View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0000322 | OpenFOAM | Bug | public | 2011-10-22 15:46 | 2012-01-05 10:33 |
Reporter | Assigned To | ||||
Priority | normal | Severity | major | Reproducibility | always |
Status | closed | Resolution | no change required | ||
Summary | 0000322: ICC Compiler Flags Incorrect | ||||
Description | The compiler flags used for ICC compilation are currently set that they will compile programs that are unusable by AMD processors. | ||||
Additional Information | /home/coffeewave/a/whitsonl/.OpenFOAM/OpenFOAM-2.0.x/wmake/rules/linux64Icc/c++Opt should be edited to remove the -SSE3 option. | ||||
Tags | No tags attached. | ||||
|
We chose the options for the Intel compiler to optimise for the Intel processors but even with these solver still run faster when compiled with gcc. Given that Icc was written for the Intel processors rather than AMD should we select default options that favor AMD over Intel? Are there option on Icc which create executables for the AMD processors which are faster than the best gcc can do? |
|
There are various indications that one _should not_ use Icc for AMD processors - one indication: http://en.wikipedia.org/wiki/Intel_C%2B%2B_Compiler#Criticism According to reference #19 in that wiki page: http://developer.amd.com/Assets/CompilerOptQuickRef-61004100.pdf - aside from Icc, it seems that Gcc, Open64 and PGI are compilers that AMD "knows" optimization flags for their CPUs. |
|
orphaned report - closing |
Date Modified | Username | Field | Change |
---|---|---|---|
2011-10-22 15:46 |
|
New Issue | |
2011-10-22 16:11 | henry | Note Added: 0000741 | |
2011-10-22 16:22 | wyldckat | Note Added: 0000743 | |
2012-01-05 10:33 |
|
Note Added: 0000899 | |
2012-01-05 10:33 |
|
Status | new => closed |
2012-01-05 10:33 |
|
Assigned To | => user2 |
2012-01-05 10:33 |
|
Resolution | open => no change required |