Mentions légales du service
Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
Chameleon
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Package Registry
Container Registry
Operate
Environments
Terraform modules
Monitor
Service Desk
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
solverstack
Chameleon
Commits
568b953b
Commit
568b953b
authored
7 years ago
by
Mathieu Faverge
Browse files
Options
Downloads
Patches
Plain Diff
Remove the unused sustained_peak functions
parent
92767cb8
No related branches found
No related tags found
1 merge request
!83
Cleanup the timing parameters and their documentation
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
runtime/starpu/include/runtime_codelet_profile.h
+1
-43
1 addition, 43 deletions
runtime/starpu/include/runtime_codelet_profile.h
with
1 addition
and
43 deletions
runtime/starpu/include/runtime_codelet_profile.h
+
1
−
43
View file @
568b953b
...
@@ -67,53 +67,11 @@
...
@@ -67,53 +67,11 @@
} \
} \
} \
} \
} \
} \
void estimate_##name##_sustained_peak(double *res_peak) \
{ \
/* We use a heuristic where we assume that all GPUs have some work, and \
* that some CPUs may not have been involved. This may not be \
* applicable to small problems where only a subset of the processing \
* units are used. We assume that all CPUs are the same, so we multiply \
* the best performance obtained on a CPU by the number of CPUs, and we \
* add this to the sum of the performance obtained by the GPUs. */
\
double peak = 0.0; \
double best_cpu = 0.0; \
unsigned ncpus = 0; \
\
unsigned worker; \
for (worker = 0; worker < starpu_worker_get_count(); worker++) \
{ \
unsigned cpu_worker = (starpu_worker_get_type(worker) == STARPU_CPU_WORKER); \
if (cpu_worker) \
ncpus++; \
\
if (name##_perf[worker].n > 0) \
{ \
long n = name##_perf[worker].n; \
double sum = name##_perf[worker].sum; \
double avg = sum / n; \
\
if (cpu_worker) \
{ \
if (avg > best_cpu) \
best_cpu = avg; \
} \
else \
{ \
peak += avg; \
} \
} \
} \
\
peak += ncpus * best_cpu; \
\
*res_peak = peak; \
} \
#define CHAMELEON_CL_CB_HEADER(name) \
#define CHAMELEON_CL_CB_HEADER(name) \
extern struct starpu_perfmodel*cl_##name##_save; \
extern struct starpu_perfmodel*cl_##name##_save; \
extern struct starpu_perfmodel cl_##name##_fake; \
extern struct starpu_perfmodel cl_##name##_fake; \
void cl_##name##_callback(); \
void cl_##name##_callback(); \
void profiling_display_##name##_info(void); \
void profiling_display_##name##_info(void);
void estimate_##name##_sustained_peak(double *res)
#endif
/* __CODELET_PROFILE_H__ */
#endif
/* __CODELET_PROFILE_H__ */
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment