| |
Home • About Us • A-Z Index • Search * Contact Us • Register • Login • Press • ShopThe Open Brand -- Problem Reporting and Interpretations System |
|
Problem Report 2579 Details
Show help | Quick Search | Submit a Test Suite Support Request | Click here to view your privileges
This page provides all information on Problem Report 2579.
Report 2579 Actions
Problem Report Number 2579 Submitter's Classification Test Suite problem State Resolved Resolution Test Suite Deficiency (TSD) Problem Resolution ID TSD.X.1339 Raised 2012-04-24 14:55 Updated 2012-04-24 16:03 Published 2012-04-24 16:03 Product Standard Internationalised System Calls and Libraries Extended V3 (UNIX 03) Certification Program The Open Brand certification program Test Suite VSX5 version 5.3.3 Test Identification C99.os/maths/llrint-tg/T.llrint(f) 1
C99.os/maths/llrint/T.llrint(f) 1
C99.os/maths/lrint-tg/T.lrint(f) 1
C99.os/maths/lrint/T.lrint(f) 1Specification Base Definitions Issue 6 Location in Spec none Problem Summary Conversion from long long int to double causes overflow Problem Text Take one case (/tset/C99.os/maths/llrint/T.llrint 1)for example.
The source codes of this case is in
C99.os/maths/llrint/llrint.c.
Between line 238 to 241, we get the input range of this
function(llrint). Because the double_max is larger than the
long_long_int_max, so we get:
high = (double)(LLONG_MAX - 1);
low = (double)(LLONG_MIN - 1);
It seems quite right, but the conversion from long long int to
double causes overflow. It is not accurate. In 64bit system
LLONG_MAX ? 1 is +9223372036854775806(2^63 -2), but after
conversion, high becomes +9223372036854775808.000000(2^63),
beyong the range of long long int.
We think that the conversion from long long int to double may
cause precision lose, and those test cases failed since got the
wrong input range.Test Output
*C99.os/maths/llrint-tg/T.llrint 1*
400|1436 1 1 00:15:23|IC Start
200|1436 1 00:15:23|TP Start
520|1436 1 00023501 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1436 1 00023501 1 2|The result was too large 21 times
520|1436 1 00023501 1 3| equal 7973 times
520|1436 1 00023501 1 4| too small 6 times
520|1436 1 00023501 1 5|The maximum relative error of 196261 occured for
value 9.22337e+18
520|1436 1 00023501 1 6|This gave a maximum loss of 71 significant
digits of base 2
520|1436 1 00023501 1 7|The maximum acceptable loss is 4 significant digits
520|1436 1 00023501 1 8|The root-mean-square relative error is 7918.04
520|1436 1 00023501 1 9|This gave an average loss of 66 significant
digits of base 2
520|1436 1 00023501 1 10|The maximum acceptable loss is 2 significant
digits
220|1436 1 1 00:15:23|FAIL
410|1436 1 1 00:15:23|IC End
*C99.os/maths/llrint-tg/T.llrintf 1*
400|1437 1 1 00:15:24|IC Start
200|1437 1 00:15:24|TP Start
520|1437 1 00023507 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1437 1 00023507 1 2|The result was too large 20 times
520|1437 1 00023507 1 3| equal 7971 times
520|1437 1 00023507 1 4| too small 9 times
520|1437 1 00023507 1 5|The maximum relative error of 196419 occured for
value 9.22337e+18
520|1437 1 00023507 1 6|This gave a maximum loss of 42 significant
digits of base 2
520|1437 1 00023507 1 7|The maximum acceptable loss is 4 significant digits
520|1437 1 00023507 1 8|The root-mean-square relative error is 7923.78
520|1437 1 00023507 1 9|This gave an average loss of 37 significant
digits of base 2
520|1437 1 00023507 1 10|The maximum acceptable loss is 2 significant digits
220|1437 1 1 00:15:24|FAIL
410|1437 1 1 00:15:24|IC End
*C99.os/maths/llrint/T.llrint 1*
400|1439 1 1 00:15:26|IC Start
200|1439 1 00:15:26|TP Start
520|1439 1 00023519 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1439 1 00023519 1 2|The result was too large 21 times
520|1439 1 00023519 1 3| equal 7973 times
520|1439 1 00023519 1 4| too small 6 times
520|1439 1 00023519 1 5|The maximum relative error of 193313 occured for
value 9.22337e+18
520|1439 1 00023519 1 6|This gave a maximum loss of 71 significant
digits of base 2
520|1439 1 00023519 1 7|The maximum acceptable loss is 4 significant digits
520|1439 1 00023519 1 8|The root-mean-square relative error is 7799.12
520|1439 1 00023519 1 9|This gave an average loss of 66 significant
digits of base 2
520|1439 1 00023519 1 10|The maximum acceptable loss is 2 significant digits
220|1439 1 1 00:15:26|FAIL
410|1439 1 1 00:15:26|IC End
*C99.os/maths/llrint/T.llrintf 1*
400|1440 1 1 00:15:27|IC Start
200|1440 1 00:15:27|TP Start
520|1440 1 00023525 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1440 1 00023525 1 2|The result was too large 20 times
520|1440 1 00023525 1 3| equal 7971 times
