg/2019-04/index.html (original)
+++ www/trunk/devmtg/2019-04/index.html Wed Feb 20 06:23:43 2019
@@ -336,10 +336,6 @@ published here as soon as we have it ava
   Cross translation unit test case reduction</a>
   <i>Réka Kovács (Eötvös Loränd University)</i>
 </li>
-<li><a href="talks.html#LightningTalk_15">
-  Utilizing performance benefits of 32-bit compiler into a 64-bit compiler.</a>
-  <i>Venkatesh (IBM)</i>
-</li>
 </ul>
 
 <b>BoFs</b>
@@ -408,12 +404,10 @@ published here as soon as we have it ava
   <i>Jack J. Garzella (University of Utah), Marek Baranowski (University of Utah),
     Shaobo He (University of Utah), Zvonimir Rakamaric (University of Utah)</i>
 </li>
-<!--
 <li><a href="talks.html#Poster_8">
   Instruction Tracing and dynamic codegen analysis to identify unique llvm performance issues.</a>
   <i>Biplob (IBM)</i>
 </li>
--->
 <li><a href="talks.html#Poster_9">
   Handling all Facebook requests with JITed C++ code</a>
   <i>Huapeng Zhou (Facebook), Yuhan Guo (Facebook)</i>
@@ -466,10 +460,6 @@ USA)</i>
   Optimizing Nondeterminacy: Exploiting Race Conditions in Parallel Programs</a>
   <i>William S. Moses (MIT CSAIL)</i>
 </li>
-<li><a href="talks.html#Poster_21">
-  Utilizing performance benefits of 32-bit compiler into a 64-bit compiler.</a>
-  <i>Venkatesh (IBM)</i>
-</li>
 </ul>
 
 <div class="www_sectiontitle" id="registration">Registration</div>

Modified: www/trunk/devmtg/2019-04/talks.html
URL: http://llvm.org/viewvc/llvm-project/www/trunk/devmtg/2019-04/talks.html?rev=354465&r1=354464&r2=354465&view=diff
==============================================================================
--- www/trunk/devmtg/2019-04/talks.html (original)
+++ www/trunk/devmtg/2019-04/talks.html Wed Feb 20 06:23:43 2019
@@ -997,26 +997,6 @@ be applied to multiple TUs in their orig
 make it possible to use C-Reduce for cross translation unit test case
 reduction.</p>
 </td></tr>
-
-<tr><td valign="top" id="LightningTalk_15">
-<b>Utilizing performance benefits of 32-bit compiler into a 64-bit compiler.</b><br>
-<i>Venkatesh (IBM)</i>
-<p>Some applications show better performance when executed on a 32-bit
-processor. Most of these applications are stack,  cache and pointer intensive
-in nature. When the same application is compiled and run on a 64-bit machine
-generally show certain performance degradation's. There are certain
-performance opportunities which can be learned and deployed in higher bit
-compilers.</p>
-<p>In this talk we analyze few applications compiled by a llvm compiler for a
-32-bit and a 64-bit processor, and analyze how certain performance gains in a
-32-bit compiled program be extracted and be used to train the llvm compiler to
-benefit a 64-bit scenario. For example a  SPEC CPU2017 INT benchmark namely
-523.xalanc_r, is a linked list based application which is high on cache misses.
-A 64-bit version of this benchmark has nearly 50% higher data cache misses as
-compared to a 32-bit one. Some of the cache misses in 64-bit compiled
-application can actually be reduced. We will discuss how this can be achieved
-and some other scenarios which will benefit a 64-bit compiler.</p>
-</td></tr>
 </table>
 
 <div class="www_sectiontitle" id="BoF">BoFs</div>
@@ -1302,7 +1282,7 @@ languages. We assess the quality of our
 through several case studies, and we describe the lessons we learned in the
 process.</p>
 </td></tr>
-<!--
+
 <tr><td valign="top" id="Poster_8">
 <b>Instruction Tracing and dynamic codegen analysis to identify unique llvm performance issues.</b><br>
 <i>Biplob (IBM)</i>
@@ -1322,7 +1302,7 @@ will look at some of these code-gen issu
 running application compiled by llvm and other compilers were traced for hot
 code sections on IBM Power9 processor.</p>
 </td></tr>
--->
+
 <tr><td valign="top" id="Poster_9">
 <b>Handling all Facebook requests with JITed C++ code</b><br>
 <i>Huapeng Zhou (Facebook), Yuhan Guo (Facebook)</i>
@@ -1577,26 +1557,6 @@ nondeterminacy through a number of examp
 their theoretical implications as well as how they perform when implemented
 atop the Tapir extension to LLVM.</p>
 </td></tr>
-
-<tr><td valign="top" id="Poster_21">
-<b>Utilizing performance benefits of 32-bit compiler into a 64-bit compiler.</b><br>
-<i>Venkatesh (IBM)</i>
-<p>Some applications show better performance when executed on a 32-bit
-processor. Most of these applications are stack,  cache and pointer intensive
-in nature. When the same application is compiled and run on a 64-bit machine
-generally show certain performance degradation's. There are certain
-performance opportunities which can be learned and deployed in higher bit
-compilers.</p>
-<p>In this talk we analyze few applications compiled by a llvm compiler for a
-32-bit and a 64-bit processor, and analyze how certain performance gains in a
-32-bit compiled program be extracted and be used to train the llvm compiler to
-benefit a 64-bit scenario. For example a  SPEC CPU2017 INT benchmark namely
-523.xalanc_r, is a linked list based application which is high on cache misses.
-A 64-bit version of this benchmark has nearly 50% higher data cache misses as
-compared to a 32-bit one. Some of the cache misses in 64-bit compiled
-application can actually be reduced. We will discuss how this can be achieved
-and some other scenarios which will benefit a 64-bit compiler.</p>
-</td></tr>
 </table>
 
 <!-- *********************************************************************** -->


