Extended HI followup: fix speed issue #10733

yujiex · 2024-09-12T17:03:27Z

Pull request overview

This PR fixes the speed problem of root solver in extended HI calculation. The original PR for extended heat index is here: Adopt the Extended Heat Index calculation in zone resilience #10548

A couple of different approaches were tried out to improve the performance

changing the root solver type from Bisection to BisectionThenRegulaFalsi: this had the biggest impact, bringing the damage from 70% to about 12%
result caching: saving Zs(Rs) result to a variable before entering the root solvers. This improved about 1-2 percentage point.
limit the domain of application to the region where the two methods differ more (red and blue region). That doesn't help much either (< 0.5% improvement).

The new extend HI method might just requires a lot of computation. So The following object is added to provide a switch from the simplified to the extended HI method.

OutputControl:ResilienceSummaries,
       \memo Specifies methods for resilience reporting variables
       \unique-object
  A1 ; \field Heat Index Algorithm
       \note Whether the simplified or the extended method should be used for heat index
       \note Simplified: based on regression analysis carried out by Lans P. Rothfusz
       \note Extended: Based on paper by Lu & Romps
       \type choice
       \key Simplified
       \key Extended
       \default Simplified

Regression diff

There's diff in one file: 1ZoneUncontrolled_Win_ASH55_Thermal_Comfort.idf. The diff is a result of the added OutputControl:ResilienceSummaries

NOTE: ENHANCEMENTS MUST FOLLOW A SUBMISSION PROCESS INCLUDING A FEATURE PROPOSAL AND DESIGN DOCUMENT PRIOR TO SUBMITTING CODE

Pull Request Author

Add to this list or remove from it as applicable. This is a simple templated set of guidelines.

Reviewer

This will not be exhaustively relevant to every PR.

…ilience"" This reverts commit a0f92d6.

github-actions · 2024-09-12T18:27:39Z

⚠️ Regressions detected on macos-14 for commit `a71113f`

Regression Summary

ESO Big Diffs: 4
Table Big Diffs: 3

github-actions · 2024-09-13T01:25:43Z

⚠️ Regressions detected on macos-14 for commit `942bb89`

Regression Summary

ESO Big Diffs: 4
Table Big Diffs: 3

github-actions · 2024-09-13T04:44:30Z

⚠️ Regressions detected on macos-14 for commit `ffd2aa0`

Regression Summary

ESO Big Diffs: 4
Table Big Diffs: 3

This reverts commit 58a77d9.

github-actions · 2024-09-13T07:58:39Z

⚠️ Regressions detected on macos-14 for commit `59b978e`

Regression Summary

ESO Big Diffs: 4
Table Big Diffs: 3

Myoldmopar · 2024-09-13T13:21:46Z

Even with the latest commit, 4f4809f, it's still causing a 10% increase in instruction count in just our performance test files. We can figure out a solution for this, but it doesn't look like it's going in this release. I'm marking this for next release and we'll deal with this later.

If this method is trimmed as far down as possible but still this much slower, I would suggest adding a switch to calculate heat index using the old way by default and then enabling the new method as an alternative. This would be a user-facing input option, so definitely not in this current release.

yujiex · 2024-09-13T15:38:38Z

Thanks, @Myoldmopar . It seems that changing the solver type had the biggest impact. Then I tried adding some result caching but those doesn't save as much (like improved 1-2 percentage point). I also tried to limit the domain of its application (in the old PR there was a plot of how results differ between the old and the new methods. I restricted it to only apply on the bottom right region where the two methods differ a lot). That doesn't help much either. Potentially the new method just requires a lot of computation.

(modify): I might have had the equation part wrong for restraining the domain where extended HI is applied (the white region in the following figures. I modified it and tried locally, roughly it improved it by about 6-7%. Might still not be good enough.

…dHIspeedFix

This reverts commit 6600069.

This reverts commit ec8f0ba.

This reverts commit 16faa31.

github-actions · 2024-09-13T17:30:19Z

⚠️ Regressions detected on macos-14 for commit `8808e86`

Regression Summary

ESO Big Diffs: 4
Table Big Diffs: 3

amirroth · 2025-02-08T00:25:56Z

src/EnergyPlus/ExtendedHeatIndex.cc

+    constexpr Real64 epsilon = 0.97;                  //           , emissivity of surface, steadman1979
+    constexpr Real64 M = 83.6;                        // kg        , mass of average US adults, fryar2018
+    constexpr Real64 H = 1.69;                        // m         , height of average US adults, fryar2018
+    Real64 A = 0.202 * pow(M, 0.425) * pow(H, 0.725); // m^2       , DuBois formula, parson2014


Don't use global variables.

also moved heat index calculation into simulation manager after the other OutputControl thing

github-actions · 2025-02-10T08:53:55Z

⚠️ Regressions detected on macos-14 for commit `03996a8`

Regression Summary

Table Big Diffs: 1

github-actions · 2025-02-10T20:43:54Z

⚠️ Regressions detected on macos-14 for commit `576adfb`

Regression Summary

Table Big Diffs: 1

amirroth · 2025-02-10T21:19:39Z

src/EnergyPlus/ExtendedHeatIndex.cc

+        Real64 Rf;
+
+        if (flux1 <= 0.0) {
+            varname = static_cast<int>(EqvarName::Phi);


Why cast this to int all the time? Just use the enum, no?

Yeah I should use the enum class instead of casting to int

amirroth · 2025-02-10T21:21:08Z

src/EnergyPlus/ExtendedHeatIndex.cc

+    }
+
+    // Convert the find_T function
+    Real64 find_T(EnergyPlusData &state, int const eqvar_name, Real64 const eqvar)


You can use enums as parameters, why make it an int here?

yeah also need to use enum class directly here as well

amirroth · 2025-02-10T21:38:39Z

src/EnergyPlus/ExtendedHeatIndex.hh

+    enum class EqvarName
+    {
+        Invalid = -1,
+        Phi = 1,


Why are these not starting at 0? Why do you have to assign explicit values to them?

This extended heat index was based on python code on this page. They have some dictionary defined the indices like this: dic = {"phi":1,"Rf":2,"Rs":3,"Rs*":3,"dTcdt":4} So I just used the same indices.

The values are meaningless unless you are using them in an array, but even then you should use them in a std::array and they should start at 0.

make sense. Let me change this to starting at 0.

amirroth · 2025-02-10T21:39:36Z

src/EnergyPlus/HeatBalanceManager.cc

+                    static_cast<HVACSystemRootSolverAlgorithm>(getEnumValue(HVACSystemRootSolverAlgorithmUC, Util::makeUPPER(AlphaName(1))));
+                if (HVACSystemRootFinding.HVACSystemRootSolverMethod == HVACSystemRootSolverAlgorithm::Invalid) {
+                    HVACSystemRootFinding.HVACSystemRootSolverMethod = HVACSystemRootSolverAlgorithm::RegulaFalsi;
+                    ShowWarningError(state,


There is a shortcut for this warning called ShowWarningInvalidKey.

Oh! Didn't realize there's this shortcut.

amirroth · 2025-02-10T21:44:24Z

src/EnergyPlus/ExtendedHeatIndex.cc

+    constexpr Real64 Za_bar = 60.6 / 11.6; // Pa m^2/W, mass transfer resistance through air, clothed part of skin
+    constexpr Real64 Za_un = 60.6 / 12.3;  // Pa m^2/W, mass transfer resistance through air, when being naked
+
+    constexpr Real64 tol = 1e-8;


This is an absolute tolerance level. What are the reasonable ranges of values here and what level of precision does this tolerance correspond to relative to those values?

The tolerance was from the python code here this page.

What someone else chose for the tolerance is not relevant. What is the magnitude of the values produced by this function? 1e-4? 1e-1? 1e+2?

If this function produces SGP (whatever that is) and the values are on the order of 1e-2 then a tolerance of 1e-8 is about six decimal places of precision which is about twice what is needed. I cannot conceive of any quantity that EnergyPlus that would need to know to a precision of more than 0.1% (1% may be okay for many thing too).

In fact, SolveRoot should probably be revised to use relative tolerance after some small fixed number of iterations. I wonder how much performance over-specific root precision is costing us. It could be quite a lot.

If I change tol to 1e-4, there will be a 0.6 difference for a target value of about 360 (0.2% relative difference). Using tol=1e-5 keeps the difference to below 0.1 (for values ranging from 200 to 400). Maybe I'll change it to tol = 1e-5?

Absolutely. Those additional digits are costing you iterations for no real benefit in terms of accuracy. Who cares if the SGP is "off" by less than 0.1%? How accurate is that metric anyway?

Normally heat index is reported at either integer or 1 decimal point. The range is mostly two three hundred. Hopefully within 0.1C accuracy is good enough.

