Comments (4)
@danielschemmel @ccadar @afd This is going to cause real problems when we replay and confirm behaviour. Although because this is a benchmark that does nothing symbolic we won't actually get any test cases which makes things a little tricky.
from fp-benchmarks-aachen.
@delcypher Yeah, we very explicitly built this test to reflect some... let's just say unused behavior
We should probably discuss this benchmark in a little more detail on friday
from fp-benchmarks-aachen.
Yeah I've hit even more issues with this test case. I've mostly fixed our version of KLEE to work with this benchmark but there are some assertions that are simply not going to pass.
This test case illustrates the problem further.
// RUN: %llvmgcc %s -O0 -g -o %t1_native
// RUN: %t1_native
// RUN: %llvmgcc %s -emit-llvm -O0 -g -c -o %t1.bc
// RUN: rm -rf %t.klee-out
// RUN: %klee --output-dir=%t.klee-out --exit-on-error %t1.bc
// REQUIRES: x86_64
#include <assert.h>
#include <inttypes.h>
#include <math.h>
#include <stdint.h>
#include <stdio.h>
#include <string.h>
void dump_long_double(long double ld) {
uint64_t data[2] = {0, 0};
memcpy(data, &ld, 10);
printf("a: 0x%.4" PRIx16 " %.16" PRIx64 "\n", (uint16_t)data[1], data[0]);
}
int main() {
// In "8.2.2 Unsupported Double
// Extended-Precision Float-Point Encodings and Pseudo-Denormals" from the
// Intel(R) 64 and IA-32 Architectures Software Developer's Manual a list
// of several classes of encodings are listed that are no longer considered
// valid. Below is an "unnormal" which is an invalid operand which should
// not compare equal to itself.
//
// It should raise the IEEE754 invalid operation exception too but we
// don't model these in KLEE right now.
//
// Significand
// Sign Exponent Integer Fraction
// [1] [100000000000001] [0] [00100000...]
const uint64_t lowBits = 0x2000000000000000;
const uint16_t highBits = 0xc001;
long double v = 0.0l;
memcpy(&v, &lowBits, sizeof(lowBits)); // 64 bits
memcpy(((uint8_t *)(&v)) + 8, &highBits, 2); // 16 bits
dump_long_double(v);
long double x = 1.0l;
// TODO: Find out where this behaviour is documented.
assert(isnan((float)v));
assert(isnan((double)v));
// Check signed casts
assert(((int8_t)v) == 0);
// FIXME: This assert below doesn't work properly. When we compile this
// program natively with Clang the casted value is zero and when we compile
// natively with gcc is -32768. We should report this!
int16_t temp = (int16_t)v;
assert((temp == 0) | (temp == -32768));
assert(((int32_t)v) == -2147483648);
assert(((int64_t)v) == INT64_MIN);
// Check signed casts
assert(((uint8_t)v) == 0);
assert(((uint16_t)v) == 0);
assert(((uint32_t)v) == 0);
uint64_t temp2 = (uint64_t)v;
// FIXME: This assert below doesn't work properly. When we compile this
// program natively with Clang the casted value is 0 and when we compile
// natively with gcc it is 0x8000000000000000. We should report this!
assert(temp2 == 0 || temp2 == 0x8000000000000000);
}
from fp-benchmarks-aachen.
Fixed by #17
from fp-benchmarks-aachen.
Related Issues (9)
- The specification for `real/sorting_doubles` is incorrect. HOT 4
- Mistakes in `aachen/syn/sqr/spec.yml` HOT 4
- Ninja & GMP HOT 2
- `ld-unnormal_default.x86_64` is mis-compiled by Clang when optimizations are enabled HOT 1
- syn/atof_default.x86_64 is incorrect despite the spec saying it should be correct HOT 2
- sqr_ family of benchmarks incorrect HOT 2
- gmp_klee_inv_arg.x86_64 crashes KLEE by calling `raise()` HOT 1
- klee_assume(f == f) is not a good way to ignore paths where symbolics are NaN HOT 4
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from fp-benchmarks-aachen.