Re: Running the bivar-stats example

[Narayanaswami, Sandeep]

That’s correct, I’m using the 0.10.0 release.

From: Glenn Weidner <gweid...@us.ibm.com>
Reply-To: "dev@systemml.incubator.apache.org" 
<dev@systemml.incubator.apache.org>
Date: Thursday, October 27, 2016 at 15:22
To: "dev@systemml.incubator.apache.org" <dev@systemml.incubator.apache.org>
Subject: Re: Running the bivar-stats example


Hi Sandeep,

No problem and thank you for trying the other scenarios. Just one quick 
question - are you using the 0.10.0 binary release?

Thanks,
Glenn


[nactive hide details for "Narayanaswami, Sandeep" ---10/27/2016 
01:14:26]"Narayanaswami, Sandeep" ---10/27/2016 01:14:26 PM---Hi, sorry I 
couldn’t get back on this sooner. @Glenn: Thanks for pointing that out. 
Unfortunately,

From: "Narayanaswami, Sandeep" <sandeep.narayanasw...@capitalone.com>
To: "dev@systemml.incubator.apache.org" <dev@systemml.incubator.apache.org>
Date: 10/27/2016 01:14 PM
Subject: Re: Running the bivar-stats example





Hi, sorry I couldn’t get back on this sooner.

@Glenn: Thanks for pointing that out. Unfortunately, switching to Spark 1.6 
didn’t make any difference: I got the same error/stacktrace again.
@Matthias: I did have the corresponding metadata files. I also tried modifying 
the script to print the matrix after reading it in, and that works as expected 
so I don’t think it’s a data read issue.

I’d appreciate any other suggestions.

Cheers,
Sandeep

On 10/20/16, 02:30, "Matthias Boehm" <mboe...@googlemail.com> wrote:

   apart from the missing support for Spark 2.x, I would recommend to
   double check your inputs. Do you have meta data files along with your
   inputs? If not, then these inputs are assumed to be in default format
   "text", i.e., matrix market ijv representation (not csv). You can either
   provide such json meta data files or simply put a 'format="csv"' into
   the read statements in bivar-stats.dml.

   Regards,
   Matthias


   On 10/20/2016 2:44 AM, Glenn Weidner wrote:
   > Hi Sandeep,
   >
   > Thank you for contacting the Apache SystemML community.  Based on the
   > following trace fragment:
   >
   > spark.SparkContext: Running Spark version 2.0.0
   >
   > it looks like you're running against Spark 2.0 which is not officially
   > supported with the current master code base.  Can you retry with an earlier
   > Spark version?  Let me know if I've misinterpreted your Spark environment.
   >
   > Cheers,
   > Glenn
   >
   >
   >
   > From: "Narayanaswami, Sandeep" <sandeep.narayanasw...@capitalone.com>
   > To: "dev@systemml.incubator.apache.org"
   > <dev@systemml.incubator.apache.org>
   > Date: 10/19/2016 05:00 PM
   > Subject: Running the bivar-stats example
   >
   >
   >
   > Hi SystemML community,
   >
   > I’m trying to run the bivariate stats example script on Spark. I’m only
   > doing a toy example with very small data, but I have been running into
   > errors.
   > My script invocation is as follows:
   >
   > spark-submit --master local[4] SystemML.jar -f
   > scripts/algorithms/bivar-stats.dml -exec hybrid_spark -nvargs
   > X=/data/iqtv.csv index1=/data/index1.csv index2=/data/index2.csv
   > types1=/data/types1.csv types2=/data/types2.csv OUTDIR=/data/outdata
   >
   > Am I calling the script incorrectly?
   > (All the files are on hdfs. Iqtv.csv is a headerless csv with two numeric
   > fields. The index1 and index2 files are text files containing only the
   > strings ‘1’ and ‘2’ respectively. The types* files are text files each
   > containing only the string ‘1’.)
   >
   > Please excuse the rather long stack trace below, but I’m hoping it will
   > help diagnose the problem.
   >
   > Cheers,
   > Sandeep
   >
   > 

   >
   >
   > 16/10/19 16:46:26 INFO api.DMLScript: BEGIN DML run 10/19/2016 16:46:26
   > ANTLR Tool version 4.3 used for code generation does not match the current
   > runtime version 4.5.3[…]
   > 16/10/19 16:46:28 INFO spark.SparkContext: Running Spark version 2.0.0
   > […]
   > 16/10/19 16:46:29 INFO api.DMLScript: END DML run 10/19/2016 16:46:29
   > 16/10/19 16:46:29 ERROR api.DMLScript: Failed to execute DML script.
   > org.apache.sysml.runtime.DMLRuntimeException:
   > org.apache.sysml.runtime.DMLRuntimeException:
   > org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile program
   > block hierarchy.
   >at org.apache.sysml.runtime.controlprogram.Program.execute
   > (Program.java:152)
   >at org.apache.sysml.api.DMLScrip

Re: Running the bivar-stats example

[Glenn Weidner]

Hi Sandeep,

No problem and thank you for trying the other scenarios.  Just one quick
question - are you using the 0.10.0 binary release?