Most thermostats used in buildings do not have this resolution.

yeah the inputs of temperature and relative humidity percentage points are integers

github-actions · 2025-02-10T23:38:13Z

⚠️ Regressions detected on macos-14 for commit `7d0f7b4`

Regression Summary

Table Big Diffs: 1

github-actions · 2025-02-11T00:11:43Z

⚠️ Regressions detected on macos-14 for commit `38a54c0`

Regression Summary

Table Big Diffs: 1

github-actions · 2025-02-11T19:34:27Z

⚠️ Regressions detected on macos-14 for commit `d9c2553`

Regression Summary

Table Big Diffs: 1

mjwitte

@yujiex Thanks for the idd changes.

mjwitte · 2025-02-12T16:47:07Z

doc/input-output-reference/src/input-for-output.tex

+
+This input specifies methods or algorithms used in certain resilience reporting variables.
+
+\subsubsection{Inputs}\label{inputs-8-020}


When viewing the changes on github, there is a warning about this line:
Label inputs-8-020' multiply defined.`

let me check on why this happens

I see, my label for the inputs section in my latex documentation collides with a existing one. I just changed the label.

github-actions · 2025-02-12T19:36:18Z

⚠️ Regressions detected on macos-14 for commit `af75116`

Regression Summary

Table Big Diffs: 1

Myoldmopar

Outside of the vector/array thing in the unit test code, I don't have any comments. And that's not worth holding this up. I will do final testing and try to get this in.

Myoldmopar · 2025-02-13T16:35:38Z

src/EnergyPlus/ExtendedHeatIndex.cc

+    }
+
+} // namespace ExtendedHI
+} // namespace EnergyPlus


The strong use of const and constexpr in this file make me very happy.

Myoldmopar · 2025-02-13T16:36:49Z

src/EnergyPlus/SimulationManager.cc

@@ -1013,6 +1013,27 @@ namespace SimulationManager {
            }
        }

+        state.dataHeatBalMgr->CurrentModuleObject = "OutputControl:ResilienceSummaries";


In the future, there's no need to use a state variable to hold this string. Just keep a constexpr string view variable of it and use that.

I'll remember and use string view in the future.

Myoldmopar · 2025-02-13T16:39:00Z

tst/EnergyPlus/unit/ExtendedHI.unit.cc

+TEST_F(EnergyPlusFixture, extendedHI_pvstar)
+{
+    Real64 tol = 1e-8;
+    std::vector<Real64> T_values = {200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 310, 320, 330, 340, 350, 360, 370};


You could use constexpr std::array all over the place in here. Since this is just in the tests, we don't have to hold this up, but always look for that in the future.

oh! I see I see. I'll take a note and remember to use constexpr next time.

Myoldmopar · 2025-02-13T16:49:08Z

Runs great here, ready to go in. Thanks @yujiex.

yujiex · 2025-02-13T17:51:22Z

Thanks @Myoldmopar

change rootsolver to BisectionThenRegulaFalsi to speed up

28bdb41

yujiex added the Defect Includes code to repair a defect in EnergyPlus label Sep 12, 2024

yujiex requested a review from Myoldmopar September 12, 2024 17:03

yujiex self-assigned this Sep 12, 2024

Yujie Xu added 3 commits September 12, 2024 10:07

Merge remote-tracking branch 'origin/develop' into extendedHIspeedFix

b9fe27e

Revert "Revert "Adopt the Extended Heat Index calculation in zone res…

6e8fcc8

…ilience"" This reverts commit a0f92d6.

change root solver to BisectionThenRegulaFalsi for speed

b23f9ad

yujiex added Performance Includes code changes that are directed at improving the runtime performance of EnergyPlus and removed Defect Includes code to repair a defect in EnergyPlus labels Sep 12, 2024

Yujie Xu added 4 commits September 12, 2024 16:19

save Zs(Rs) value in a variable

bd97716

find eqvar name and value together

5e96092

update unit test find_eqvar_name and value

b96c3a5

save Ra_bar calculation result

6600069

restrict domain of extended HI application

58a77d9

yujiex added 2 commits September 12, 2024 23:57

Revert "restrict domain of extended HI application"

16faa31

This reverts commit 58a77d9.

try using a shorter termination for bisection solver

4f4809f

Myoldmopar added this to the EnergyPlus 25.1 IO Freeze milestone Sep 13, 2024

Yujie Xu added 5 commits September 13, 2024 09:07

Merge remote-tracking branch 'origin/extendedHIspeedFix' into extende…

aa7352d

…dHIspeedFix

Try using RegulaFalsiThenBisection

ec8f0ba

Revert "save Ra_bar calculation result"

9e32af5

This reverts commit 6600069.

Revert "Try using RegulaFalsiThenBisection"

f488176

This reverts commit ec8f0ba.

Revert "Revert "restrict domain of extended HI application""

b813d6e

This reverts commit 16faa31.

fix application domain equation

23c3005

amirroth reviewed Feb 8, 2025

View reviewed changes

Yujie Xu added 3 commits February 9, 2025 23:28

Change idd obj name to OutputControl:ResilienceSummaries w doc change

5c7d6fa

also moved heat index calculation into simulation manager after the other OutputControl thing

use getEnumValue for HVACSystemRootSolverMethod

70091eb

remove global variables

d7d4539

Yujie Xu added 2 commits February 10, 2025 10:30

clang-format

02510f2

license year in .hh

ec56408

amirroth reviewed Feb 10, 2025

View reviewed changes

Yujie Xu added 4 commits February 10, 2025 14:06

directly use the enum class EqvarName instead of casting

53491a1

use ShowWarningInvalidKey for warning message

962608f

more removal of global vars

7e51320

clang-format, change enum EqvarName to start from 0

6241263

change root finding tolerance, clang-format

756096b

mjwitte reviewed Feb 12, 2025

View reviewed changes

doc input section latex label fix

7cc4fd6

mjwitte added the IDDChange Code changes impact the IDD file (cannot be merged after IO freeze) label Feb 12, 2025

Myoldmopar approved these changes Feb 13, 2025

View reviewed changes

Myoldmopar merged commit 6342ea8 into develop Feb 13, 2025
10 checks passed

Myoldmopar deleted the extendedHIspeedFix branch February 13, 2025 16:49


		This input specifies methods or algorithms used in certain resilience reporting variables.

		\subsubsection{Inputs}\label{inputs-8-020}

Extended HI followup: fix speed issue #10733

Extended HI followup: fix speed issue #10733

Conversation

yujiex commented Sep 12, 2024 • edited Loading

Pull request overview

Regression diff

Pull Request Author

Reviewer

github-actions bot commented Sep 12, 2024

⚠️ Regressions detected on macos-14 for commit a71113f

github-actions bot commented Sep 13, 2024

⚠️ Regressions detected on macos-14 for commit 942bb89

github-actions bot commented Sep 13, 2024

⚠️ Regressions detected on macos-14 for commit ffd2aa0

github-actions bot commented Sep 13, 2024

⚠️ Regressions detected on macos-14 for commit 59b978e

Myoldmopar commented Sep 13, 2024

yujiex commented Sep 13, 2024 • edited Loading

github-actions bot commented Sep 13, 2024

⚠️ Regressions detected on macos-14 for commit 8808e86

Choose a reason for hiding this comment

github-actions bot commented Feb 10, 2025

⚠️ Regressions detected on macos-14 for commit 03996a8

github-actions bot commented Feb 10, 2025

⚠️ Regressions detected on macos-14 for commit 576adfb

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Feb 10, 2025

⚠️ Regressions detected on macos-14 for commit 7d0f7b4

github-actions bot commented Feb 11, 2025

⚠️ Regressions detected on macos-14 for commit 38a54c0

github-actions bot commented Feb 11, 2025

⚠️ Regressions detected on macos-14 for commit d9c2553

mjwitte left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

github-actions bot commented Feb 12, 2025

⚠️ Regressions detected on macos-14 for commit af75116

Myoldmopar left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Myoldmopar commented Feb 13, 2025

yujiex commented Feb 13, 2025

yujiex commented Sep 12, 2024 •

edited

Loading

⚠️ Regressions detected on macos-14 for commit `a71113f`

⚠️ Regressions detected on macos-14 for commit `942bb89`

⚠️ Regressions detected on macos-14 for commit `ffd2aa0`

⚠️ Regressions detected on macos-14 for commit `59b978e`

yujiex commented Sep 13, 2024 •

edited

Loading

⚠️ Regressions detected on macos-14 for commit `8808e86`

⚠️ Regressions detected on macos-14 for commit `03996a8`

⚠️ Regressions detected on macos-14 for commit `576adfb`

⚠️ Regressions detected on macos-14 for commit `7d0f7b4`

⚠️ Regressions detected on macos-14 for commit `38a54c0`

⚠️ Regressions detected on macos-14 for commit `d9c2553`

⚠️ Regressions detected on macos-14 for commit `af75116`