[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17107040#comment-17107040 ] Chinna Rao Lalam commented on HIVE-23133: - LGTM +1 > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch, HIVE-23133.6.patch > > > Use java.lang.StrictMath rather than java.lang.Math for compatibility in the > following UDFs: > UDFCos.java > UDFExp.java > UDFLn.java > UDFLog.java > UDFLog2.java > UDFLog10.java > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17107039#comment-17107039 ] Chinna Rao Lalam commented on HIVE-23133: - This issue fixes math functions which are failing in ARM platform like log, exp, cos, toRadians. [~Kevin_Zheng] please check other math functions if any in follow up jira. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch, HIVE-23133.6.patch > > > Use java.lang.StrictMath rather than java.lang.Math for compatibility in the > following UDFs: > UDFCos.java > UDFExp.java > UDFLn.java > UDFLog.java > UDFLog2.java > UDFLog10.java > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17106213#comment-17106213 ] Hive QA commented on HIVE-23133: Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/13002798/HIVE-23133.6.patch {color:red}ERROR:{color} -1 due to no test(s) being added or modified. {color:green}SUCCESS:{color} +1 due to 17268 tests passed Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/22307/testReport Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/22307/console Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-22307/ Messages: {noformat} Executing org.apache.hive.ptest.execution.TestCheckPhase Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.YetusPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase {noformat} This message is automatically generated. ATTACHMENT ID: 13002798 - PreCommit-HIVE-Build > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch, HIVE-23133.6.patch > > > Use java.lang.StrictMath rather than java.lang.Math for compatibility in the > following UDFs: > UDFCos.java > UDFExp.java > UDFLn.java > UDFLog.java > UDFLog2.java > UDFLog10.java > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17106197#comment-17106197 ] Hive QA commented on HIVE-23133: | (/) *{color:green}+1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | || || || || {color:brown} master Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 2m 25s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 9m 28s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 27s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 1s{color} | {color:green} master passed {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 4m 35s{color} | {color:blue} ql in master has 1527 extant Findbugs warnings. {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 16s{color} | {color:green} master passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 31s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 52s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 23s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m 23s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 2s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 44s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 15s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 17s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black} 31m 58s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Optional Tests | asflicense javac javadoc findbugs checkstyle compile | | uname | Linux hiveptest-server-upstream 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u5 (2017-09-19) x86_64 GNU/Linux | | Build tool | maven | | Personality | /data/hiveptest/working/yetus_PreCommit-HIVE-Build-22307/dev-support/hive-personality.sh | | git revision | master / 9ffbbdc | | Default Java | 1.8.0_111 | | findbugs | v3.0.1 | | modules | C: vector-code-gen ql U: . | | Console output | http://104.198.109.242/logs//PreCommit-HIVE-Build-22307/yetus.txt | | Powered by | Apache Yetushttp://yetus.apache.org | This message was automatically generated. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch, HIVE-23133.6.patch > > > Use java.lang.StrictMath rather than java.lang.Math for compatibility in the > following UDFs: > UDFCos.java > UDFExp.java > UDFLn.java > UDFLog.java > UDFLog2.java > UDFLog10.java > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] >
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17106042#comment-17106042 ] Zhenyu Zheng commented on HIVE-23133: - [~gopalv] Thanks for the review, in previous patches (1-3) we didn't add config, if that is the better way, we can keep it that way. I will add the funcs that we changed to the ticket description. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17105599#comment-17105599 ] Gopal Vijayaraghavan commented on HIVE-23133: - Adding a configuration is probably a bad idea, considering this is for transcendental functions & not for commonly used OLAP functions (like avg or sum). The missing part of this ticket is what functions are modified for compatibility, instead of just linking to test runs. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17105393#comment-17105393 ] Hive QA commented on HIVE-23133: Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/13002694/HIVE-23133.5.patch {color:red}ERROR:{color} -1 due to no test(s) being added or modified. {color:red}ERROR:{color} -1 due to 4 failed/errored test(s), 17265 tests executed *Failed tests:* {noformat} org.apache.hadoop.hive.cli.TestCompareCliDriver.testCliDriver[vectorized_math_funcs] (batchId=229) org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[vectorized_math_funcs] (batchId=54) org.apache.hadoop.hive.ql.txn.compactor.TestWorker2.minorTableWithBase (batchId=256) org.apache.hive.jdbc.TestActivePassiveHA.testActivePassiveHA (batchId=215) {noformat} Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/22283/testReport Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/22283/console Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-22283/ Messages: {noformat} Executing org.apache.hive.ptest.execution.TestCheckPhase Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.YetusPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase Tests exited with: TestsFailedException: 4 tests failed {noformat} This message is automatically generated. ATTACHMENT ID: 13002694 - PreCommit-HIVE-Build > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17105369#comment-17105369 ] Hive QA commented on HIVE-23133: | (x) *{color:red}-1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | || || || || {color:brown} master Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m 53s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 8m 17s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 31s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 9s{color} | {color:green} master passed {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 0m 33s{color} | {color:blue} common in master has 63 extant Findbugs warnings. {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 3m 45s{color} | {color:blue} ql in master has 1527 extant Findbugs warnings. {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 18s{color} | {color:green} master passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 29s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 55s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 28s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m 28s{color} | {color:green} the patch passed {color} | | {color:red}-1{color} | {color:red} checkstyle {color} | {color:red} 0m 39s{color} | {color:red} ql: The patch generated 2 new + 2 unchanged - 0 fixed = 4 total (was 2) {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 32s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 16s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 14s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black} 30m 18s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Optional Tests | asflicense javac javadoc findbugs checkstyle compile | | uname | Linux hiveptest-server-upstream 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u5 (2017-09-19) x86_64 GNU/Linux | | Build tool | maven | | Personality | /data/hiveptest/working/yetus_PreCommit-HIVE-Build-22283/dev-support/hive-personality.sh | | git revision | master / 8bfdd18 | | Default Java | 1.8.0_111 | | findbugs | v3.0.1 | | checkstyle | http://104.198.109.242/logs//PreCommit-HIVE-Build-22283/yetus/diff-checkstyle-ql.txt | | modules | C: common vector-code-gen ql U: . | | Console output | http://104.198.109.242/logs//PreCommit-HIVE-Build-22283/yetus.txt | | Powered by | Apache Yetushttp://yetus.apache.org | This message was automatically generated. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch, HIVE-23133.5.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] >
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17104571#comment-17104571 ] Hive QA commented on HIVE-23133: Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/13002631/HIVE-23133.4.patch {color:red}ERROR:{color} -1 due to no test(s) being added or modified. {color:red}ERROR:{color} -1 due to 2 failed/errored test(s), 17253 tests executed *Failed tests:* {noformat} org.apache.hadoop.hive.cli.TestMiniDruidCliDriver.testCliDriver[druid_materialized_view_rewrite_ssb] (batchId=128) org.apache.hadoop.hive.metastore.txn.TestTxnHandler.testReplTimeouts (batchId=244) {noformat} Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/22267/testReport Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/22267/console Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-22267/ Messages: {noformat} Executing org.apache.hive.ptest.execution.TestCheckPhase Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.YetusPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase Tests exited with: TestsFailedException: 2 tests failed {noformat} This message is automatically generated. ATTACHMENT ID: 13002631 - PreCommit-HIVE-Build > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17104537#comment-17104537 ] Hive QA commented on HIVE-23133: | (/) *{color:green}+1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 1s{color} | {color:green} The patch does not contain any @author tags. {color} | || || || || {color:brown} master Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 2m 4s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 8m 6s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 29s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 6s{color} | {color:green} master passed {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 0m 32s{color} | {color:blue} common in master has 63 extant Findbugs warnings. {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 3m 39s{color} | {color:blue} ql in master has 1527 extant Findbugs warnings. {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 16s{color} | {color:green} master passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 29s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 55s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 27s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m 27s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 1m 7s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 4m 32s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 17s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 14s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black} 30m 5s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Optional Tests | asflicense javac javadoc findbugs checkstyle compile | | uname | Linux hiveptest-server-upstream 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u5 (2017-09-19) x86_64 GNU/Linux | | Build tool | maven | | Personality | /data/hiveptest/working/yetus_PreCommit-HIVE-Build-22267/dev-support/hive-personality.sh | | git revision | master / 7e77f25 | | Default Java | 1.8.0_111 | | findbugs | v3.0.1 | | modules | C: common vector-code-gen ql U: . | | Console output | http://104.198.109.242/logs//PreCommit-HIVE-Build-22267/yetus.txt | | Powered by | Apache Yetushttp://yetus.apache.org | This message was automatically generated. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Zhenyu Zheng >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch, HIVE-23133.4.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] >
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17095838#comment-17095838 ] Chinna Rao Lalam commented on HIVE-23133: - Thanks for the patch. StrictMath helps portability of java programs, so using strictmath is good. I am considering these 2 things about strictmath. # Performance issue, as published jdk performance comparison differences are small. # Backward compatibility , but with strictmath it is fixing the problem completely. Any thoughts on this [~hashutosh] [~gopalv] [~jcamachorodriguez] > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17092970#comment-17092970 ] Zhenyu Zheng commented on HIVE-23133: - For those that might have concerns on the performance degration of changing from Math to StrictMath, here is some testing data: [https://bugs.openjdk.java.net/browse/JDK-8210416] from the data, we can see that the difference between Math and StrictMath function is actually very very small and StrictMath provides correct answers. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch, > HIVE-23133.3.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17081162#comment-17081162 ] Hive QA commented on HIVE-23133: Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12999530/HIVE-23133.2.patch {color:red}ERROR:{color} -1 due to no test(s) being added or modified. {color:green}SUCCESS:{color} +1 due to 18211 tests passed Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/21559/testReport Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/21559/console Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-21559/ Messages: {noformat} Executing org.apache.hive.ptest.execution.TestCheckPhase Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.YetusPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase {noformat} This message is automatically generated. ATTACHMENT ID: 12999530 - PreCommit-HIVE-Build > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17081138#comment-17081138 ] Hive QA commented on HIVE-23133: | (/) *{color:green}+1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 0s{color} | {color:green} The patch does not contain any @author tags. {color} | || || || || {color:brown} master Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m 51s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 8m 6s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 13s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 53s{color} | {color:green} master passed {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 3m 43s{color} | {color:blue} ql in master has 1528 extant Findbugs warnings. {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 6s{color} | {color:green} master passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 28s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 34s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 9s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m 9s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 54s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 3m 58s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 3s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 16s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black} 26m 53s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Optional Tests | asflicense javac javadoc findbugs checkstyle compile | | uname | Linux hiveptest-server-upstream 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u5 (2017-09-19) x86_64 GNU/Linux | | Build tool | maven | | Personality | /data/hiveptest/working/yetus_PreCommit-HIVE-Build-21559/dev-support/hive-personality.sh | | git revision | master / 5f47808 | | Default Java | 1.8.0_111 | | findbugs | v3.0.1 | | modules | C: vector-code-gen ql U: . | | Console output | http://104.198.109.242/logs//PreCommit-HIVE-Build-21559/yetus.txt | | Powered by | Apache Yetushttp://yetus.apache.org | This message was automatically generated. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch, HIVE-23133.2.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17080013#comment-17080013 ] Hive QA commented on HIVE-23133: Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12999397/HIVE-23133.1.patch {color:red}ERROR:{color} -1 due to no test(s) being added or modified. {color:red}ERROR:{color} -1 due to 7 failed/errored test(s), 18207 tests executed *Failed tests:* {noformat} org.apache.hadoop.hive.cli.TestCliDriver.testCliDriver[vectorized_math_funcs] (batchId=23) org.apache.hadoop.hive.cli.TestCompareCliDriver.testCliDriver[vectorized_math_funcs] (batchId=303) org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[subquery_select] (batchId=178) org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[vector_decimal_udf2] (batchId=188) org.apache.hadoop.hive.cli.TestMiniLlapLocalCliDriver.testCliDriver[vectorized_math_funcs] (batchId=171) org.apache.hadoop.hive.cli.TestSparkCliDriver.testCliDriver[subquery_select] (batchId=136) org.apache.hadoop.hive.cli.TestSparkCliDriver.testCliDriver[vectorized_math_funcs] (batchId=127) {noformat} Test results: https://builds.apache.org/job/PreCommit-HIVE-Build/21536/testReport Console output: https://builds.apache.org/job/PreCommit-HIVE-Build/21536/console Test logs: http://104.198.109.242/logs/PreCommit-HIVE-Build-21536/ Messages: {noformat} Executing org.apache.hive.ptest.execution.TestCheckPhase Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.YetusPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase Tests exited with: TestsFailedException: 7 tests failed {noformat} This message is automatically generated. ATTACHMENT ID: 12999397 - PreCommit-HIVE-Build > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote on X86 machines), this is because of we > use [Math > Library|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html] for > these kind of calculations. > and according to the > [illustration|https://docs.oracle.com/javase/6/docs/api/java/lang/Math.html]: > _Unlike some of the numeric methods of class StrictMath, all implementations > of the equivalent functions of class Math are not_ > _defined to return the bit-for-bit same results. This relaxation permits > better-performing implementations where strict reproducibility_ > _is not required._ > _By default many of the Math methods simply call the equivalent method in > StrictMath for their implementation._ > _Code generators are encouraged to use platform-specific native libraries or > microprocessor instructions, where available,_ > _to provide higher-performance implementations of Math methods._ > so the result will have difference across hardware archs. > On the other hand, JAVA provided another library > [StrictMath|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html] > that will not have this kind of problem as according to its' > [reference|https://docs.oracle.com/javase/6/docs/api/java/lang/StrictMath.html]: > _To help ensure portability of Java programs, the definitions of some of the > numeric functions in this package require that they produce_ > _the same results as certain published algorithms._ > So in order to fix the above mentioned problem, we have to consider switch to > use StrictMath instead of Math. > > > -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-23133) Numeric operations can have different result across hardware archs
[ https://issues.apache.org/jira/browse/HIVE-23133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17079995#comment-17079995 ] Hive QA commented on HIVE-23133: | (/) *{color:green}+1 overall{color}* | \\ \\ || Vote || Subsystem || Runtime || Comment || || || || || {color:brown} Prechecks {color} || | {color:green}+1{color} | {color:green} @author {color} | {color:green} 0m 1s{color} | {color:green} The patch does not contain any @author tags. {color} | || || || || {color:brown} master Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 1m 48s{color} | {color:blue} Maven dependency ordering for branch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 7m 52s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 9s{color} | {color:green} master passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 54s{color} | {color:green} master passed {color} | | {color:blue}0{color} | {color:blue} findbugs {color} | {color:blue} 3m 43s{color} | {color:blue} ql in master has 1527 extant Findbugs warnings. {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 1s{color} | {color:green} master passed {color} | || || || || {color:brown} Patch Compile Tests {color} || | {color:blue}0{color} | {color:blue} mvndep {color} | {color:blue} 0m 29s{color} | {color:blue} Maven dependency ordering for patch {color} | | {color:green}+1{color} | {color:green} mvninstall {color} | {color:green} 1m 35s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} compile {color} | {color:green} 1m 5s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javac {color} | {color:green} 1m 5s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} checkstyle {color} | {color:green} 0m 52s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} whitespace {color} | {color:green} 0m 0s{color} | {color:green} The patch has no whitespace issues. {color} | | {color:green}+1{color} | {color:green} findbugs {color} | {color:green} 3m 54s{color} | {color:green} the patch passed {color} | | {color:green}+1{color} | {color:green} javadoc {color} | {color:green} 1m 3s{color} | {color:green} the patch passed {color} | || || || || {color:brown} Other Tests {color} || | {color:green}+1{color} | {color:green} asflicense {color} | {color:green} 0m 14s{color} | {color:green} The patch does not generate ASF License warnings. {color} | | {color:black}{color} | {color:black} {color} | {color:black} 26m 17s{color} | {color:black} {color} | \\ \\ || Subsystem || Report/Notes || | Optional Tests | asflicense javac javadoc findbugs checkstyle compile | | uname | Linux hiveptest-server-upstream 3.16.0-4-amd64 #1 SMP Debian 3.16.43-2+deb8u5 (2017-09-19) x86_64 GNU/Linux | | Build tool | maven | | Personality | /data/hiveptest/working/yetus_PreCommit-HIVE-Build-21536/dev-support/hive-personality.sh | | git revision | master / 796a9c5 | | Default Java | 1.8.0_111 | | findbugs | v3.0.1 | | modules | C: vector-code-gen ql U: . | | Console output | http://104.198.109.242/logs//PreCommit-HIVE-Build-21536/yetus.txt | | Powered by | Apache Yetushttp://yetus.apache.org | This message was automatically generated. > Numeric operations can have different result across hardware archs > -- > > Key: HIVE-23133 > URL: https://issues.apache.org/jira/browse/HIVE-23133 > Project: Hive > Issue Type: Sub-task >Reporter: Zhenyu Zheng >Assignee: Yikun Jiang >Priority: Major > Attachments: HIVE-23133.1.patch > > > Currently, we have set up an ARM CI to test out how Hive works on ARM > platform: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/] > Among the failures, we have observed that some numeric operations can have > different result across hardware archs, such as: > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestCliDriver/testCliDriver_vector_decimal_udf2_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_subquery_select_/] > [https://builds.apache.org/view/H-L/view/Hive/job/Hive-linux-ARM-trunk/25/testReport/org.apache.hadoop.hive.cli/TestSparkCliDriver/testCliDriver_vectorized_math_funcs_/] > we can see that the calculation results of log, exp, cos, toRadians etc is > slitly different than the .out file results that we are > comparing(they are tested and wrote