[llvm-dev] [Help] Problem on running SPECInt 2006 with LLVM

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[llvm-dev] [Help] Problem on running SPECInt 2006 with LLVM

David Jones via llvm-dev
Hi, all. I'm trying to analyze SPECInt 2006 with LLVM and face errors on some programs. 
I modified the configuration in makefile tree so that each source code generates its byte code instead of an object file. 
Then, I combined all the generated byte codes into a single byte code and run it directly with lli. 
The approach works with several benchmarks (e.g. bzip) but not all of them. 
Certain benchmarks like "471.omnetpp" or "483.xalancbmk" halt their execution with segmentation fault. Interestingly, both benchmarks are implemented C++. 

Do I need to add certain optimization flag? 
I'm only using the flags as follows: -static -emit-llvm -O2

Any advice will be a great help.
Thank you!
--
Best, Sung

_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
Reply | Threaded
Open this post in threaded view
|

Re: [llvm-dev] [Help] Problem on running SPECInt 2006 with LLVM

David Jones via llvm-dev

Hi Sung-

 

Have a look at https://www.spec.org/cpu2017/results/res2018q3/cpu2017-20180626-07434.pdf to see how llvm can be used to run SPEC – the listed page is for CPU2017 but it is similar to how you compile and run CPU2006.

 

-Thx

dibyendu

 

From: llvm-dev [mailto:[hidden email]] On Behalf Of Sunghyun Park via llvm-dev
Sent: Tuesday, October 23, 2018 9:23 AM
To: LLVM Dev <[hidden email]>
Subject: [llvm-dev] [Help] Problem on running SPECInt 2006 with LLVM

 

Hi, all. I'm trying to analyze SPECInt 2006 with LLVM and face errors on some programs. 

I modified the configuration in makefile tree so that each source code generates its byte code instead of an object file. 

Then, I combined all the generated byte codes into a single byte code and run it directly with lli. 

The approach works with several benchmarks (e.g. bzip) but not all of them. 

Certain benchmarks like "471.omnetpp" or "483.xalancbmk" halt their execution with segmentation fault. Interestingly, both benchmarks are implemented C++. 

 

Do I need to add certain optimization flag? 

I'm only using the flags as follows: -static -emit-llvm -O2

 

Any advice will be a great help.

Thank you!

--

Best, Sung


_______________________________________________
LLVM Developers mailing list
[hidden email]
http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev