feat: add per-process and per-photon G4 timing instrumentation#280
Open
ggalgoczi wants to merge 5 commits into
Open
feat: add per-process and per-photon G4 timing instrumentation#280ggalgoczi wants to merge 5 commits into
ggalgoczi wants to merge 5 commits into
Conversation
Add optical photon performance profiling to GPURaytrace: - Per-process step timing (Transportation, OpWLS, OpRayleigh, OpAbsorption) using std::chrono::steady_clock between consecutive SteppingAction calls - Per-photon lifetime timing with median, percentiles, and histogram - Per-photon step count statistics (avg, median, p10/p50/p90/p99) - --skip-gpu flag to measure G4-only photon propagation without GPU Add benchmark script (examples/photontiming_geant4/photontimingandsteps.sh) that correctly measures GPU vs G4 speedup using three runs: 1. G4 with photons, no GPU (--skip-gpu) 2. G4 baseline with setStackPhotons false 3. Normal GPU run Measured on apex.gdml (10 MeV electron, ~250k photons, RTX 4090): - Transportation: 0.77 us/step, 114M steps, 99.8% of total time - OpWLS: 0.85 us/invocation, 42k invocations, 0.07% of total - Median photon: 2.1 us (UV, exits world quickly) - Average photon: 306 us (skewed by 4% WLS-converted tail at 1-10ms) - Speedup: ~1,400x (single-threaded G4 vs GPU)
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Cpp-linter Review
Used clang-format v20.1.2
Click here for the full clang-format patch
diff --git a/src/GPURaytrace.cpp b/src/GPURaytrace.cpp
index 64386ca..be1dd60 100644
--- a/src/GPURaytrace.cpp
+++ b/src/GPURaytrace.cpp
@@ -78,3 +78 @@ int main(int argc, char **argv)
- program.add_argument("--skip-gpu")
- .help("skip GPU photon propagation (for measuring G4-only photon time)")
- .flag();
+ program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag();
diff --git a/src/GPURaytrace.h b/src/GPURaytrace.h
index b61f0d6..ccad899 100644
--- a/src/GPURaytrace.h
+++ b/src/GPURaytrace.h
@@ -483,4 +483,20 @@ struct SteppingAction : G4UserSteppingAction
- if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; }
- else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; }
- else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; }
- else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; }
+ if (pname == "Transportation")
+ {
+ fTimeTransport += dt;
+ fCountTransport++;
+ }
+ else if (pname == "OpWLS")
+ {
+ fTimeOpWLS += dt;
+ fCountOpWLS++;
+ }
+ else if (pname == "OpRayleigh")
+ {
+ fTimeOpRayleigh += dt;
+ fCountOpRayleigh++;
+ }
+ else if (pname == "OpAbsorption")
+ {
+ fTimeOpAbsorption += dt;
+ fCountOpAbsorption++;
+ }
@@ -507 +523,2 @@ struct SteppingAction : G4UserSteppingAction
- if (fSkipGenstep) return; // skip genstep collection for timing-only runs
+ if (fSkipGenstep)
+ return; // skip genstep collection for timing-only runs
@@ -637 +654,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -639 +658,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -641,6 +662,12 @@ struct TrackingAction : G4UserTrackingAction
- if (dt < 1000) fTimeBucket0++;
- else if (dt < 10000) fTimeBucket1++;
- else if (dt < 100000) fTimeBucket2++;
- else if (dt < 1000000) fTimeBucket3++;
- else if (dt < 10000000) fTimeBucket4++;
- else fTimeBucket5++;
+ if (dt < 1000)
+ fTimeBucket0++;
+ else if (dt < 10000)
+ fTimeBucket1++;
+ else if (dt < 100000)
+ fTimeBucket2++;
+ else if (dt < 1000000)
+ fTimeBucket3++;
+ else if (dt < 10000000)
+ fTimeBucket4++;
+ else
+ fTimeBucket5++;
@@ -649 +676,5 @@ struct TrackingAction : G4UserTrackingAction
- { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); }
+ {
+ std::lock_guard<std::mutex> lock(fTimesMutex);
+ fAllTimes.push_back(dt);
+ fAllSteps.push_back(nsteps);
+ }
@@ -656 +687,2 @@ struct TrackingAction : G4UserTrackingAction
- if (n == 0) return;
+ if (n == 0)
+ return;
@@ -691 +723,2 @@ struct TrackingAction : G4UserTrackingAction
- for (int s : fAllSteps) step_sum += s;
+ for (int s : fAllSteps)
+ step_sum += s;
@@ -699,4 +732,2 @@ struct TrackingAction : G4UserTrackingAction
- << "p10=" << fAllSteps[ssz / 10]
- << ", p50=" << fAllSteps[ssz / 2]
- << ", p90=" << fAllSteps[ssz * 9 / 10]
- << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
+ << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2]
+ << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
@@ -718,6 +749,4 @@ inline void RunAction::PrintTimingReport()
- std::cout << "Geant4: StepTime " << std::setw(15) << name
- << ": count=" << std::setw(10) << count
- << " avg=" << std::fixed << std::setprecision(2) << std::setw(8)
- << total_ns / 1000.0 / count << " us"
- << " total=" << std::setprecision(3) << std::setw(8)
- << total_ns / 1e9 << " s" << std::endl;
+ std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count
+ << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count
+ << " us"
+ << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl;
Have any feedback or feature suggestions? Share it here.
