4.4.1 Out of Memory & 4.2.3.2 Move Delay Due to Supply Calculation

[rkr1958]

4.1.1 Out of Memory & 4.2.3.2 Move Delay Due to Supply Calculation

I know my title is a mouth full but I always try to capture my issue in the title. I've got a lot going on but please bear with me.

By the way I need to amend my title to,"4.4.1 Out of Memory & 4.2.3.2 Move Delay Due to Supply Calculation NEITHER of which are issues with 3.2.0."

In my next post I will attach a zip with 2 games associated with the 2 screen caps below. Both game files will cause 4.4.1 to crash with what I suspect is an out of memory error in the supply calculation (though no crash report is produced). I've documented this crash/bug before in https://www.matrixgames.com/forums/tm.asp?m=5127784

The loss of the allied (Free French) supply CP in the Arabian Sea (bottom cap) introduces significant delays when you move certain AXIS units with 4.2.3.2. But no delays with 3.2.0. I will quantify this in my next post.

We did find it interesting, and very frustrating, that the loss of allied supply through the Arabian sea impacted the times it took to move axis units.

[rkr1958]

Zip file of CP-ArabianSea-Typical.GAM & NoCP-ArabianSea-Delays.GAM games attached.

To "quantify" the move delays we're seeing I timed how long it took to move 6 German units in Russia and 4 Italian units in France. The specific moves were:

Russia. German Units.
(1) V Mot corps. 40,50 to 38,52.
(2) Rommel HQ-A. 44,55 yo 47,54.
(3) XLVII Arm corps. 53,58 to Nikolayev 57,57.
(4) XLI Arm corps. 55,58 to Nikolayev 57,57.
(5) Mot div. 56,58 to Nikolayev 57,57.
(6) 1st Arm div. 51,56 to 56,58.

France. Italian Units.
(7) Cavallero HQ-I. 59,28 to 58,28.
(8) Badoglio HQ-I. 58,28 to 57,27.
(9) African Mech corps. 58,30 to 58,28.
(10) 3rd Inf div. 58,28 to 59,28.

I need to caveat the following timing results. There's learning involved that affects times as I went through these 10 moves at least a half of dozen times. There's inherent "noise" is pointing-clicking-moving-pointing-clicking too. However; with all that said I think the timing differences are stark and informative.

4.2.3.2
(1) CP-ArabianSea-Typical.GAM, 2 minutes, 40 seconds, Avg = 16 sec/move
(2) NoCP-ArabianSea-Delays.GAM, 5 minutes, 22 seconds, Avg = 32 sec/move (100% longer or double the time to make these 10 moves)

3.2.0.
(3) CP-ArabianSea-Typical.GAM, 1 minute, 40 seconds, Avg = 10 sec/move
(4) NoCP-ArabianSea-Delays.GAM, 1 minute, 30 seconds, Avg = 9 sec/move (considering "noise" no difference between these two cases with 3.2.0. Also, it took 60% to 320% longer to make these moves with 4.2.3.2.)

[michaelbaldur]

sorry to say that this is a really old issue ..

do you know that the calculation times are shown

the circled number is the time in Mili seconds

[rkr1958]

ORIGINAL: michaelbaldur
do you know that the calculation times are shown

Is that with the debug version? Unless I'm dense, which I may be, I'm not seeing those numbers with the version I have.

[Centuur]

ORIGINAL: rkr1958

ORIGINAL: michaelbaldur
do you know that the calculation times are shown

Is that with the debug version? Unless I'm dense, which I may be, I'm not seeing those numbers with the version I have.

Yes, it's only available in debug versions.

[rkr1958]

I thought I'd repeat my timing experiment to include the latest MWIF version 5.0.0.1 (Beta) with the caveat that I've upgraded my PC since the original timing experiment. So I ran the experiment on the two game file configurations (CP & No CP in the Arabian Sea) for MWIF versions 3.2.0, 4.2.3.2, 5.0.0.1.

New & Old PC Specs:

Processor:
(New) Intel i9-12900, 2.4 GHz w/16 cores & 4 threads
(Old) Intel i5-3340, 3.10 GHz w/4 cores & 4 threads

RAM:
(New) 64 GB
(Old) 8 GB

Main Harddrive.
(New) Solid State
(Old) SATA

O/S
(New) Windows Home 11, 64-bit
(Old) Windows Home 10, 64 bit

Timing Summary

Timing-Summary.jpg (31.77 KiB) Viewed 309 times

Summary
(1) Maybe a little crow eating on my part is due?
(2) My new PC configuration definitely speeds things up significantly. Supply calculation times are cut in half with 3.2.0 and are within 15% of 3.2.0 with 4.2.3.2 & 5.0.0.1 using the "CP in Arabian Sea" game file.
(3) However; supply calculation times are still almost doubled (1.8 x) for the "No CP in Arabian Sea" game file for 4.2.3.2 & 5.0.0.1 but remain the same for 3.2.0.
(4) So I think it's definitely time for me to play with the latest (5.0.0.1). That is, all the latest fixes and additionally coded optional rules are worth it to me for the 15% slower supply calculation times.
(5) Can anyone recommend a good brand of crow to eat?