520|1440 1 00023525 1 4| too small 9 times
520|1440 1 00023525 1 5|The maximum relative error of 194259 occured for
value 9.22337e+18
520|1440 1 00023525 1 6|This gave a maximum loss of 42 significant
digits of base 2
520|1440 1 00023525 1 7|The maximum acceptable loss is 4 significant digits
520|1440 1 00023525 1 8|The root-mean-square relative error is 7836.62
520|1440 1 00023525 1 9|This gave an average loss of 37 significant
digits of base 2
520|1440 1 00023525 1 10|The maximum acceptable loss is 2 significant digits
220|1440 1 1 00:15:27|FAIL
410|1440 1 1 00:15:27|IC End
*C99.os/maths/lrint-tg/T.lrint 1*
400|1474 1 1 00:16:01|IC Start
200|1474 1 00:16:01|TP Start
520|1474 1 00023796 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1474 1 00023796 1 2|The result was too large 21 times
520|1474 1 00023796 1 3| equal 7973 times
520|1474 1 00023796 1 4| too small 6 times
520|1474 1 00023796 1 5|The maximum relative error of 194944 occured for
value 9.22337e+18
520|1474 1 00023796 1 6|This gave a maximum loss of 71 significant
digits of base 2
520|1474 1 00023796 1 7|The maximum acceptable loss is 4 significant digits
520|1474 1 00023796 1 8|The root-mean-square relative error is 7864.9
520|1474 1 00023796 1 9|This gave an average loss of 66 significant
digits of base 2
520|1474 1 00023796 1 10|The maximum acceptable loss is 2 significant digits
220|1474 1 1 00:16:01|FAIL
*C99.os/maths/lrint-tg/T.lrintf 1*
400|1475 1 1 00:16:02|IC Start
200|1475 1 00:16:02|TP Start
520|1475 1 00023802 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1475 1 00023802 1 2|The result was too large 20 times
520|1475 1 00023802 1 3| equal 7971 times
520|1475 1 00023802 1 4| too small 9 times
520|1475 1 00023802 1 5|The maximum relative error of 196063 occured for
value 9.22337e+18
520|1475 1 00023802 1 6|This gave a maximum loss of 42 significant
digits of base 2
520|1475 1 00023802 1 7|The maximum acceptable loss is 4 significant digits
520|1475 1 00023802 1 8|The root-mean-square relative error is 7909.42
520|1475 1 00023802 1 9|This gave an average loss of 37 significant
digits of base 2
520|1475 1 00023802 1 10|The maximum acceptable loss is 2 significant digits
220|1475 1 1 00:16:02|FAIL
410|1475 1 1 00:16:02|IC End
*C99.os/maths/lrint/T.lrint 1*
400|1477 1 1 00:16:04|IC Start
200|1477 1 00:16:04|TP Start
520|1477 1 00023814 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1477 1 00023814 1 2|The result was too large 21 times
520|1477 1 00023814 1 3| equal 7973 times
520|1477 1 00023814 1 4| too small 6 times
520|1477 1 00023814 1 5|The maximum relative error of 194818 occured for
value 9.22337e+18
520|1477 1 00023814 1 6|This gave a maximum loss of 71 significant
digits of base 2
520|1477 1 00023814 1 7|The maximum acceptable loss is 4 significant digits
520|1477 1 00023814 1 8|The root-mean-square relative error is 7859.83
520|1477 1 00023814 1 9|This gave an average loss of 66 significant
digits of base 2
520|1477 1 00023814 1 10|The maximum acceptable loss is 2 significant digits
220|1477 1 1 00:16:04|FAIL
410|1477 1 1 00:16:04|IC End
*C99.os/maths/lrint/T.lrintf 1*
400|1478 1 1 00:16:05|IC Start
200|1478 1 00:16:05|TP Start
520|1478 1 00023820 1 1|Random arguments were tested from the interval
[-9.22337e+18, 9.22337e+18]
520|1478 1 00023820 1 2|The result was too large 20 times
520|1478 1 00023820 1 3| equal 7971 times
520|1478 1 00023820 1 4| too small 9 times
520|1478 1 00023820 1 5|The maximum relative error of 192317 occured for
value 9.22337e+18
520|1478 1 00023820 1 6|This gave a maximum loss of 42 significant
digits of base 2
520|1478 1 00023820 1 7|The maximum acceptable loss is 4 significant digits
520|1478 1 00023820 1 8|The root-mean-square relative error is 7758.29
520|1478 1 00023820 1 9|This gave an average loss of 37 significant
digits of base 2
520|1478 1 00023820 1 10|The maximum acceptable loss is 2 significant digits
220|1478 1 1 00:16:05|FAILReview Information
Review Type TSMA Review Start Date 2012-04-24 14:55 Last Updated 2012-04-24 12:26 Completed 2012-04-24 12:26 Status Complete Review Recommendation Test Suite Deficiency (TSD) Review Response This is accepted as a fault in the test suite.
Review Type SA Review Start Date 2012-04-24 20:26 Last Updated 2012-04-24 16:00 Completed 2012-04-24 16:00 Status Complete Review Resolution Test Suite Deficiency (TSD) Review Conclusion A test suite deficiency is granted.
Problem Reporting System Options:
- View Report 2579
- List All PRs
- Search Reports
- Email the System Administrator
- View the The Open Brand Interpretations Database User Manual
Contact the Certification Authority