Comment on lines
+78
to
+80
| program.add_argument("--skip-gpu") | ||
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | ||
| .flag(); |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| program.add_argument("--skip-gpu") | |
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | |
| .flag(); | |
| program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag(); |
Comment on lines
+483
to
+486
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | ||
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | ||
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | ||
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | |
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | |
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | |
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } | |
| if (pname == "Transportation") | |
| { | |
| fTimeTransport += dt; | |
| fCountTransport++; | |
| } | |
| else if (pname == "OpWLS") | |
| { | |
| fTimeOpWLS += dt; | |
| fCountOpWLS++; | |
| } | |
| else if (pname == "OpRayleigh") | |
| { | |
| fTimeOpRayleigh += dt; | |
| fCountOpRayleigh++; | |
| } | |
| else if (pname == "OpAbsorption") | |
| { | |
| fTimeOpAbsorption += dt; | |
| fCountOpAbsorption++; | |
| } |
| } | ||
| } | ||
|
|
||
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs | |
| if (fSkipGenstep) | |
| return; // skip genstep collection for timing-only runs |
| fPhotonCount++; | ||
|
|
||
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | ||
| cur = fMaxPhotonTime.load(); | ||
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| else fTimeBucket5++; | ||
|
|
||
| int nsteps = track->GetCurrentStepNumber(); | ||
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } | |
| { | |
| std::lock_guard<std::mutex> lock(fTimesMutex); | |
| fAllTimes.push_back(dt); | |
| fAllSteps.push_back(nsteps); | |
| } |
| void PrintPhotonTiming() | ||
| { | ||
| int n = fPhotonCount.load(); | ||
| if (n == 0) return; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (n == 0) return; | |
| if (n == 0) | |
| return; |
| if (ssz > 0) | ||
| { | ||
| long long step_sum = 0; | ||
| for (int s : fAllSteps) step_sum += s; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| for (int s : fAllSteps) step_sum += s; | |
| for (int s : fAllSteps) | |
| step_sum += s; |
Comment on lines
+699
to
+702
| << "p10=" << fAllSteps[ssz / 10] | ||
| << ", p50=" << fAllSteps[ssz / 2] | ||
| << ", p90=" << fAllSteps[ssz * 9 / 10] | ||
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| << "p10=" << fAllSteps[ssz / 10] | |
| << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] | |
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; | |
| << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Comment on lines
+718
to
+723
| std::cout << "Geant4: StepTime " << std::setw(15) << name | ||
| << ": count=" << std::setw(10) << count | ||
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | ||
| << total_ns / 1000.0 / count << " us" | ||
| << " total=" << std::setprecision(3) << std::setw(8) | ||
| << total_ns / 1e9 << " s" << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| std::cout << "Geant4: StepTime " << std::setw(15) << name | |
| << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | |
| << total_ns / 1000.0 / count << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) | |
| << total_ns / 1e9 << " s" << std::endl; | |
| std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count | |
| << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl; |
The genstep collection (Cerenkov/Scintillation CollectGenstep calls and associated mutex locks) adds overhead to the G4-only timing run. Guard it with fSkipGenstep flag so Run 1 measures pure G4 photon propagation without GPU-related bookkeeping.
ggalgoczi
force-pushed
the
photontiming_geant4
branch
from
b799d72 to
9bfd6fa
Compare
Contributor
Cpp-Linter Report
|
Contributor
There was a problem hiding this comment.
Cpp-linter Review
Used clang-format v20.1.2
Click here for the full clang-format patch
diff --git a/src/GPURaytrace.cpp b/src/GPURaytrace.cpp
index 64386ca..be1dd60 100644
--- a/src/GPURaytrace.cpp
+++ b/src/GPURaytrace.cpp
@@ -78,3 +78 @@ int main(int argc, char **argv)
- program.add_argument("--skip-gpu")
- .help("skip GPU photon propagation (for measuring G4-only photon time)")
- .flag();
+ program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag();
diff --git a/src/GPURaytrace.h b/src/GPURaytrace.h
index b61f0d6..ccad899 100644
--- a/src/GPURaytrace.h
+++ b/src/GPURaytrace.h
@@ -483,4 +483,20 @@ struct SteppingAction : G4UserSteppingAction
- if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; }
- else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; }
- else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; }
- else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; }
+ if (pname == "Transportation")
+ {
+ fTimeTransport += dt;
+ fCountTransport++;
+ }
+ else if (pname == "OpWLS")
+ {
+ fTimeOpWLS += dt;
+ fCountOpWLS++;
+ }
+ else if (pname == "OpRayleigh")
+ {
+ fTimeOpRayleigh += dt;
+ fCountOpRayleigh++;
+ }
+ else if (pname == "OpAbsorption")
+ {
+ fTimeOpAbsorption += dt;
+ fCountOpAbsorption++;
+ }
@@ -507 +523,2 @@ struct SteppingAction : G4UserSteppingAction
- if (fSkipGenstep) return; // skip genstep collection for timing-only runs
+ if (fSkipGenstep)
+ return; // skip genstep collection for timing-only runs
@@ -637 +654,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -639 +658,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -641,6 +662,12 @@ struct TrackingAction : G4UserTrackingAction
- if (dt < 1000) fTimeBucket0++;
- else if (dt < 10000) fTimeBucket1++;
- else if (dt < 100000) fTimeBucket2++;
- else if (dt < 1000000) fTimeBucket3++;
- else if (dt < 10000000) fTimeBucket4++;
- else fTimeBucket5++;
+ if (dt < 1000)
+ fTimeBucket0++;
+ else if (dt < 10000)
+ fTimeBucket1++;
+ else if (dt < 100000)
+ fTimeBucket2++;
+ else if (dt < 1000000)
+ fTimeBucket3++;
+ else if (dt < 10000000)
+ fTimeBucket4++;
+ else
+ fTimeBucket5++;
@@ -649 +676,5 @@ struct TrackingAction : G4UserTrackingAction
- { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); }
+ {
+ std::lock_guard<std::mutex> lock(fTimesMutex);
+ fAllTimes.push_back(dt);
+ fAllSteps.push_back(nsteps);
+ }
@@ -656 +687,2 @@ struct TrackingAction : G4UserTrackingAction
- if (n == 0) return;
+ if (n == 0)
+ return;
@@ -691 +723,2 @@ struct TrackingAction : G4UserTrackingAction
- for (int s : fAllSteps) step_sum += s;
+ for (int s : fAllSteps)
+ step_sum += s;
@@ -699,4 +732,2 @@ struct TrackingAction : G4UserTrackingAction
- << "p10=" << fAllSteps[ssz / 10]
- << ", p50=" << fAllSteps[ssz / 2]
- << ", p90=" << fAllSteps[ssz * 9 / 10]
- << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
+ << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2]
+ << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
@@ -718,6 +749,4 @@ inline void RunAction::PrintTimingReport()
- std::cout << "Geant4: StepTime " << std::setw(15) << name
- << ": count=" << std::setw(10) << count
- << " avg=" << std::fixed << std::setprecision(2) << std::setw(8)
- << total_ns / 1000.0 / count << " us"
- << " total=" << std::setprecision(3) << std::setw(8)
- << total_ns / 1e9 << " s" << std::endl;
+ std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count
+ << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count
+ << " us"
+ << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl;
Have any feedback or feature suggestions? Share it here.
Comment on lines
+78
to
+80
| program.add_argument("--skip-gpu") | ||
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | ||
| .flag(); |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| program.add_argument("--skip-gpu") | |
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | |
| .flag(); | |
| program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag(); |
Comment on lines
+483
to
+486
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | ||
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | ||
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | ||
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | |
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | |
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | |
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } | |
| if (pname == "Transportation") | |
| { | |
| fTimeTransport += dt; | |
| fCountTransport++; | |
| } | |
| else if (pname == "OpWLS") | |
| { | |
| fTimeOpWLS += dt; | |
| fCountOpWLS++; | |
| } | |
| else if (pname == "OpRayleigh") | |
| { | |
| fTimeOpRayleigh += dt; | |
| fCountOpRayleigh++; | |
| } | |
| else if (pname == "OpAbsorption") | |
| { | |
| fTimeOpAbsorption += dt; | |
| fCountOpAbsorption++; | |
| } |
| } | ||
| } | ||
|
|
||
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs | |
| if (fSkipGenstep) | |
| return; // skip genstep collection for timing-only runs |
| fPhotonCount++; | ||
|
|
||
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | ||
| cur = fMaxPhotonTime.load(); | ||
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| else fTimeBucket5++; | ||
|
|
||
| int nsteps = track->GetCurrentStepNumber(); | ||
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } | |
| { | |
| std::lock_guard<std::mutex> lock(fTimesMutex); | |
| fAllTimes.push_back(dt); | |
| fAllSteps.push_back(nsteps); | |
| } |
| void PrintPhotonTiming() | ||
| { | ||
| int n = fPhotonCount.load(); | ||
| if (n == 0) return; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (n == 0) return; | |
| if (n == 0) | |
| return; |
| if (ssz > 0) | ||
| { | ||
| long long step_sum = 0; | ||
| for (int s : fAllSteps) step_sum += s; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| for (int s : fAllSteps) step_sum += s; | |
| for (int s : fAllSteps) | |
| step_sum += s; |
Comment on lines
+699
to
+702
| << "p10=" << fAllSteps[ssz / 10] | ||
| << ", p50=" << fAllSteps[ssz / 2] | ||
| << ", p90=" << fAllSteps[ssz * 9 / 10] | ||
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| << "p10=" << fAllSteps[ssz / 10] | |
| << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] | |
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; | |
| << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Comment on lines
+718
to
+723
| std::cout << "Geant4: StepTime " << std::setw(15) << name | ||
| << ": count=" << std::setw(10) << count | ||
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | ||
| << total_ns / 1000.0 / count << " us" | ||
| << " total=" << std::setprecision(3) << std::setw(8) | ||
| << total_ns / 1e9 << " s" << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| std::cout << "Geant4: StepTime " << std::setw(15) << name | |
| << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | |
| << total_ns / 1000.0 / count << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) | |
| << total_ns / 1e9 << " s" << std::endl; | |
| std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count | |
| << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl; |
Contributor
There was a problem hiding this comment.
Cpp-linter Review
Used clang-format v20.1.2
Click here for the full clang-format patch
diff --git a/src/GPURaytrace.cpp b/src/GPURaytrace.cpp
index 64386ca..be1dd60 100644
--- a/src/GPURaytrace.cpp
+++ b/src/GPURaytrace.cpp
@@ -78,3 +78 @@ int main(int argc, char **argv)
- program.add_argument("--skip-gpu")
- .help("skip GPU photon propagation (for measuring G4-only photon time)")
- .flag();
+ program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag();
diff --git a/src/GPURaytrace.h b/src/GPURaytrace.h
index b61f0d6..ccad899 100644
--- a/src/GPURaytrace.h
+++ b/src/GPURaytrace.h
@@ -483,4 +483,20 @@ struct SteppingAction : G4UserSteppingAction
- if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; }
- else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; }
- else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; }
- else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; }
+ if (pname == "Transportation")
+ {
+ fTimeTransport += dt;
+ fCountTransport++;
+ }
+ else if (pname == "OpWLS")
+ {
+ fTimeOpWLS += dt;
+ fCountOpWLS++;
+ }
+ else if (pname == "OpRayleigh")
+ {
+ fTimeOpRayleigh += dt;
+ fCountOpRayleigh++;
+ }
+ else if (pname == "OpAbsorption")
+ {
+ fTimeOpAbsorption += dt;
+ fCountOpAbsorption++;
+ }
@@ -507 +523,2 @@ struct SteppingAction : G4UserSteppingAction
- if (fSkipGenstep) return; // skip genstep collection for timing-only runs
+ if (fSkipGenstep)
+ return; // skip genstep collection for timing-only runs
@@ -637 +654,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -639 +658,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -641,6 +662,12 @@ struct TrackingAction : G4UserTrackingAction
- if (dt < 1000) fTimeBucket0++;
- else if (dt < 10000) fTimeBucket1++;
- else if (dt < 100000) fTimeBucket2++;
- else if (dt < 1000000) fTimeBucket3++;
- else if (dt < 10000000) fTimeBucket4++;
- else fTimeBucket5++;
+ if (dt < 1000)
+ fTimeBucket0++;
+ else if (dt < 10000)
+ fTimeBucket1++;
+ else if (dt < 100000)
+ fTimeBucket2++;
+ else if (dt < 1000000)
+ fTimeBucket3++;
+ else if (dt < 10000000)
+ fTimeBucket4++;
+ else
+ fTimeBucket5++;
@@ -649 +676,5 @@ struct TrackingAction : G4UserTrackingAction
- { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); }
+ {
+ std::lock_guard<std::mutex> lock(fTimesMutex);
+ fAllTimes.push_back(dt);
+ fAllSteps.push_back(nsteps);
+ }
@@ -656 +687,2 @@ struct TrackingAction : G4UserTrackingAction
- if (n == 0) return;
+ if (n == 0)
+ return;
@@ -691 +723,2 @@ struct TrackingAction : G4UserTrackingAction
- for (int s : fAllSteps) step_sum += s;
+ for (int s : fAllSteps)
+ step_sum += s;
@@ -699,4 +732,2 @@ struct TrackingAction : G4UserTrackingAction
- << "p10=" << fAllSteps[ssz / 10]
- << ", p50=" << fAllSteps[ssz / 2]
- << ", p90=" << fAllSteps[ssz * 9 / 10]
- << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
+ << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2]
+ << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
@@ -718,6 +749,4 @@ inline void RunAction::PrintTimingReport()
- std::cout << "Geant4: StepTime " << std::setw(15) << name
- << ": count=" << std::setw(10) << count
- << " avg=" << std::fixed << std::setprecision(2) << std::setw(8)
- << total_ns / 1000.0 / count << " us"
- << " total=" << std::setprecision(3) << std::setw(8)
- << total_ns / 1e9 << " s" << std::endl;
+ std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count
+ << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count
+ << " us"
+ << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl;
Have any feedback or feature suggestions? Share it here.
Comment on lines
+78
to
+80
| program.add_argument("--skip-gpu") | ||
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | ||
| .flag(); |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| program.add_argument("--skip-gpu") | |
| .help("skip GPU photon propagation (for measuring G4-only photon time)") | |
| .flag(); | |
| program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag(); |
Comment on lines
+483
to
+486
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | ||
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | ||
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | ||
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; } | |
| else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; } | |
| else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; } | |
| else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; } | |
| if (pname == "Transportation") | |
| { | |
| fTimeTransport += dt; | |
| fCountTransport++; | |
| } | |
| else if (pname == "OpWLS") | |
| { | |
| fTimeOpWLS += dt; | |
| fCountOpWLS++; | |
| } | |
| else if (pname == "OpRayleigh") | |
| { | |
| fTimeOpRayleigh += dt; | |
| fCountOpRayleigh++; | |
| } | |
| else if (pname == "OpAbsorption") | |
| { | |
| fTimeOpAbsorption += dt; | |
| fCountOpAbsorption++; | |
| } |
| } | ||
| } | ||
|
|
||
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (fSkipGenstep) return; // skip genstep collection for timing-only runs | |
| if (fSkipGenstep) | |
| return; // skip genstep collection for timing-only runs |
| fPhotonCount++; | ||
|
|
||
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| long long cur = fMinPhotonTime.load(); | ||
| while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {} | ||
| cur = fMaxPhotonTime.load(); | ||
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {} | |
| while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) | |
| { | |
| } |
| else fTimeBucket5++; | ||
|
|
||
| int nsteps = track->GetCurrentStepNumber(); | ||
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); } | |
| { | |
| std::lock_guard<std::mutex> lock(fTimesMutex); | |
| fAllTimes.push_back(dt); | |
| fAllSteps.push_back(nsteps); | |
| } |
| void PrintPhotonTiming() | ||
| { | ||
| int n = fPhotonCount.load(); | ||
| if (n == 0) return; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| if (n == 0) return; | |
| if (n == 0) | |
| return; |
| if (ssz > 0) | ||
| { | ||
| long long step_sum = 0; | ||
| for (int s : fAllSteps) step_sum += s; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| for (int s : fAllSteps) step_sum += s; | |
| for (int s : fAllSteps) | |
| step_sum += s; |
Comment on lines
+699
to
+702
| << "p10=" << fAllSteps[ssz / 10] | ||
| << ", p50=" << fAllSteps[ssz / 2] | ||
| << ", p90=" << fAllSteps[ssz * 9 / 10] | ||
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| << "p10=" << fAllSteps[ssz / 10] | |
| << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] | |
| << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; | |
| << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2] | |
| << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl; |
Comment on lines
+718
to
+723
| std::cout << "Geant4: StepTime " << std::setw(15) << name | ||
| << ": count=" << std::setw(10) << count | ||
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | ||
| << total_ns / 1000.0 / count << " us" | ||
| << " total=" << std::setprecision(3) << std::setw(8) | ||
| << total_ns / 1e9 << " s" << std::endl; |
Contributor
There was a problem hiding this comment.
clang-format suggestion
Suggested change
| std::cout << "Geant4: StepTime " << std::setw(15) << name | |
| << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) | |
| << total_ns / 1000.0 / count << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) | |
| << total_ns / 1e9 << " s" << std::endl; | |
| std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count | |
| << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count | |
| << " us" | |
| << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl; |
plexoos
requested changes
Apr 13, 2026
| @@ -0,0 +1,114 @@ | |||
| # GPU vs G4 Optical Photon Benchmark | |||
|
|
|||
| Measures the speedup of GPU (OptiX 7) optical photon propagation compared to | |||
Member
There was a problem hiding this comment.
Is this benchmark really for OptiX 7 only?
Member
There was a problem hiding this comment.
This new script looks like a benchmark measurement. Let's move it to benchmarks/
| cudaDeviceSynchronize(); | ||
| auto end = std::chrono::high_resolution_clock::now(); | ||
| std::chrono::duration<double> elapsed = end - start; | ||
| std::cout << "Simulation time: " << elapsed.count() << " seconds" << std::endl; |
Member
There was a problem hiding this comment.
Make up your mind about whether to use std::cout or G4cout consistently. I would suggest using std::cout.
…on_timing_summary.py Move the inline python3 -c "..." block out of examples/photontiming_geant4/photontimingandsteps.sh into optiphy/ana/photon_timing_summary.py. Shell now only orchestrates the three benchmark runs and parses log values; Python module owns the speedup formatting and is testable standalone. Behavior verified identical against same input values for both typical and zero cases.
Contributor
There was a problem hiding this comment.
Cpp-linter Review
Used clang-format v20.1.2
Click here for the full clang-format patch
diff --git a/src/GPURaytrace.cpp b/src/GPURaytrace.cpp
index 64386ca..be1dd60 100644
--- a/src/GPURaytrace.cpp
+++ b/src/GPURaytrace.cpp
@@ -78,3 +78 @@ int main(int argc, char **argv)
- program.add_argument("--skip-gpu")
- .help("skip GPU photon propagation (for measuring G4-only photon time)")
- .flag();
+ program.add_argument("--skip-gpu").help("skip GPU photon propagation (for measuring G4-only photon time)").flag();
diff --git a/src/GPURaytrace.h b/src/GPURaytrace.h
index b61f0d6..ccad899 100644
--- a/src/GPURaytrace.h
+++ b/src/GPURaytrace.h
@@ -483,4 +483,20 @@ struct SteppingAction : G4UserSteppingAction
- if (pname == "Transportation") { fTimeTransport += dt; fCountTransport++; }
- else if (pname == "OpWLS") { fTimeOpWLS += dt; fCountOpWLS++; }
- else if (pname == "OpRayleigh"){ fTimeOpRayleigh += dt; fCountOpRayleigh++; }
- else if (pname == "OpAbsorption"){ fTimeOpAbsorption += dt; fCountOpAbsorption++; }
+ if (pname == "Transportation")
+ {
+ fTimeTransport += dt;
+ fCountTransport++;
+ }
+ else if (pname == "OpWLS")
+ {
+ fTimeOpWLS += dt;
+ fCountOpWLS++;
+ }
+ else if (pname == "OpRayleigh")
+ {
+ fTimeOpRayleigh += dt;
+ fCountOpRayleigh++;
+ }
+ else if (pname == "OpAbsorption")
+ {
+ fTimeOpAbsorption += dt;
+ fCountOpAbsorption++;
+ }
@@ -507 +523,2 @@ struct SteppingAction : G4UserSteppingAction
- if (fSkipGenstep) return; // skip genstep collection for timing-only runs
+ if (fSkipGenstep)
+ return; // skip genstep collection for timing-only runs
@@ -637 +654,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt < cur && !fMinPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -639 +658,3 @@ struct TrackingAction : G4UserTrackingAction
- while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt)) {}
+ while (dt > cur && !fMaxPhotonTime.compare_exchange_weak(cur, dt))
+ {
+ }
@@ -641,6 +662,12 @@ struct TrackingAction : G4UserTrackingAction
- if (dt < 1000) fTimeBucket0++;
- else if (dt < 10000) fTimeBucket1++;
- else if (dt < 100000) fTimeBucket2++;
- else if (dt < 1000000) fTimeBucket3++;
- else if (dt < 10000000) fTimeBucket4++;
- else fTimeBucket5++;
+ if (dt < 1000)
+ fTimeBucket0++;
+ else if (dt < 10000)
+ fTimeBucket1++;
+ else if (dt < 100000)
+ fTimeBucket2++;
+ else if (dt < 1000000)
+ fTimeBucket3++;
+ else if (dt < 10000000)
+ fTimeBucket4++;
+ else
+ fTimeBucket5++;
@@ -649 +676,5 @@ struct TrackingAction : G4UserTrackingAction
- { std::lock_guard<std::mutex> lock(fTimesMutex); fAllTimes.push_back(dt); fAllSteps.push_back(nsteps); }
+ {
+ std::lock_guard<std::mutex> lock(fTimesMutex);
+ fAllTimes.push_back(dt);
+ fAllSteps.push_back(nsteps);
+ }
@@ -656 +687,2 @@ struct TrackingAction : G4UserTrackingAction
- if (n == 0) return;
+ if (n == 0)
+ return;
@@ -691 +723,2 @@ struct TrackingAction : G4UserTrackingAction
- for (int s : fAllSteps) step_sum += s;
+ for (int s : fAllSteps)
+ step_sum += s;
@@ -699,4 +732,2 @@ struct TrackingAction : G4UserTrackingAction
- << "p10=" << fAllSteps[ssz / 10]
- << ", p50=" << fAllSteps[ssz / 2]
- << ", p90=" << fAllSteps[ssz * 9 / 10]
- << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
+ << "p10=" << fAllSteps[ssz / 10] << ", p50=" << fAllSteps[ssz / 2]
+ << ", p90=" << fAllSteps[ssz * 9 / 10] << ", p99=" << fAllSteps[ssz * 99 / 100] << std::endl;
@@ -718,6 +749,4 @@ inline void RunAction::PrintTimingReport()
- std::cout << "Geant4: StepTime " << std::setw(15) << name
- << ": count=" << std::setw(10) << count
- << " avg=" << std::fixed << std::setprecision(2) << std::setw(8)
- << total_ns / 1000.0 / count << " us"
- << " total=" << std::setprecision(3) << std::setw(8)
- << total_ns / 1e9 << " s" << std::endl;
+ std::cout << "Geant4: StepTime " << std::setw(15) << name << ": count=" << std::setw(10) << count
+ << " avg=" << std::fixed << std::setprecision(2) << std::setw(8) << total_ns / 1000.0 / count
+ << " us"
+ << " total=" << std::setprecision(3) << std::setw(8) << total_ns / 1e9 << " s" << std::endl;
Have any feedback or feature suggestions? Share it here.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Add optical photon performance profiling to a new example based GPURaytrace:
Add benchmark script (examples/photontiming_geant4/photontimingandsteps.sh) that correctly measures GPU vs G4 speedup using three runs:
Measured on apex.gdml (10 MeV electron, ~250k photons, RTX 4090):
Shall be merged after wavelength shifting is