Thanks,
Glenn




From:   "Narayanaswami, Sandeep" <sandeep.narayanasw...@capitalone.com>
To: "dev@systemml.incubator.apache.org"
<dev@systemml.incubator.apache.org>
Date:   10/27/2016 01:14 PM
Subject:    Re: Running the bivar-stats example



Hi, sorry I couldn’t get back on this sooner.

@Glenn: Thanks for pointing that out. Unfortunately, switching to Spark 1.6
didn’t make any difference: I got the same error/stacktrace again.
@Matthias: I did have the corresponding metadata files. I also tried
modifying the script to print the matrix after reading it in, and that
works as expected so I don’t think it’s a data read issue.

I’d appreciate any other suggestions.

Cheers,
Sandeep

On 10/20/16, 02:30, "Matthias Boehm" <mboe...@googlemail.com> wrote:

apart from the missing support for Spark 2.x, I would recommend to
double check your inputs. Do you have meta data files along with your
inputs? If not, then these inputs are assumed to be in default format
"text", i.e., matrix market ijv representation (not csv). You can
either
provide such json meta data files or simply put a 'format="csv"' into
the read statements in bivar-stats.dml.

Regards,
Matthias


On 10/20/2016 2:44 AM, Glenn Weidner wrote:
> Hi Sandeep,
>
> Thank you for contacting the Apache SystemML community.  Based on the
> following trace fragment:
>
> spark.SparkContext: Running Spark version 2.0.0
>
> it looks like you're running against Spark 2.0 which is not
officially
> supported with the current master code base.  Can you retry with an
earlier
> Spark version?  Let me know if I've misinterpreted your Spark
environment.
>
> Cheers,
> Glenn
>
>
>
> From:  "Narayanaswami, Sandeep"
<sandeep.narayanasw...@capitalone.com>
> To:"dev@systemml.incubator.apache.org"
> <dev@systemml.incubator.apache.org>
> Date:  10/19/2016 05:00 PM
> Subject:   Running the bivar-stats example
>
>
>
> Hi SystemML community,
>
> I’m trying to run the bivariate stats example script on Spark. I’m
only
> doing a toy example with very small data, but I have been running
into
> errors.
> My script invocation is as follows:
>
> spark-submit --master local[4] SystemML.jar -f
> scripts/algorithms/bivar-stats.dml -exec hybrid_spark -nvargs
> X=/data/iqtv.csv index1=/data/index1.csv index2=/data/index2.csv
> types1=/data/types1.csv types2=/data/types2.csv OUTDIR=/data/outdata
>
> Am I calling the script incorrectly?
> (All the files are on hdfs. Iqtv.csv is a headerless csv with two
numeric
> fields. The index1 and index2 files are text files containing only
the
> strings ‘1’ and ‘2’ respectively. The types* files are text files
each
> containing only the string ‘1’.)
>
> Please excuse the rather long stack trace below, but I’m hoping it
will
> help diagnose the problem.
>
> Cheers,
> Sandeep
>
>


>
>
> 16/10/19 16:46:26 INFO api.DMLScript: BEGIN DML run 10/19/2016
16:46:26
> ANTLR Tool version 4.3 used for code generation does not match the
current
> runtime version 4.5.3[…]
> 16/10/19 16:46:28 INFO spark.SparkContext: Running Spark version
2.0.0
> […]
> 16/10/19 16:46:29 INFO api.DMLScript: END DML run 10/19/2016 16:46:29
> 16/10/19 16:46:29 ERROR api.DMLScript: Failed to execute DML script.
> org.apache.sysml.runtime.DMLRuntimeException:
> org.apache.sysml.runtime.DMLRuntimeException:
> org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile
program
> block hierarchy.
>at org.apache.sysml.runtime.controlprogram.Program.execute
> (Program.java:152)
>at org.apache.sysml.api.DMLScript.execute(DMLScript.java:677)
>at org.apache.sysml.api.DMLScript.executeScript
(DMLScript.java:337)
>at org.apache.sysml.api.DMLScript.main(DMLScript.java:192)
>at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
>at sun.reflect.NativeMethodAccessorImpl.invoke
> (NativeMethodAccessorImpl.java:62)
>at sun.reflect.DelegatingMethodAccessorImpl.invoke
> (DelegatingMethodAccessorImpl.java:43)
>at java.lang.reflect.Method.invoke(Method.jav

Re: Running the bivar-stats example

[Narayanaswami, Sandeep]

Hi, sorry I couldn’t get back on this sooner. 

@Glenn: Thanks for pointing that out. Unfortunately, switching to Spark 1.6 
didn’t make any difference: I got the same error/stacktrace again. 
@Matthias: I did have the corresponding metadata files. I also tried modifying 
the script to print the matrix after reading it in, and that works as expected 
so I don’t think it’s a data read issue. 

I’d appreciate any other suggestions. 

Cheers,
Sandeep

On 10/20/16, 02:30, "Matthias Boehm"  wrote:

apart from the missing support for Spark 2.x, I would recommend to 
double check your inputs. Do you have meta data files along with your 
inputs? If not, then these inputs are assumed to be in default format 
"text", i.e., matrix market ijv representation (not csv). You can either 
provide such json meta data files or simply put a 'format="csv"' into 
the read statements in bivar-stats.dml.

Regards,
Matthias


On 10/20/2016 2:44 AM, Glenn Weidner wrote:
> Hi Sandeep,
>
> Thank you for contacting the Apache SystemML community.  Based on the
> following trace fragment:
>
> spark.SparkContext: Running Spark version 2.0.0
>
> it looks like you're running against Spark 2.0 which is not officially
> supported with the current master code base.  Can you retry with an 
earlier
> Spark version?  Let me know if I've misinterpreted your Spark environment.
>
> Cheers,
> Glenn
>
>
>
> From: "Narayanaswami, Sandeep" 
> To:   "dev@systemml.incubator.apache.org"
> 
> Date: 10/19/2016 05:00 PM
> Subject:  Running the bivar-stats example
>
>
>
> Hi SystemML community,
>
> I’m trying to run the bivariate stats example script on Spark. I’m only
> doing a toy example with very small data, but I have been running into
> errors.
> My script invocation is as follows:
>
> spark-submit --master local[4] SystemML.jar -f
> scripts/algorithms/bivar-stats.dml -exec hybrid_spark -nvargs
> X=/data/iqtv.csv index1=/data/index1.csv index2=/data/index2.csv
> types1=/data/types1.csv types2=/data/types2.csv OUTDIR=/data/outdata
>
> Am I calling the script incorrectly?
> (All the files are on hdfs. Iqtv.csv is a headerless csv with two numeric
> fields. The index1 and index2 files are text files containing only the
> strings ‘1’ and ‘2’ respectively. The types* files are text files each
> containing only the string ‘1’.)
>
> Please excuse the rather long stack trace below, but I’m hoping it will
> help diagnose the problem.
>
> Cheers,
> Sandeep
>
> 

>
>
> 16/10/19 16:46:26 INFO api.DMLScript: BEGIN DML run 10/19/2016 16:46:26
> ANTLR Tool version 4.3 used for code generation does not match the current
> runtime version 4.5.3[…]
> 16/10/19 16:46:28 INFO spark.SparkContext: Running Spark version 2.0.0
> […]
> 16/10/19 16:46:29 INFO api.DMLScript: END DML run 10/19/2016 16:46:29
> 16/10/19 16:46:29 ERROR api.DMLScript: Failed to execute DML script.
> org.apache.sysml.runtime.DMLRuntimeException:
> org.apache.sysml.runtime.DMLRuntimeException:
> org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile program
> block hierarchy.
>at org.apache.sysml.runtime.controlprogram.Program.execute
> (Program.java:152)
>at org.apache.sysml.api.DMLScript.execute(DMLScript.java:677)
>at org.apache.sysml.api.DMLScript.executeScript(DMLScript.java:337)
>at org.apache.sysml.api.DMLScript.main(DMLScript.java:192)
>at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
>at sun.reflect.NativeMethodAccessorImpl.invoke
> (NativeMethodAccessorImpl.java:62)
>at sun.reflect.DelegatingMethodAccessorImpl.invoke
> (DelegatingMethodAccessorImpl.java:43)
>at java.lang.reflect.Method.invoke(Method.java:497)
>at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy
> $SparkSubmit$$runMain(SparkSubmit.scala:729)
>at org.apache.spark.deploy.SparkSubmit$.doRunMain$1
> (SparkSubmit.scala:185)
>at org.apache.spark.deploy.SparkSubmit$.submit
> (SparkSubmit.scala:210)
>at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:124)
>at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)
> Caused by: org.apache.sysml.runtime.DMLRuntimeException:
> org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile program
> block hierarchy.
>at
> 

Re: Running the bivar-stats example

[Glenn Weidner]

Hi Sandeep,

Thank you for contacting the Apache SystemML community.  Based on the
following trace fragment:

spark.SparkContext: Running Spark version 2.0.0

it looks like you're running against Spark 2.0 which is not officially
supported with the current master code base.  Can you retry with an earlier
Spark version?  Let me know if I've misinterpreted your Spark environment.

Cheers,
Glenn



From:   "Narayanaswami, Sandeep" 
To: "dev@systemml.incubator.apache.org"

Date:   10/19/2016 05:00 PM
Subject:Running the bivar-stats example



Hi SystemML community,

I’m trying to run the bivariate stats example script on Spark. I’m only
doing a toy example with very small data, but I have been running into
errors.
My script invocation is as follows:

spark-submit --master local[4] SystemML.jar -f
scripts/algorithms/bivar-stats.dml -exec hybrid_spark -nvargs
X=/data/iqtv.csv index1=/data/index1.csv index2=/data/index2.csv
types1=/data/types1.csv types2=/data/types2.csv OUTDIR=/data/outdata

Am I calling the script incorrectly?
(All the files are on hdfs. Iqtv.csv is a headerless csv with two numeric
fields. The index1 and index2 files are text files containing only the
strings ‘1’ and ‘2’ respectively. The types* files are text files each
containing only the string ‘1’.)

Please excuse the rather long stack trace below, but I’m hoping it will
help diagnose the problem.

Cheers,
Sandeep




16/10/19 16:46:26 INFO api.DMLScript: BEGIN DML run 10/19/2016 16:46:26
ANTLR Tool version 4.3 used for code generation does not match the current
runtime version 4.5.3[…]
16/10/19 16:46:28 INFO spark.SparkContext: Running Spark version 2.0.0
[…]
16/10/19 16:46:29 INFO api.DMLScript: END DML run 10/19/2016 16:46:29
16/10/19 16:46:29 ERROR api.DMLScript: Failed to execute DML script.
org.apache.sysml.runtime.DMLRuntimeException:
org.apache.sysml.runtime.DMLRuntimeException:
org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile program
block hierarchy.
   at org.apache.sysml.runtime.controlprogram.Program.execute
(Program.java:152)
   at org.apache.sysml.api.DMLScript.execute(DMLScript.java:677)
   at org.apache.sysml.api.DMLScript.executeScript(DMLScript.java:337)
   at org.apache.sysml.api.DMLScript.main(DMLScript.java:192)
   at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
   at sun.reflect.NativeMethodAccessorImpl.invoke
(NativeMethodAccessorImpl.java:62)
   at sun.reflect.DelegatingMethodAccessorImpl.invoke
(DelegatingMethodAccessorImpl.java:43)
   at java.lang.reflect.Method.invoke(Method.java:497)
   at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy
$SparkSubmit$$runMain(SparkSubmit.scala:729)
   at org.apache.spark.deploy.SparkSubmit$.doRunMain$1
(SparkSubmit.scala:185)
   at org.apache.spark.deploy.SparkSubmit$.submit
(SparkSubmit.scala:210)
   at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:124)
   at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)
Caused by: org.apache.sysml.runtime.DMLRuntimeException:
org.apache.sysml.runtime.DMLRuntimeException: Unable to recompile program
block hierarchy.
   at
org.apache.sysml.runtime.controlprogram.parfor.opt.OptimizationWrapper.optimize
(OptimizationWrapper.java:299)
   at
org.apache.sysml.runtime.controlprogram.parfor.opt.OptimizationWrapper.optimize
(OptimizationWrapper.java:172)
   at
org.apache.sysml.runtime.controlprogram.ParForProgramBlock.execute
(ParForProgramBlock.java:550)
   at org.apache.sysml.runtime.controlprogram.Program.execute
(Program.java:145)
   ... 12 more
Caused by: org.apache.sysml.runtime.DMLRuntimeException: Unable to
recompile program block hierarchy.
   at
org.apache.sysml.hops.recompile.Recompiler.recompileProgramBlockHierarchy
(Recompiler.java:541)
   at
org.apache.sysml.runtime.controlprogram.parfor.opt.OptimizationWrapper.optimize
(OptimizationWrapper.java:278)
   ... 15 more
Caused by: org.apache.sysml.hops.HopsException: ERROR: line 342, column 4
-- error constructing Lops for UnaryOp Hop --

   at org.apache.sysml.hops.UnaryOp.constructLops(UnaryOp.java:177)
   at org.apache.sysml.hops.DataOp.constructLops(DataOp.java:252)
   at org.apache.sysml.hops.recompile.Recompiler.recompileHopsDag
(Recompiler.java:209)
   at org.apache.sysml.hops.recompile.Recompiler.rRecompileProgramBlock
(Recompiler.java:881)
   at org.apache.sysml.hops.recompile.Recompiler.rRecompileProgramBlock
(Recompiler.java:839)
   at org.apache.sysml.hops.recompile.Recompiler.rRecompileProgramBlock
(Recompiler.java:839)
   at org.apache.sysml.hops.recompile.Recompiler.rRecompileProgramBlock
(Recompiler.java:841)
   at